Overunity.com Archives

Hi everyone,

I have finally created a Ultra lowest simple Joule Thief circuit which self oscillate\start with mere 25mV in my quest for free energy.

Our body radiate heat 24/7 even if we are sleeping.By using peltier we could tap into this heat constantly being generated by our body and the cold side is just the surrounding temperature with small heatsink.
Since i am living in a hot country the temperature difference between my body and room is only 6.5 degrees celsius.
With such a small differenece in temperature i noticed that peltier alone is able to output around 70mV by connecting to multimeter.If you are living in a cooler country then the temperature difference between your body and room temperature would be greater hence peltier would able to produce more power.
My joule Thief boost the tiny voltage to a usable voltage with the possibility tricke charge any portable device.The open circuit voltage from my joule is 8volts output for 50mV input at around 4mA current draw for input.

The peltier could be strapped onto wrist or place into hat for example.

I have posted my Ultra lowest 25mv self starting Joule Thief circuit in youtube under user "sanjev21"

http://www.youtube.com/watch?v=Fzww6yPQMrg (http://www.youtube.com/watch?v=Fzww6yPQMrg)

Very nice but i think a small solar panel gives you much more free energy and you can store it in a battery so you can light the led 24/7.

nice work!

may i suggest something?  take 2 peltiers use one to generate power and use the other to heat, and put them back to back. ie: so the one generating power is feeding its power to the one generating heat. place these both in a well insulated box that you can place a 'bootstrap' heat source in, like a small coal or a hot rock.

i added a led bank to the first one of these i built and used a lm7812 to control the voltage being applied from the one peltier to the other. you can apply this same principle to heat and light a small, well insulated shelter/shed. of course, you need a few peltier combos to do this, but it does work... i have done it. around here the temperature differential between inside and outside can be as much as 70 degrees F... i simply put peltiers in the walls, with one side exposed to the elements.

you can capture solar energy during the day to create a 'bootstrap' temperature differential, or as scratchrobot noted, you could use a solar panel to charge a battery and apply that to the peltier that generates the heat for a 'bootstap'. there are lots of different possibilities for 'bootstrapping'... an earth battery could be used, a simple windmill or a hand cranked generator. just use your imagination!

again, nice work!

Quote from: scratchrobot on December 20, 2012, 10:23:21 AM
Very nice but i think a small solar panel gives you much more free energy and you can store it in a battery so you can light the led 24/7.

What if there was no sun or charge at the time of need when its dark?  ;)

I think its a great idea. And the small difference in temperature that Mag is talking about is small.

Very nice.  ;D

Mags

Quote from: Magluvin on December 20, 2012, 06:51:08 PM
What if there was no sun or charge at the time of need when its dark?  ;)

I think its a great idea. And the small difference in temperature that Mag is talking about is small.

Very nice.  ;D

Mags

You are right and i agree very nice idea and build  :)

Hi magpwr,
Nice idea, i too would like to experiment with peltiers, can you try and put a heater coil from the output of your jt to the peltier, and see if you can loop it with its own power?


Regards
Cc
@Mags
Thought it was you who started this thread  ;D

Hi,

We can use external heat source but it will defeat the meaning of "Free energy" which my circuit was intended for .Circuit output can't generate enough power to heat peltier again.
Peltier are only 50% efficient if used as a cooler.

Although there could be another possibility like using "crystal radio" to tap into "natural radio wave" and broadcast radio wave which may produce in milli volts to power my circuit to boost voltage
to useable level.

Quote from: magpwr on December 21, 2012, 01:36:47 AM
Hi,

We can use external heat source but it will defeat the meaning of "Free energy" which my circuit was intended for .Circuit output can't generate enough power to heat peltier again.
Peltier are only 50% efficient if used as a cooler.

Although there could be another possibility like using "crystal radio" to tap into "natural radio wave" and broadcast radio wave which may produce in milli volts to power my circuit to boost voltage
to useable level.

I did not even notice the 25mv, that's very low!
I also have a Peltier somewhere lying around and maybe i can power my crystal with it, your idea to tap into some wave is also nice.
Can you tell me why there are 4 fet's in your circuit?

Thanks

Quote from: magpwr on December 21, 2012, 01:36:47 AM
Hi,

We can use external heat source but it will defeat the meaning of "Free energy" which my circuit was intended for .Circuit output can't generate enough power to heat peltier again.
Peltier are only 50% efficient if used as a cooler.

Although there could be another possibility like using "crystal radio" to tap into "natural radio wave" and broadcast radio wave which may produce in milli volts to power my circuit to boost voltage
to useable level.

Have you looked at the JT threads? Where gadgetmalls jt circuit 1.5v AA battery can light modified cfl's, christmas lights etc...if your jt is like that maybe its posible?

Hi,

To Crazycut06,
This joule thief could output around 30v at 1V input. But my version 1.0 joule thief is design for low power/current.

I have already created a high power version but i realise peltier can't produce high current at such a low voltage like body heat.

If you are talking about "ultimate high power joule thief" to run from battery down to below <100mV .Yes it is already made &tested but still in improvement stage.
But that  is reserve for  future  posting version 2.0 but input current is higher at 150mA . I have also created version 3.0 which have certain improvement over 2.0.
The toroid remain the same in any case.

To Scratchrobot,
The reason why i am using 4JFET in parallel is to increase current output.Using 1 fet current consumption is only around 1mA and obvioiusly the output current is lower.THE jfet was not design to handle high current.

Quote from: magpwr on December 21, 2012, 07:15:42 AM

To Scratchrobot,
The reason why i am using 4JFET in parallel is to increase current output.Using 1 fet current consumption is only around 1mA and obvioiusly the output current is lower.THE jfet was not design to handle high current.

So I could also use just  one jfet with higher current rating and the result would be the same?

To Scratchrobot,

You can use 1 JFet and get lower current ouput.These FET are not design for high current.

You will need to experiment to understand the logic on why i was using 4 JFET.

serves no purpose in the real world.

To cuttinedge2005,

I think you have not calculated all the random possibility before replying to me.

Let me tell you where this circuit could be applied not in urban location where you assume there  is power even during disaster.
1)Industry application-monitor furance or engine temperature remotely without battery then transmit the information to a central location or control room remotely on interval basis.
2)For military or for hikers out in forested area where is no direct sunlight if you are on the move.
Gps or or other equipment could be trickle charged.
3)If peltier is installed into miners hat.It could serve as permanent emergency light even if the battery pack fails.

Quote from: magpwr on December 21, 2012, 10:41:05 PM
To Scratchrobot,

You can use 1 JFet and get lower current ouput.These FET are not design for high current.

You will need to experiment to understand the logic on why i was using 4 JFET.

Thanks i will as soon i have the jfets. Can you tell me the values of the caps you used because i cant read them in your video.

Quote from: onthecuttingedge2005 on December 22, 2012, 12:49:59 AM
serves no purpose in the real world.

At least this circuit is working and gives energy unlike most circuits here.
Not everything has to have a purpose.

Hi Scratchrobot,

The input capacitor is 100uf and output i'm using 1uf(105 written on cap) for quick increase in voltage you could use higher cap like 100uf at output stage as well and notice how fast the voltage will increase.
Or if your are just powering led 100nf caps will do as well at the output stage.

The toroid and circuit could be shrunk alot more ,it's up to individual builder. Prototype is always bigger/ugly initially since we don't know if it would work the way we want it.


I am indeed  bad at hand writing  rarely get the chance to write these days unless i am filling in some official document.

Quote from: magpwr on December 22, 2012, 09:21:50 PM
Hi Scratchrobot,

The input capacitor is 100uf and output i'm using 1uf(105 written on cap) for quick increase in voltage you could use higher cap like 100uf at output stage as well and notice how fast the voltage will increase.
Or if your are just powering led 100nf caps will do as well at the output stage.

The toroid and circuit could be shrunk alot more ,it's up to individual builder. Prototype is always bigger/ugly initially since we don't know if it would work the way we want it.

I am indeed  bad at hand writing  rarely get the chance to write these days unless i am filling in some official document.

I'm waiting for my fets to arrive and when i have them i will build your circuit and report back here.
Thanks again for sharing your circuit.

The Fets arrived and i tried replicating your circuit but without succes. I will try making another coil.

Be sure to try all combinations of coil connections before you change other stuff. Say you have coil A with connections 1 and 2, and Coil B ditto.
So you have 4 possible ways to hook up the coils; two should work and two won't, and of the two that do, one might be a little better.

I like this circuit and the use of peltier elements.

Unfortunately I have to get this special transistors before I can start experimenting. I have a good collection of peltier elements. My idea is to stick the peltier element to a cold window glass and a longish aluminium "heater" warmed up by the air in the warm room.

The use of 4 transistors is still a riddle for me: my guess is that one transistor would suffice once the one amongst the four is found which switches most easily with this low Voltage and low current at the base. But I could be wrong. The argument, that 4 transistors give more power than 1 in case they do not switch on nicely is not very good, because 4 times very low power would still not be enough. The trick is to make the transistor switch on inspite of low Voltage and low current at the base. Once that is achieved, it does not matter whether it is 1 or 4 transistors. There could be a strange interaction between the bases of the four transistors, but this seems to be esotheric. What ever, in case 4 work, why not.

Greetings, Conrad

Theory concerning 4 transistors in parallel versus just 1 transistor:

The base of a transistor is also a small parasitic capacitor. Having 4 transistors (4 bases) in parallel will multiply this parasitic capacitor by 4. A charge will accumulate in this capacitor and it will flow into the one transitor that triggers by chance as the first one (the other 3 will not trigger and at each cycle an other transistor of the four transistors could trigger but my guess is that always the same triggers due to very fine diferences in the production process).

The charge is sufficient to trigger one transistor if it is accumulated in 4 bases but is not sufficient if it can only accumulate in one base.

Putting a 1nF capacitor in parallel to the 1-8K base resistor should do the same. May be a smaller capacitor of e.g. 500 pF will be better.

Greetings, Conrad

Hi Conrad,

If I may chime in here,  no need for thinking that the 4 transistors in parallel may behave esotheric. The 4 in parallel has 2 useful benefits: the resultant drain-source channel ON resistance becomes 1/4 of that of a single transistor so that the power loss (I*I*R) due to the smaller channel resistance is also 1/4 only.
The second benefit is that the resultant forward transconductance (or forward transfer admittance) (dVgs/dId) becomes 4 times as high than that of a single transistor. 
So the reason for paralleling them is to improve manufacturing "deficiences" on certain important parameters for an oscillator. Of course unwanted parameters also occur when paralleling: input and (output) capacitances get also 4 times as much, this would not be needed at all...  should a device in a circuit need higher input capacitance, you could always add some to see any benefit.

Here is a data sheet for the 2sk170:  http://www.audiodesignguide.com/DACend2/2sk170.pdf (http://www.audiodesignguide.com/DACend2/2sk170.pdf)  and you can order cheaply from futurlec.com for instance.

Greetings,
Gyula

Quote from: gyulasun on December 30, 2012, 12:27:05 PM
Hi Conrad,

Here is a data sheet for the 2sk170:  http://www.audiodesignguide.com/DACend2/2sk170.pdf (http://www.audiodesignguide.com/DACend2/2sk170.pdf)  and you can order cheaply from futurlec.com for instance.

Greetings,
Gyula

@Gyula: thank you for your input. Looking at the data sheet I see a "reverse transfer capacitance CRSS" of 6 pF. So, a 22 pF capacitor parallel to the 1-8 K base resistor should do the same. But there might be the other considerations you mentioned. It has to be tested, theory is just words.

Greetings, Conrad

Quote from: TinselKoala on December 30, 2012, 09:07:43 AM
Be sure to try all combinations of coil connections before you change other stuff. Say you have coil A with connections 1 and 2, and Coil B ditto.
So you have 4 possible ways to hook up the coils; two should work and two won't, and of the two that do, one might be a little better.

Thanks but i tried every possible way but nothing was working for me. I will try again tomorrow and also hook up a scope to see if i get oscilations. Not easy testing this with one finger on the peltier, Magpwr where are you  ;-)

Quote from: scratchrobot on December 30, 2012, 01:24:23 PM
Thanks but i tried every possible way but nothing was working for me. I will try again tomorrow and also hook up a scope to see if i get oscilations. Not easy testing this with one finger on the peltier, Magpwr where are you  ;-)

Test with a stable voltage source like a regulated power supply or a battery. When the circuit works, then sub in the Peltier.

Hi scratchrobot,

I would like to check with you only 3 things to find out why circuit isn't working for you.

1)Are you using the most ideal N Channel Juntion FET-2SK170 (1 or 4 should work as well).I have not mentioned in video.Please don't bother trying with other N-Channel JFets(>30 varieties ) the result isn't good (not as low startup/osc)or worst don't even osc at 100mV or can't simply handle low impedance at primary(4 turns).

2)Is your primary (4 TURNS) and Secondary is around 80...100(Even 70 turns do work on smaller toroid,but open circuit voltage would be lower,with 10mm led voltage remains same around 2.3v ...2.4v)
May need to swap secondary turns ( One of the 2 output to positive)

3)Have you verifed the output voltage from peltier(+ - output with hand on itand heatsink) before even connecting to circuit.Please google datasheet of N-Channel J-FET to identify pinout(Drain ,Gate ,Source = C ,B ,E)Lastly peltier in itself(not with hand or on heatsink) ,out of the box should have low impedance reading <5Ohms.
   My peltier is 3Ohms .My toroid primary is <1Ohms with 4 turns.I think in theory peltier which got lower impedance may produce good result.


Lastly you could check toroid connection merely using common NPN transistor with 1.5V battery if there is no output just swap secondary (one of the 2 wires to + and primary one of 2 wire to +)
 
Have a Happy New Year.

Quote from: magpwr on December 31, 2012, 08:40:24 AM
Hi scratchrobot,

I would like to check with you only 3 things to find out why circuit isn't working for you.

1)Are you using the most ideal N Channel Juntion FET-2SK170 (1 or 4 should work as well).I have not mentioned in video.Please don't bother trying with other N-Channel JFets(>30 varieties ) the result isn't good (not as low startup/osc)or worst don't even osc at 100mV or can't simply handle low impedance at primary(4 turns).

2)Is your primary (4 TURNS) and Secondary is around 80...100(Even 70 turns do work on smaller toroid,but open circuit voltage would be lower,with 10mm led voltage remains same around 2.3v ...2.4v)
May need to swap secondary turns ( One of the 2 output to positive)

3)Have you verifed the output voltage from peltier(+ - output with hand on itand heatsink) before even connecting to circuit.Please google datasheet of N-Channel J-FET to identify pinout(Drain ,Gate ,Source = C ,B ,E)Lastly peltier in itself(not with hand or on heatsink) ,out of the box should have low impedance reading <5Ohms.
   My peltier is 3Ohms .My toroid primary is <1Ohms with 4 turns.I think in theory peltier which got lower impedance may produce good result.


Lastly you could check toroid connection merely using common NPN transistor with 1.5V battery if there is no output just swap secondary (one of the 2 wires to + or swap primary one of 2 wire to +).
If works then use JFET with lower voltage or ultra cap eg:50F 2.5v charged using 1AA for 1..3 seconds to get around 100mV,if you are not using peltier to test.Since Ultra caps eg:20F...50F got very low impedance as well.
 
Have a Happy New Year.

Quote from: magpwr on December 31, 2012, 08:40:24 AM
Hi scratchrobot,

I would like to check with you only 3 things to find out why circuit isn't working for you.

1)Are you using the most ideal N Channel Juntion  (1 or 4 should work as well).I have not mentioned in video.Please don't bother trying with other N-Channel JFets(>30 varieties ) the result isn't good (not as low startup/osc)or worst don't even osc at 100mV or can't simply handle low impedance at primary(4 turns).

2)Is your primary (4 TURNS) and Secondary is around 80...100(Even 70 turns do work on smaller toroid,but open circuit voltage would be lower,with 10mm led voltage remains same around 2.3v ...2.4v)
May need to swap secondary turns ( One of the 2 output to positive)

3)Have you verifed the output voltage from peltier(+ - output with hand on itand heatsink) before even connecting to circuit.Please google datasheet of N-Channel J-FET to identify pinout(Drain ,Gate ,Source = C ,B ,E)Lastly peltier in itself(not with hand or on heatsink) ,out of the box should have low impedance reading <5Ohms.
   My peltier is 3Ohms .My toroid primary is <1Ohms with 4 turns.I think in theory peltier which got lower impedance may produce good result.


Lastly you could check toroid connection merely using common NPN transistor with 1.5V battery if there is no output just swap secondary (one of the 2 wires to + and primary one of 2 wire to +)
 
Have a Happy New Year.

Hi Magpwr,

My mosfet's are J175 so maybe they are not the one's you used and they are the reason the circuit isn't working for me.
My peltier without heatsink reads 5 ohms without a heatsink and with heatsink and one finger gives 25mv easely.
I will try with normal transistor to see if i can get it to work.

Thanks and happy new year to you too.

Hi Scratchrobot,

I'm sorry you had to find out the hard way.I did mentioned somehwhere in youtube description or comment to another interested person whom wish to construct the ultra low voltage circuit.
Never to try any Jxxx series all those don't work.I have done stimulation testing using trial software to find out the best component(It's on you to find out.I won't reveal it).
Please don't bother trying other Junction FETs you won't get ultra low self startup voltage.I have wasted 3 days to accidently discover a self starting joule thief at below <30mV while trying to optimise joule Thief.

I bought 2SK170 from  http://www.futurlec.com (http://www.futurlec.com)  (After searching in internet high and low) for $0.20ea (Bought 10 pieces for $2 USD) Since i'm living in Asia shipment is only around $5.50USD for me.Got it within  2 weeks.

Recent discovery my circuit could be powered by solar cell in room merely shine a led torchlight to part of the 1.7v solar cell(400mA) to light a led again from my ultra low joule thief(NOT OVERUNITY).

I am still waiting for delivery of some 2SK170 transistors. But I tested magpwr's circuit with transistors I had (see attached circuit diagram and photo, photo shows test with a BC547B).

2N3904 : self starts to oscillate at 0.52 Volt  and can then be turned down to 0.17 Volt

2N4401 : self starts to oscillate at 0.5 Volt  and can then be turned down to 0.1 Volt

BC574B : self starts to oscillate at 0.49 Volt  and can then be turned down to 0.14 Volt

I also tried with 10 turns for the collector coil, but 4 turns worked better (same self start voltages, but with 4 turns the LED showed some light with less supply Voltage). 4 and 100 turns seems to be a good ratio for the coils.

The 1 µF capacitor over the LED had no effect (also tried other values like 100 nF).

The resistor at the base influences power consumption very much. With 1 K the power consumption at 0.5 Volt was around 15 mA. With 10 K it was too low to be measured reliably, less than 1 mA.

I have redrawn magpwrs's circuit to give it a more conventional look. The LED is driven by the negative halve of the base coil spike. The positive halve of the spike triggers the transistor.

So, it really seems to be the transistor which is the determining component. On Monday I should get the 2SK170 transistors.

Greetings, Conrad

Hi conradelektro,

Nice work man.This is how 1st started with <100mV Joule Thief using transistor 2SD1450 (0.13v..0.3v VCE(voltage drop between emitter and collector at 400mA.By referring to datasheet and looking at vce
for expected circuit current draw.

If you refer to my youtube channel  "sanjev21" "45mV Joule Thief" was my 1st low power joule thief  using  "2SD1450" Japanese transistor.But it was not self starting and i had to manually short emitter and collector of transistor to kick start oscillation.

My other experiment which i have not posted in youtube using the best NPN transistor i could find in digikey "ZTX1048A" which have the lowest Voltage drop across emitter&collector at mere(27mV...45mV) at 500mA.Little smaller than TO92 and could handle 4Amps or 20Amps pulse no "heat sink".
Using this ultra low drop transistor i could achieve oscillation as low as "30mV" but as usual i have to kickstart oscillation at higher voltage eg:60mV by temporary shorting emitter&collector.

I have also combine ZTX1048A transistor(To boost output current) and (2SK170 or 2SK364 or 3SK366 .But take note 2SK170 do have better lower starting voltage by 1..3mV)  Junction-FET(To self start osc at low voltage).Connecting 22ohms from coil to transistor and 1.8k coil to JFET Gate while connecting (JFET Source to transistor emitter,J-FET Drain to transistor Collector).Design for <100mV.

For this above transistor and JFET combined circuit i was able to charge 1000uf cap in matter of <5sec to reach around 2volts.

But the current drain for above transistor with JFET circuit  is  around 100mA .
I was using 50FARAD 2.5v Ultra capacitor(charged with 1AA battery for few seconds to reach 100mV) as power source.Since it's known ultra cap got very low internal resistance.

The above circuit will work very nice on 1.5V battery but need to tweak 22ohms resistor to higher value to maintain current draw at decent level eg:100mA.
Then again i realised it's pointless using JFET for 1.5v battery since we could use transistor which is more efficient at higher current.

Peltier only produce <5mA if I connected to multimeter if i use my hand as heat source(36.5c) against room temperature in my case(30c for a hot country) 6.5C difference.
The peltier i'm using do have internal resistance of around 3ohms(Out of the box /Not in use).
In theory i'd believe if we use lower internal resistance for peltier <2ohms we may get better current output but then again.I 'm not sure if multi meter could properly detect "current" at such low voltage produce from peltier 70mV..90mV.

Solar cell do work to power my JFET base joule thief in room lighting or led light.I'm not sure if using other power source like from crystal radio could work.

If we match "source and circuit impedance" then we could  further optimize on efficiency.


Lastly the circuit do work on 1 x 2SK170 as well with lower Circuit current draw at 1mA using peltier at around 60mV. I used 4 JFET  since i noticed there is improvement with output voltage especially with Led as load current draw increases to 4.2mA(Mentioned in youtube description).

Lastly if anyone wish to go for extreme low starting voltage at 22mV instead of 25mV got to try 2n3972 JFET (But very low current handling in itself output capacitor cannot be more than 100nf) and combine with 2SK170 x 3 or 4.

If anyone is a expert with SMD base tiny circuits there is UK base company which manufacture 2SK170 as LSK170  (TO92 package and SMD tiny version).
http://www.micross.com/packaged-parts-plastic-discretes.aspx (http://www.micross.com/packaged-parts-plastic-discretes.aspx)

@magpwr:

I just realised that I can start oscillation of your circuit with a 2N3904 transistor at 0.2 Volt by briefly shorting collector / emitter.

I found a source for the 2SK170 transistor in Germany

http://www.reichelt.de/ (http://www.reichelt.de/)

and I would be able to get the ZTX1048A transistor from Farnell.

At what Voltage does the ZTX1048A self start? (You mentioned that you could start it at 60 mV by briefly shorting collector / emitter.)

I did not find a source for the 2n3972?

There should be a replacement for the 2SK170, but I did not find any?

Greetings, Conrad

http://www.utsource.net/ic-datasheet/2N3972-677148.html
9 dollars US !!

http://www.utsource.net/ic-datasheet/2SK170-1553154.html
2 dollars.

I've made 2 orders with these people, for mosfets and strange transistors for my F43 FG.... the ordering is painless and the shipping for small quantities is only 4 dollars, but it takes 7-10 business days for the parts to arrive. The parts I ordered were cheaper here than anywhere else for small quantities, and they have stuff that nobody else seems to have, like the Motorola transistors I needed.

I did some power consumption measurements:

Magpwr's circuit as shown in myReply #29 with a 2N3904 transistor uses about 1 mA at 0.2 Volt, which is about 200 µW. The white LED glows dimly.

This is the lowest supply Voltage I personally could ever achieve, but it has to be started by briefly shorting collector / emitter. But I could achieve 50 µW with a 1 Volt power supply with an other circuit, see my Reply #15947 at http://www.overunity.com/6123/jule-thief/15945/#.UOcYyne7kaw (http://www.overunity.com/6123/jule-thief/15945/#.UOcYyne7kaw) (Jule Thief thread page 1064).

I am really looking forward to test the 2SK170 transistor. A self starting Joule Thief type circuit with a power supply of less than 0.1 Volt (even if it needs 1 mA) would be really good for my self made batteries (salt water as electrolyte, silver coated copper wire or graphite chunk versus magnesium chunk or aluminium plate or galvanised steel plate) or for Peltier Elements. Magpwr has shown it and I hardly can wait to see it myself. And I have to try air cores with the 2SK170.

It is of course of little use to dimly light a LED, but it is a nice experiment to make a tiny amount of power visible in the form of a glowing LED.

I never observed OU with a Joule Thief type circuit but one can do amazing stuff like High Voltage at high frequency (Slayer Exciter type circuits, which I like to call modern Tesla Coil Towers), or strong 110V or 220V LED lamps (LaserSabers Joule Ringer) and many other variants all showing some interesting features. Just a few components and so many possibilities.

For me the nice thing about the Joule Thief type circuits is that they work, the LED or the lamp lights up, instant gratification. An other "field" that works for me are the pulse motors (also not OU, but they really spin).

@TinselKoala: thank you for mentioning utsource.net, it is always interesting to know where to get stuff.

Greetings, Conrad

Quote from: magpwr on January 02, 2013, 06:49:07 AM
Hi Scratchrobot,

I'm sorry you had to find out the hard way.I did mentioned somehwhere in youtube description or comment to another interested person whom wish to construct the ultra low voltage circuit.
Never to try any Jxxx series all those don't work.I have done stimulation testing using trial software to find out the best component(It's on you to find out.I won't reveal it).
Please don't bother trying other Junction FETs you won't get ultra low self startup voltage.I have wasted 3 days to accidently discover a self starting joule thief at below <30mV while trying to optimise joule Thief.

I bought 2SK170 from  http://www.futurlec.com (http://www.futurlec.com)  (After searching in internet high and low) for $0.20ea (Bought 10 pieces for $2 USD) Since i'm living in Asia shipment is only around $5.50USD for me.Got it within  2 weeks.

Recent discovery my circuit could be powered by solar cell in room merely shine a led torchlight to part of the 1.7v solar cell(400mA) to light a led again from my ultra low joule thief(NOT OVERUNITY).

No problem, i'm used to learn the hard way  ;)
I just ordered some 2SK170 from http://www.reichelt.de/ (http://www.reichelt.de/) thanks Conradelektro for the link.
So I will try your circuit again when they arrive.

Keep up the good work.

Hi,

There is a plan which just appeared in my mind.If anyone make it works then it would be the "1st crystal powered fm radio receiver" ever.

I'd believe almost everyone have heard of "crystal detector" use to receive AM and SW broadcast.

Since 25mV joule thief is technically a voltage booster.

I wonder if we combine geranium crystal with the usual antenna and ground or earth then use that to power 25mV joule thief with 1..4 x 2sk170 to power ready made 1.5V fm radio(digital or analog).
The output speaker might need to be change to piezo speaker to lower 1.5v radio current consumption.
In theory may work.Please take note of output voltage from joule thief may need to use zener 2.4v maybe or small red led to prevent voltage from rising beyond  1.8v.

I know the above suggestion sounds like a crazy idea to achieve a "world 1st batteryless  1.5v fm radio which will have clear reception".

Quote from: magpwr on January 05, 2013, 10:36:23 AM
Hi,

There is a plan which just appeared in my mind.If anyone make it works then it would be the "1st crystal powered fm radio receiver" ever.

I'd believe almost everyone have heard of "crystal detector" use to receive AM and SW broadcast.

Since 25mV joule thief is technically a voltage booster.

I wonder if we combine geranium crystal with the usual antenna and ground or earth then use that to power 25mV joule thief with 1..4 x 2sk170 to power ready made 1.5V fm radio(digital or analog).
The output speaker might need to be change to piezo speaker to lower 1.5v radio current consumption.
In theory may work.Please take note of output voltage from joule thief may need to use zener 2.4v maybe or small red led to prevent voltage from rising beyond  1.8v.

I know the above suggestion sounds like a crazy idea to achieve a "world 1st batteryless  1.5v fm radio which will have clear reception".

Nice idea and maybe it is possible with an antenna and a ground conection?
Would also be nice to just charge a cap with the output of a cristal radio with your circuit and blink a led now and then.

http://solomonsmusic.net/FM_CrystalRadio.html
http://www.somerset.net/arm/fm_only_lowtech.html
http://www.vacuumtubesinc.com/TubeTidbits/AnFMCrystalSet.aspx
http://electronbunker.ca/FMCrystalSet.html
http://www.youtube.com/watch?v=IUaZlCeEMpY


Not FM.... but no external antenna or ground needed:
http://www.arrl.org/files/file/Technology/tis/info/pdf/culter.pdf

Quote from: magpwr on January 05, 2013, 10:36:23 AM
Hi,

There is a plan which just appeared in my mind. If anyone make it works then it would be the "1st crystal powered fm radio receiver" ever.

I'd believe almost everyone have heard of "crystal detector" use to receive AM and SW broadcast.

Since 25mV joule thief is technically a voltage booster.

I wonder if we combine geranium crystal with the usual antenna and ground or earth then use that to power 25mV joule thief with 1..4 x 2sk170 to power ready made 1.5V fm radio(digital or analog).
The output speaker might need to be change to piezo speaker to lower 1.5v radio current consumption.
In theory may work.Please take note of output voltage from joule thief may need to use zener 2.4v maybe or small red led to prevent voltage from rising beyond  1.8v.

I know the above suggestion sounds like a crazy idea to achieve a "world 1st batteryless  1.5v fm radio which will have clear reception".

@magpwr:   At what Voltage does the ZTX1048A transistor self start in your circuit? (You mentioned that you could start it at 60 mV by briefly shorting collector / emitter.)

@crystal power idea:   I did extensive tests and found that the "field strength" of radio, TV, cell phone and WLAN transmissions is way too low to receive any meaningful power. People who lived only a few kilometres away from big AM radio transmitters could receive some power, but most AM stations have shut down in Europe. Also very long antennas and a good earth ground are needed (which is not practical). But I might have done everything wrong and more clever people can find a way.

One can get some power in the µA from the 50 Hz 220 V AC net if one places the "antenna" (just a long cable) very close to appliances which consume power (lamps, refrigerator, washing machines, etc.).

Some people draw mA from the neutral of the 50Hz 220V AC net. But this is cheating. If the 50Hz 220V AC net is not balanced (unequal load on the three phases which are usually distributed) the neutral wanders away from earth ground, and one can draw power from this (between neutral and earth ground).

I see the application of your circuit with Peltier Elements, self made low power batteries and may be even Piezo Crystals.

What I like to dream about is a "neutrino receiver" because the earth is constantly bombarded with neutrinos from the sun. The problem is, neutrinos do not like to interact with any material. But may be some strange material or "field" can be found, that catches neutrinos. But this is just a wild speculation.

Greetings, Conrad

Quote from: TinselKoala on January 05, 2013, 11:36:49 AM
http://solomonsmusic.net/FM_CrystalRadio.html (http://solomonsmusic.net/FM_CrystalRadio.html)
http://www.somerset.net/arm/fm_only_lowtech.html (http://www.somerset.net/arm/fm_only_lowtech.html)
http://www.vacuumtubesinc.com/TubeTidbits/AnFMCrystalSet.aspx (http://www.vacuumtubesinc.com/TubeTidbits/AnFMCrystalSet.aspx)
http://electronbunker.ca/FMCrystalSet.html (http://electronbunker.ca/FMCrystalSet.html)
http://www.youtube.com/watch?v=IUaZlCeEMpY (http://www.youtube.com/watch?v=IUaZlCeEMpY)


Not FM.... but no external antenna or ground needed:
http://www.arrl.org/files/file/Technology/tis/info/pdf/culter.pdf (http://www.arrl.org/files/file/Technology/tis/info/pdf/culter.pdf)

@TinselKoala: very nice circuits. What I always wanted to do is getting away from tuning into a single transmitter or radio station. I want a "broad band receiver" which catches all transmissions from 60 MHz to 10 GHz and then just rectifies all these waves into DC with a full bridge rectifier using HF diodes.

This needs a special "broadband antenna" which catches all electromagnetic waves in a wide spectrum. The problem is the "geometry" of this antenna because different wave lengths need different antenna lengths (for a dipole antenna it is lambda/4).

Greetings, Conrad

Hi Conradelektro and TinselKoala,

To conradelektro,
You asked At what Voltage does the ZTX1048A transistor self start in your circuit?
The answer i only get oscillation to start(temp shorting emitter & collector) 60mV which i was comparing against the old Japanese transistor 2SD1450.I could manually get oscillation to start at 50mV.
But my objective was to see how low voltage a good transistor could continue to oscillate base on low vce(found in datasheet for lower current draw <200mA)
The improvement i noted is oscillation lasted as low as 30mV for ZTX1048A compared to Japanese transistor 2SD1450(Oscillation stop around 43mV) which had slightly higher VCE drop.
If you checked out my youtube channel search "Sanjev21" title would be "Ultra Low voltage 45mV Joule Thief -Potential to be powered by peltier using body heat only" .This was 1st original non self starting version joule thief using the Japanese transistor.I have not posted ZTX version video since by then i had received the 2SK170 to work with.

Please try not to think about "neutrino detector" just typing this word in google yield huge collection of image all of which are built deep underground.I mean to say it's huge.
Seen a documentary about capturing 1 or 2 neutrinos which orignates from the sun years ago just to study it.Can't recall fully.

To TinselKoala,

Thanks  for providing interesting links which makes our eyes open even wider.
But i'm more interested in this particular link you sent.
http://www.arrl.org/files/file/Technology/tis/info/pdf/culter.pdf (http://www.arrl.org/files/file/Technology/tis/info/pdf/culter.pdf)

The author mentioned-
It only takes 5 to 10 mV peak-to-peak on the
source of the MOSFET to provide an adequate
volume level.

There is a another line in pdf where he mentioned combining 4 mosfet to improve performance-
The detector consists of four MOSFETs
in parallel using the quad version of the
detector IC in order to better match the lower
resonator impedance

AL D110800A  (Dual Mosfet version) 
ALD110800APCL(Quad mosfet version)

I will reading up the datasheet to check to current draw capability.Who knows it might be even better than 2SK170 ultra low 25mV.

Lastly,

There is another plan that popped in my mind.How about we use "motor as micro power generator through vibration" as we walk.This idea would be better than peltier since
voltage output could be around 500mV to 1v around 10x times better in terms of voltage output for joule thief.

This is sketchy plan in words as from my mind.Use a small palm size rectangular plastic box as the housing unit for "motor used as micro power generator".
The motor will be the weight on the center and motor is connected to 2 springs(One on the upper portion of the box) and another spring from motor to lower half of box.
The "motor needs to be dual long shaft type" with gears on both end of motor shaft.In the box a line of teeth(From top to bottom) to be lined both side of box(Have not figured out how to balance motor parallely).

If we place the rectangular box vertically in our bag then as we walk the motor would vibrate up and down therefore generating good amount of electricity "on the move".
In theory the power output would be at least 10x time better than peltier design.To power joule thief.

Just a concept.

Quote from: magpwr on January 06, 2013, 07:45:06 AM

There is a another line in pdf where he mentioned combining4 mosfetto improve performance-
The detector consists of four MOSFETs
in parallel using the quad version of the
detector IC in order to better match the lower
resonator impedance

AL D110800A (Dual Mosfet version) 
ALD110800APCL (Quad mosfet version)

@magpwr:  I ordered some Quad ALD110800 and Dual ALD110900 transistors. Mouser rips me off for them (about EUR 3.-- a piece and specially the delivery costs are remarkable), but these transistors just sound right for extreme low power applications. This fascinates me at the moment. Thank you magpwr for sharing your findings.

Interesting is also the use of more than one transistor in parallel. That seems to be less exotic than I thought.

Once I have a very low power Joule Thief working I also have to carefully select the LED. Some LEDs are brighter at low power. Conventional wisdom suggests red LEDs, but the white LEDs have become rather bright at low power. And at the specified power (e.g. 20 mA) they almost hurt the eye.

There is also much work to do with the coils. I guess one wants high grade Ferrite toroids (high induction) and that the oscillation frequency will be low with low power sources. But the 4 : 100 ration suggested by magpwr is a very good starting point.  I tried to measure frequency at my test circuit (2N3904 transistor) at 0.15 Volt, but the DMM influenced the circuit and I did not get a meaningful result. One has to do it "indirectly", but at low power everything one does will influence the circuit (e.g. a measurement winding over the toroid).

I also saw that one can put away the resistor at the base at very low Voltages (e.g. less than 0.2 Volt). That makes the LED brighter. Specially with very low power sources that becomes an option. The transistor will be completely unprotected, therefore one has to be careful that the power source does not increase over a certain threshold (as could happen with a Peltier Element) or it will burn the transistor.

Greetings, Conrad

P-S.: I never thought about a conventional Neutrino tank in my basement, I am looking for a strange field or material that interacts better with neutrinos than the liquid in the big conventional tanks. But it is only a wild thought experiment and not a serious project.

To conradelektro,

It is a good idea to start working on toroid while waiting for parts to be delivered.

Smaller toroid will work just as efficient as well base on my findings.But it's bit more challenging on the mind to wind the toroid.I used a small hard plastic which could pass through the small toroid.
A deep V is cut on plastic on top and bottom to hold the copper wires.The length of plastic is around 1/2 index finger to hold copper wire.Someone in youtube even used ice-cream stick with V cutting to wind toroid.

Another findings related to toroid big or small if your secondary winding exceed the 1/2 way mark\circle\toroid i noticed the efficiency drops.Keep it within the 1/2 as shown in my 25mV Joule Thief video.

The "someone" may have been me, as I posted such a video. I was surprised to learn that people were not already using this method. The length of the spindle or bobbin (i'm not sure what to call it) should be slightly more than one full wrap of the wire around the toroid. Then you can estimate the amount of wire to wrap onto the spindle by the desired turn count on the toroid very easily. Use a material that won't scratch off the enamel insulation; I have used wood and scraps of circuit board material. I've also used aluminum but this can scratch the insulation, if the spindle is wide enough to contact the already-wrapped wire as you push it thru the hole in the toroid.
Another way to do it is to make a splittable circular form that you can assemble onto the toroid through the hole in the toroid. Then you wind your wire from the supply spool onto the circular form,  and then off the circular form onto the toroid.

Update:

I got some 2SK170 transistors and my replication of magpwr's circuit works down to 50 mV. It seems to draw about 0.5 mA at 50 mV.

And it is self starting all the way down to 50 mV. The white LED glows dimmly at 50 mV.

At 100 mV the LED is fairly bright and the current draw seems to be near 1 mA.

The resistor at the Gate is 2.2 KOhm, just one transistor and my coils on a toroid are 10 : 100 at the moment (just to try out some ratios).

I have to take of a few turns to get to 4:100 and will report more tests soon. With four transistors and a ration of 4:100 it should go down to the 25 mV reported by magpwr.

The 50 mV are the lowest supply Voltage I have ever seen for a Joule Thief on my workbench. Thank you magpwr for finding and disclosing the 2SK170.

Greetings, Conrad

I also got my 2SK170 transistors and finally it's working now ;D
The led seems to start glowing around 35mv, will do some more testing with this nice circuit.
Thanks Magpwr and others for helping me getting started with this.

http://www.youtube.com/watch?v=bWqo4akFdVk (http://www.youtube.com/watch?v=bWqo4akFdVk)

Regards,

Hi,

Glad to know the circuit is working for you guys to validate my ultra low voltage joule thief claims.

Just to clarify the "25mV" i meant was the JFET base circuit starts to "self oscillate" it's not about lighting led at 25mV.

It's slighthy higher at around 50mV input for lighting white led.

50 mV is the best I can do. Coil ratio 10:100 seems to be marginally better than ratio 4:100. I see not much difference between using one or two 2SK170 transistors.

I did some tests with home made batteries and with a 50 x 50 mm Peltier Element. See the attached photos and the circuit diagram.

This Joule Thief circuit with the 2SK170 transistor does what I hoped it would do:

1) It works with just one cell of a home made battery. The two batteries I tested today each hold a Voltage of about 150 mV when connected to the circuit and the white LED glows nicely.

2) It also works with just one Peltier Element at a very moderate temperature differential. I had an aluminium plate at both sides of the Peltier Element. The cold side aluminium reached into cold tab water and the hot side aluminium was warmed up by hand. It also works without the cold water reservoir (the cold side at room temperature), but after some minutes the temperature of the cold side will warm up towards the temperature of the hand.

Will try with a different Toroid and with 4 transistors in parallel. In the coming days I also want to do tests with an air coil and some ALD110800 and ALD110900 transistors should arrive soon.

It might be necessary to change the circuit for the ALD110900 or ALD110800 because the Gate has a diode towards the negative rail in order to protect the transistor from a wrong potential between Gate and Source (Gate should not go below Source). But it might be possible to leave the diode pin unconnected (pin V- of the component left open)? We will see whether the component needs a connection to the negative rail? If not, the circuit can stay as it is.

Greetings, Conrad

I used a small 10mm toroid 100:4 and there is not much difference when using 1 or 4 2SK170 transistors but with 4 the led glows a little better. I did a test in the dark and the led starts glowing at around 30ma input from the peltier. I made another video, the led is above the multimeter. http://www.youtube.com/watch?v=ak-aFHztZLg (http://www.youtube.com/watch?v=ak-aFHztZLg)

Quote from: scratchrobot on January 09, 2013, 01:12:56 PM
I used a small 10mm toroid 100:4 and there is not much difference when using 1 or 4 2SK170 transistors but with 4 the led glows a little better. I did a test in the dark and the led starts glowing at around 30ma input from the peltier. I made another video, the led is above the multimeter. http://www.youtube.com/watch?v=ak-aFHztZLg (http://www.youtube.com/watch?v=ak-aFHztZLg)

@scratchrobot: following your idea and video of doing a test in the dark, I see a very faint glow in the white LED at 36 mV (using the Peltier Element as shown in the photo in may last post). Because my DMM is not a top quality instrument, let's say it happens at 40 mV. (Circuit shown in my last post.)

I am impressed by the 2SK170 transistor, it seems to allow oscillation at very low Voltages (which I even can not measure with my instruments). It already oscillates well below the Voltage at which the LED starts to glow.

I am very happy that all seems to work very well at 100 mV (LED sufficiently bright to impress onlookers), which is good enough for most home made very simple battery cells and for simple Peltier experiments with body warmth and ambient temperature.

I plan a "coffee or tea cup temperature indicator" on top of which one places a hot or warm cup or mug and the gadget inicates whether it is till warmer than room temperature by the glowing LED. Not realy a world saving invention, but it could serve as a novelty item for science geeks.

Tomorrow I will test an earth battery  in the garden (copper and aluminium tube driven into the ground in close proximity).


@magpqr: thank you for clearyfying the 25 mV result, I thought I did something wrong because my white LED only started to shine at 50 mV. With a good Voltmeter (very high impedance) one could may be see oscillation with a 2SK170 at even lower Voltages than 25 mV.

Greetings, Conrad

 a "coffee or tea cup temperature indicator"

Nice idea!

Hi,

I figured out in order to find out at which low voltage the JFET base joule thief starts oscillation if we don't have an oscilloscope.We could use piezo speaker only provided the frequency is running at <20KHZ for sound to be heard.
It doesn't matter if we don't hear anything since we could use Led as simple indicator.

There is another plan which i have in my mind which i'm planning on doing to increase output power from 25mV joule thief with <4mA current draw.
Please look at the youtube videos by searching for the 2 youtube user "name" indicated below first(Please jump to last paragraph at the bottom then come back here) .To understand i'm really talking about.
The plan is to connect the copper pipe with toroid like shown in video.One end of copper pipe with 6 toroid around it goes to the JFET Drain and other end of copper pipe goes to primary coil.

The reason why i'm planning this is because if you think "copper pipe" by itself which means impedance is very very low likely below 0.01 Ohms for a short copper pipe.right...
The Larskro circuit with copper pipe\toroid combination as shown below somehow reveal that with even such a very low Ohms between copper pipe not to mention the micro voltage if any exist on copper pipe during operation is very very tiny.Yet the output from inside the copper pipe could light Leds brightly.
My observation of those video as shown below from my perspective the circuit using 1.5volt merely serve as "oscillator" for the copper pipe/toroid combination.
Since 25mV JFET base circuit\Joule Thief is also a Oscillator/ultra low voltage version . If we connect to (one end of copper pipe between primary coil and JFET drain).This pipe with toroid around it in theory would "resonate" as well and in theory produce more power to light LED or many leds .Maybe we could even stepped up voltage from there from the pipe,but not figured out yet if need too.

Hence no power need to be tapped from the Gate of JFET like in current circuit example to power Led since this time more power is produce from within the copper pipe/toroid to light more leds (theory).





I have received the toroids but i need to find time over weekend to go to hardware shop to get  2 type of smaller copper pipe to fit 2 version of toroid shown in video link below.In theory this may be even more efficient.

Youtube user "Larskro" Title "Resonant LCR Amazing Circuit 2" or same tube user- Title "Amazing Circuit duplication of Larskro YT-video"
Youtube "capindres" "Double Amazing Joule Thief video" (A big improvement over Larskro's version 100++ really bright Led on 1xAA)

Those toroids base on the 3 videos are purchased from here.
http://www.cwsbytemark.com  (http://www.cwsbytemark.com)

@magpwr: please look at the attached circuit. Is this what you imagine?

Greetings, Conrad

Hi Conradelektro,


It should still be based on the original circuit design for 25mV joule thief.The one you have posted recently in this forum.But this  time connection will  be "JFET (drain\collector) to (one end of copper pipe ,then another end of copper pipe)  to toroid 4 turns " in series.

Original connection is JFET Drain\collector directly connected to toroid 4 turns(Which is already less than 1 Ohms).

The (100:4)toroid is still needed to serve as oscillator to resonate copper pipe with 6 toroid around it.

Quote from: magpwr on January 10, 2013, 11:28:38 AM
Hi Conradelektro,

It should still be based on the original circuit design for 25mV joule thief.The one you have posted recently in this forum.But this  time connection will  be "JFET (drain\collector) to (one end of copper pipe ,then another end of copper pipe)  to toroid 4 turns " in series.

Original connection is JFET Drain\collector directly connected to toroid 4 turns(Which is already less than 1 Ohms).

The (100:4)toroid is still needed to serve as oscillator to resonate copper pipe with 6 toroid around it.

@magpwr: in case the toroid should stay in the circuit as you describe, I would put the usual bifilar 10 turns on the toroid, because the 100 turns for the Gate are only needed to lift the Voltage above the 2.2 Volt level for the diode. Since the diode will be fed via the copper tube, this Voltage lift at the Gate of the transistor is not needed any more.

I am always a bit reluctant to let the Gate of a transistor go much below the Source (or the Base below the Emmiter in an NPN transistor) because this can cause damage (at least conventional transitor wisdom claims that).

Of course, only a test can give clarification, this is just theoretical waffling.

Greetings, Conrad

Hi Conradelektro,

I have to agree with you as this is just theory.


The original circuit version by larskro may not be efficient enough which you have embedded into the circuit diagram to show me.Just realised it hours later after i woke up.
Since it's looks like shorting the transistor  by directly connecting to pipe from transistor collector.There will be a usual voltage drop maybe 0.9v across BDX33 and current would be very high could drain 1xAA very easily.

The circuit i'm referring to is a better version where copper pipe is connected in series with coil.Just refer to the circuit diagram shown below.

A another version from the original larkro circuit using copper pipe.
https://www.youtube.com/watch?v=WlVkfaFiKiE (http://www.youtube.com/watch?v=WlVkfaFiKiE)

That's the reason why i plan to connect the 25mv Joule thief toroid 4 turns in series with copper pipe.In theory may work ,maybe.
Base on the above youtube i think there will be higher voltage drop across coil since there is more turns in the coil base on youtube link.We need to be aware copper pipe resistance is very very low <0.05Ohms.And yet this above video work still amaze me.

@magpwr: the video you mentioned (https://www.youtube.com/watch?v=WlVkfaFiKiE (http://www.youtube.com/watch?v=WlVkfaFiKiE)) is very interesting.

I attach a drawing that depicts how I understand your idea (toroid + copper pipe with many toroids).

Greetings, Conrad

Hi conradelektro,

Yes that is the plan in my mind.Thanks for drawing out the circuit diagram.

May i ask what software you use.

@magpwr: I just used Windows Paint (Windows 7).

Windows Paint is good enough for simple diagrams and if you have a Windows System you have Paint for free.

For more complicated drawing jobs I use TurboCAD Delux V12 (rather old version). Today they sell Version 19 http://www.turbocad.com/TurboCAD/TurboCADWindows/TurboCADDeluxe19/tabid/1869/Default.aspx (http://www.turbocad.com/TurboCAD/TurboCADWindows/TurboCADDeluxe19/tabid/1869/Default.aspx).

Many people use circuit design programms, some of them are free of charge. But they are usually too complicated for simple circuits.

One could use Gimp on a Linux System. I also have Gimp for Windows http://www.gimp.org/downloads/ (http://www.gimp.org/downloads/) and sometimes I use it too.


Tomorrow I will look through my copper pipes and toroids to test your idea. From the video of Eva Aboriđanin (http://www.youtube.com/user/Zdravko1394?feature=watch) I conclude that it migh just work.

Greetings, Conrad

Update "many toroids over copper pipe" and transistor ALD110800 or ALD110900:

1) Many Toroids Over Copper Pipe: Total Failure!

I fabricated such a strange "many toroids over copper pipe" (see attached photo) and put it into magpwr's circuit as indicated in the schematics published with my Reply #56 in this thread (my second last post). If I leave the LED in its original place one sees that the circuit oscillates and the LED glows. But I can not get anything out of the "many toroids over copper pipe". May be it only works with high power (e.g. 500 mA flowing through).

But, please, I do not want to discourage anybody. It is important that others try as well.

2) Transistor ALD110800 or ALD110900: These transistors are not a one to one replacement of the 2SK170. So, do not buy them yet.

I put an ALD110900 in magpwr's circuit instead of two 2SK170 transistors. The circuit starts to oscillate (LED glowing) at about 0.4 Volt but stops after a few seconds. Changing the resistor in front of the gate does not help, it does the same without any resistor.

I have to do more tests and experiments. My impression is that the component (the ALD110900) needs a connection to the negative rail which implies that the LED has to go elsewhere. In case the V- pin is not connected some strange charge seems to build up inside the component. In other words, the Gate needs the internal diode connection to the negative rail in order to function properly.

There might be some other problems, I am not really an expert. But I will play with it some more. It shows once more how lucky we are that magpwr identified the 2SK170.

Greetings, Conrad

Hi magpwr,

Your idea on the copper tube and toroid cores around it reminded me of mainly Russian youtube videos where such transformer assembly was shown BUT there were a multiturn coil around the outer surface of the toroidal cores (because it has a normal cylindrical surface, such as Conrad shows above in his photo too). 
So I think the 100 turn may go onto the outer surface of the toroidal cores put close to each other, the 4 or 5 turn feedback coil would come on top of the 100 turns and the output may come from the thick piece of wire put through the copper tube as shown in this video:
http://www.youtube.com/watch?v=eH2TWPJiwEA (http://www.youtube.com/watch?v=eH2TWPJiwEA) 

I hope you understand how I mean the 100 turn coil, and I do not mean any extra output from such a transformer could be gained, I just trying to show a variant with the copper tube + toroid core type transformer.  By the way no any extra output is suggested from such "magic" transformer on the Russian sites as far I know.

rgds, Gyula

PS Here is a short video where a coil is shown as wound onto the outside cylindrical surface of a longer toroidal core, see from 0:17 to 0:23 :  http://www.youtube.com/watch?v=2qtVR1LkLSc (http://www.youtube.com/watch?v=2qtVR1LkLSc)

Hi gyulasum and conradelektro,

From the russian demo by looking dates of posting youtube.It looks like original design came from Russia.
But video was not easy to understand,but some in youtubers managed to figured how it works and showed a better demo which is easy to replicate.

I'ts very interesting seeing a bulb lighting from a drill bit or copper pipe or screwdriver version from youtube user "Larskro".
Knowing all this metal got very very low resistance.How did bulb light from a straight metal still pretty much puzzles me with such a low current.

Since i was using the "Larskro's approach" and observed the "double amazing joule thief" video from a other youtubers.
I understand oscillator is needed to "resonate" the copper pipe(very low resistance,low voltage would definately be measured across pipes).Unfortunately i have not got the time to head down to hardware store to get myself copper pipe which toroids could be fittted.

Once i got copper pipe i will start with scaled up version(1.5V) then scale down from there to <100mV .Then figure out if need to stepped voltage from copper wire inside pipe since i believe the output is likely not rectified or DC output.

To conradelektro,
Please try higher voltage first 1.5v.Please take note the if your running something inside pipe may need to "insulate straight wire inside,unless it's insulated 14 or 18AWG copper wire" to prevent it from touching the inner walls of pipe like shown in double amazing video.

Quote from: magpwr on January 11, 2013, 11:22:11 PM
To conradelektro,
Please try higher voltage first 1.5v.Please take note the if your running something inside pipe may need to "insulate straight wire inside,unless it's insulated 14 or 18AWG copper wire" to prevent it from touching the inner walls of pipe like shown in double amazing video.

@magpwr: I tried the insulated wire through the copper pipe and Voltages up to 1.5 Volt.

@gyulasum: these strange transformers with toroids over a copper pipe are an interesting thing to explore, but I think they are not suitable for very low Voltage and low power. In the videos which magpwr cites there seem to be quite high currents (several hundred mA) flowing through the circuit. That is why the presenters in the Videos change the 1.5 Volt battery or 1.2 Volt rechargeable battery to have a fresh one.

Transistors ALD110800 and ALD110900:

With an ordinary Joule Thief circuit (bifilar windings over a toroid) I get oscillations down to 0,1 Volt and 10 µA. This is the lowest supply power I have ever seen in a JT (selfstarting). But I can not do much with these oscillations, wherever I place the LED, it will not glow. I will play a bit more with the ALD110800 and ALD110900 transistors.

It seems the ALD110800 and ALD110900 transistors need a connection to the negative rail (pin V-), then they behave well. The very low power consumption seems to stem from the "Drain Source On Resistance" of 500 Ohm. This opens interesting possibilities (one has great freedom with the coil at the Drain without causing too much current through Drain / Source) but also limitations. One also does not need any resistor at the Gate. May be that is bad for making LEDs glow, but for building a very low Voltage and very low power oscillator, this helps a lot.

Greetings, Conrad

Some success with the ALD110900 transistor in a Slayer Exciter type circuit:

The ALD110900 just wants to oscillate, no matter how badly matched the coils are.

Look at http://www.youtube.com/user/GBluer  (http://www.youtube.com/user/GBluer),  GBluer shows in his videos many different Slayer Exciter type circuits and many different coils. I used the toroid I had from the 2SK170 experiments ( 4:100 coils) which is a very bad choice, and still it works fairly well.

See the attached circuit (hardly any components) and the photos. I will wind a tuneable coil (with a Ferrite rod that slides in and out), which should be more suitable for this type of circuit. The toroid dampens the oscillations down to about 100 KHz (due to its high inductance), a Slayer Exciter wants MHz.

The 2SK170 is still better for low Voltage (LED glows at 50 mV), but the 200 mV and 200µA I achieve with the ALD110900 are pretty good (given that it needs no additional components besides the coils, because the 100 µF capacitor is optional).

The ALD110900 pin V- is connected to the negative rail. Note the internal diode from the Gate to the negative rail. In a Slayer Exciter there always has to be "something" (diode, capacitor, inductance) between Gate and negative rail.

Greetings, Conrad

Hi Conrad,


Data sheet for your ALD MOSFETs says:

"For most applications, connect the V+ pin to the most positive voltage and the V- pins to the most negative voltage in the
system. All other pins must have voltages within these voltage limits at all times."   

I know you mentioned you connected V- pin to the Source which is at the most negative voltage in your schematic and maybe it would be good to connect V+ pin to the positive pole of the supply voltage. Of course this latter is valid for quad device ALD110800 (its V+  is Pin 13) and there is no such V+  pin for the ALD110900.
One more thing: you surely noticed the pins with asterisk symbol as (IC*) and these pins are also recommended to connect to V-  i.e. to the most negative voltage rail.  Such pins for the quad device ALD110800 are Pin 1, 8, 9 and 16 and for the dual device ALD110900 these are Pin 1 and 8.

You mentioned the Drain Source On resistance is 500 Ohm, however this is valid whenever the Gate Source voltage is +4V (at Vds=0.1V). And how well this condition is fulfilled in your shown oscillator circuit, an oscilloscope could be used to see the Gate source AC voltage waves with respect to the Source pin as being the most negative point.
OF course if you connect the dual device in parallel, then you would have about 250 Ohm RdsON and for the quad device this would be about 125 Ohm.

You also mentioned the internal diode between the Gate and the V- pins and it occured to me that while current can only circulate via this internal diode and via the 100 turns coil and via the LED when the induced AC voltage polarity across the 100 turn coil corresponds to the forward direction of both the LED and the internal diode (this is why the LED direction is important) so IN CASE the internal diode is a normal Si diode (albeit a very fast one) the forward voltage drop across it may be at least 0.6V and this lowers the useful voltage that may feed the LED from the total induced AC voltage peak so I wonder if connecting a Germanium or Schottky diode across the Gate Source pins in parallel with the internal diode (anode to anode and cathode to cathode) you may be able to gain some more brightness from the LED. Maybe this extra diode influences oscillation too much and you get just a dimmer LED, well I hope not... and if you agree and feel like testing it, please do. Even another Si diode like 1N914 or 1N4148 might be worth trying if you do not have any Germanium or Schottky diode at hand.

You have very nice results!

rgds,  Gyula

Quote from: gyulasun on January 12, 2013, 02:55:48 PM
Hi Conrad,


Data sheet for your ALD MOSFETs says:

One more thing: you surely noticed the pins with asterisk symbol as (IC*) and these pins are also recommended to connect to V-  i.e. to the most negative voltage rail.  Such pins for the quad device ALD110800 are Pin 1, 8, 9 and 16 and for the dual device ALD110900 these are Pin 1 and 8.

You mentioned the Drain Source On resistance is 500 Ohm, however this is valid whenever the Gate Source voltage is +4V (at Vds=0.1V). And how well this condition is fulfilled in your shown oscillator circuit, an oscilloscope could be used to see the Gate source AC voltage waves with respect to the Source pin as being the most negative point.
OF course if you connect the dual device in parallel, then you would have about 250 Ohm RdsON and for the quad device this would be about 125 Ohm.

You also mentioned the internal diode between the Gate and the V- pins and it occured to me that while current can only circulate via this internal diode and via the 100 turns coil and via the LED when the induced AC voltage polarity across the 100 turn coil corresponds to the forward direction of both the LED and the internal diode (this is why the LED direction is important) so IN CASE the internal diode is a normal Si diode (albeit a very fast one) the forward voltage drop across it may be at least 0.6V and this lowers the useful voltage that may feed the LED from the total induced AC voltage peak so I wonder if connecting a Germanium or Schottky diode across the Gate Source pins in parallel with the internal diode (anode to anode and cathode to cathode) you may be able to gain some more brightness from the LED. Maybe this extra diode influences oscillation too much and you get just a dimmer LED, well I hope not... and if you agree and feel like testing it, please do. Even another Si diode like 1N914 or 1N4148 might be worth trying if you do not have any Germanium or Schottky diode at hand.

rgds,  Gyula

@Gyula: thank you for the helpful advice.

I connected a 1N5711 diode (forward Voltage 0.38 V)  from Gate to negative rail (in parallel to the internal diode) and it made the LED a little brighter at very low Voltages. I now see the LED glowing at 180 mV (before it was 200 mV). Starting at 250 mV the difference in brightness is clearly visible.

Connecting the IC* pins to negative rail did not change anything. I guess because they are internally connected to the V- pin which I already had connected to negative rail, it might reduce noise.

I use the two transistors in the in ALD110900 in parallel and I will try all four transistors in a ALD110800 in parallel. I agree, the V+ pin will be important in a ALD110800. I think V+ and V- get rid of stray charges in the component.

Just now I am preparing the winding of a 500 turn Slayer Tower. The winding is about 120 mm high, 30 mm diameter, plastic tube, 0.22 mm diameter enamelled wire. The primary will have 12 turns because that worked well with Slayer Towers I tested some years ago.

I am a big fan of the Slayer Exciter because it reminds me of a Tesla Coil. For me the Slayer Exciter is a Tesla Coil built with modern components (transistor instead of a spark gap).

At 1 V supply voltage a LED could be lit from the top wire of the 100 turns coil with an Avramenko plug (without connecting it to negative rail). So, it is a genuine Slayer Exciter.

Greetings, Conrad

Hi Conradelektro,

I just placed order on ALD110800 last night.I could only advise any finding once i received those.

My plan  would be using 3 tap/output from toroid somewhat similar connection to get negative for mosfet  http://www.arrl.org/files/file/Technology/tis/info/pdf/culter.pdf  (http://www.arrl.org/files/file/Technology/tis/info/pdf/culter.pdf)
I would like to imagine that as a unique osillator or s/w radio base on above link using no (antenna or ground or battery).

Honestly 500ohms from single mosfet doesn't sound good even combining 4 mosfet to lower on resistance(RDS) to 125Ohms at around 5v.
The reason being 4 turns or 10 turns on primary winding of toroid would have resistance below "1" Ohms .
This itself create a huge instances of Impedance mismatch  between mosfet and coil which would lower efficiency further.

The only other way i know to increase Joule thief "current output" at around 50mV input is to connect good npn transistor ZTX1048A(22Ohms ... 100Ohms to base from secondary coil) in parallel  with one 2SK170(1.8k to gate from secondary like the original layout in 25mV joule  thief).Hint "The JFET will kick start npn transistor through emitter & collector once input voltage reach around 30mV(lowest possible osc to sustain for npn transistor ZTX1048A)    ...40mV"
This is tested to be working great at higher current draw depending on resistor value on npn transistor base which you need to tally with "source of power/current capability at 50mV".
These circuit is my version 2.0 using 50farad ultra cap quick charged in 1...3sec to get around 50mV which i have not posted yet.The reason why i have not posted ver 2.0 is because peltier measured output using body heat is only 4.2mA.

I will update finding about copper pipes/tubes with torrid once i purchased small copper pipes normally used in Air-con.

info only related to diodes-
1N5817 at 1A is 450mV  (Datasheet)     (Using multimeter to test forward voltage on my available Schottky diode 1N5817 is 0.145v)
1N5711 at 0.6mA is at 0.2volts (Looking at graph in datasheet) .I do not have this diode to test on multimeter  mode-Diode Tester

Some more tests with the Slayer Exciter using ALD110900 transistor:

The new coil (see photo) was not better than the toroid from my Reply #63. The Ferrite Core was necessary, without it at least a 1 Volt supply Voltage was needed. I tried many coil ratios from 8:500 to 25:500, 13:500 was best.

The lower limit to make the LED glow was still 180 mV and I have the impression that the LED was a little brighter with the Toroid.

I will change the ratio of windings on the toroid from 4:100 to 4:200 with a few taps to see if that leads further.

So, the 2SK170 is still the best for low Voltages, do not buy the ALDs yet.

I will try the Slayer Exciter type circuit with the 2SK170 and other Joule Thief variants with the ALD110900 and ALD110800.

Greetings, Conrad

Some more tests with the ALD110900 transistor:

See the attached drawing which shows all the tets I did with the ALD110900 transistor (2 transistors in parallel). Only the "Slayer Exciter" variant comes close to the 2SK170 transistor (180 mV), all other circuits need at least 700 mV. The toroid was better than an air coil.

One can make a JT with the ALD110900, two windings on a toroid and a LED, the electrolytic capacitor and the 1N5711 diode being optional. And it uses very little current, namely about 1.2 mA at 1 Volt or 1.5 Volt, although the LED is fairly bright at this level.

I now leave the ALDs be till I get new ideas and will concentrate on the 2SK170 transistor which could make a LED glow at 50 mV with magpwr's circuit.

What are the essential characteristics of the 2SK170? I want to find replacements, because it seems odd that only the 2SK170 can make a LED glow at 50 mV?

Greetings, Conrad

hi conradelektro,

This is a huge list of N-channel Junction FET which will self osc\start at low voltage.But 2SK170 is the best over all.
I have already taken the trouble find out the best self starting component.Please note those i put those *high current out* are the better ones notably 2SK..series as listed below.Some 2sk jfet don't work so i exclude those and rest of unwantd JFETS already.

Please look no further unless it some JFET which is released recently,unknown at this moment.

N channel JFET which could self start at low voltage <100mV-----------

2N4861A      *35mv startup
2N4393   *    *60mv startup
2N3972           *33mv startup    ***********Could self start at 22mV (or 1mV lower than 2SK170)but much lower current output capability.
2SK117              *31mv startup -high current output
2SK170      *30mv startup -high current output ******final testing reveal this is the best self starting at 23mV.
2SK364               *31mv startup -high current output
2SK366               *31mv startup -high current output

2N4093       *35mv startup
2N4223               *35mv startup
2N5484               *60mV startup
2N5103               *60mv startup

BF510                 *45mv startup
BF410B               *60mv

BC264C               *60mV
BC264A               *60mV
BC264D               *60mV

@magpwr: thank you, great list, that helps a lot.

I guess you wrote "mV" where it should be "V", see this list which I will keep for reference:

N channel JFET which could self start at low voltage <100mV (according to magpwr):

2N4861A   0.35 V     start up
2N4393     0.060 V   start up  with BD329
2N3972     0.033 V   start up    *********** Could self start at 22mV (or 1mV lower than 2SK170)but much lower current output capability
2SK117     0.031 V  start up - high current output

2SK170      0.030 V  start up - high current output  ******final testing reveal this is the best self starting at 23mV

2SK364      0.031 V  start up - high current output
2SK366      0.031 V  start up - high current output

2N4093       0.35 V    start up
2N4223       0.35 V    start up
2N5484       0.060 V  start up
2N5103       0.060 V  start up

BF510          0.045 V  start up
BF410B        0.060 V

BC264C        0.060 V
BC264A        0.060 V
BC264D        0.060 V

Magpwr please verify, this list is very helpful for everybody playing with Joule Thief type circuits. I will put it into a spread sheet once verified by you.

I do not understand "2N4393     0.060 V   start up with BD329" (how is it started with the BD329)?

Mouser sells the 2SK3666-2-TB-E and the STD35N3LH5. Are they replacements for the 2SK364 and 2SK366?
Mouser also sells 512-MMBF5103, a replacement of BF510?
Mouser sells the 771-BF510215 , a replacement of BF510?

The 2SK170 is fairly easy to get, all the other are not widely sold.

Sorry to ask again about the ZTX1048: I understood that one could start the ZTX1048 with a 2SK170, but at what Voltage would the ZTX1048 start alone (just the ZTX1048), at 40 mV ?

Greetings, Conrad

Some results with 4 transistors 2SK170 in parallel in magpwr's circuit: 2.2 K resistor in front of the Gate (magpwr specified 1.8 K), 2N5711 diode in series with a white LED.

The white Led starts to glow at 37 mV and is clearly visible at 50 mV.

Power usage is rather high, that is why the LED glows stronger with 4 transistors in parallel:

2.2 mA at 50 mV
6 mA at 100 mV
14 ma at 200 mV (LED is very bright at 200 mV)
34 mA at 500 mV (this is too high, needs a 10 K resistor at the Gate to reduce current to 20 mA)
(do not try with 1 Volt, could destroy the transistors)

It works fine with a 50 mm x 50 mm Peltier Element: cold side at room temperature (about 22 degrees Celsius) and the hot side touched with the hand. Again I see that the LED starts to glow around 37 mV and is nicely visible at 50 mV which is reached within seconds when touching the Peltier Elemet with the hand.

I could also make a LED shine with 4 transistors 2SK170 in parallel and the Slayer Exciter circuit I showed in this thread. But it needs at least 100 mV to make the LED glow, so magpwr's circuit is better.

In case one wants to use 1.5 Volt batteries (till they are deeply discharged down to 200 mV) one could use the ALD110900 transistor in the Slayer Exciter circuit I showed earlier in this thread because it does not need any additional components (just the toroid with the windings and the white LED) and uses very little power (1 mA at 1 V).

But for a really low Voltage power supply (50 mV to 200 mV) magpwr's circuit with four 2SK170 transistors in parallel is great. It needs a few mA to work, about 100 µW at 50 mV, and it is astonishing  that one can make a LED glow with such little power and such little Voltage. With one 2SK170 transistor the demand is only 500µA in order to see the LED dimly glow at 50 mV.

Greetings, Conrad

hi Conradelektro,


I have admended the voltage in earlier posting(hours ago) eg:0.035mV changed to 35mV.


The one which i combine JFET with BD329 was at early stages of testing.Could ignore those old value which i forgot to remove as ZTX1048A combine with 2SK170 in parallel configuration using 1 resistor 2.2kOhms from secondary coil still produce better result at higher current draw.This could be the version 2.0 circuit maybe not suitable for peltier.I  noticed  peltier raw output is around 10mA for me with 6c temperature difference.


ZTX1048A(4amps direct,20A pulsing for one tiny transistor with no heatsink ,refer to datasheet) could self start at usual around 0.6volts like all silicon transistor.But voltage drop across emitter and collector at below 200mA is around 50mV base on datasheet ,VCE drop across common average transistor is around 0.3...0.5v .
ZTX1047A should be slighter better but i can't find those online.

Hi Folks,

I have made some comparison on JFETs as per electrical data and I think the BF861B from NXP (earlier Philips type) may be worth a test because most of its features are the same or  a bit better than that of the 2SK170. Here is a data sheet for it:
http://www.nxp.com/documents/data_sheet/BF861A_BF861B_BF861C.pdf (http://www.nxp.com/documents/data_sheet/BF861A_BF861B_BF861C.pdf)  and Farnell has it http://uk.farnell.com/nxp/bf861b-215/jet-n-ch-25v-sot-23/dp/1758057 (http://uk.farnell.com/nxp/bf861b-215/jet-n-ch-25v-sot-23/dp/1758057) 
The highest transconductance of this family is BF861C but it should be looked for where it is avalable  EDIT RS Components seem to have it http://uk.rs-online.com/web/c/?searchTerm=bf861&sra=oss&r=t  a package of 10 should be ordered...

I do not know this seller but they seem to sell ZTX1047A: http://elcodis.com/searchparts/ztx1047.html (http://elcodis.com/searchparts/ztx1047.html)   Farnell and others has the ZTX1048A.

rgds,  Gyula

@gyulasun: thank you for searching the BF861B, since it is not expensive I might order it.
@magpwr: thank you for updating the transistor list.

I looked for the transistors named by magpwr and found the following sources:


2SK364 (31mv start up -high current output)  511-STD35N3LH5  Mouser (might not be the same, replacement type)

2SK366 (31mv start up -high current output)  863-2SK3666-2-TB-E  Mouser

BF510   (45mv start up)  512-MMBF5103  or  771-BF510215  Mouser

2SK170 (30 mV start up - high current output, best so far)  Mouser,  Reichelt (Germany),  eBay seller diy-audio4you (Germany)

ZTX1048A (600 mV startup, combine with 2SK170)  522-ZTX1048A  Mouser,  Farnell

BF861B (not tested, Gyulasun)  Mouser,  Farnell


I attached a photo of my Peltier Element Test (4 transistors 2SK170 in parallel, steel cooking pan at room temperature about 22° Celsius, Peltier Element 50 mm x 50 mm).

Greetings, Conrad

regarding light power

..... the led has 10 times the efficiency of the candle ....

http://www.youtube.com/watch?v=0_GqE5aDR68 (http://www.youtube.com/watch?v=0_GqE5aDR68)


http://www.buytpod.com/


http://www.tellurex.com/products/power.php

Quote from: wings on January 15, 2013, 05:56:21 AM
regarding light power

..... the led has 10 times the efficiency of the candle ....

http://www.youtube.com/watch?v=0_GqE5aDR68 (http://www.youtube.com/watch?v=0_GqE5aDR68)


http://www.buytpod.com/ (http://www.buytpod.com/)


http://www.tellurex.com/products/power.php (http://www.tellurex.com/products/power.php)

I never test it but

... fire emit ions ....

http://www.youtube.com/watch?feature=player_embedded&v=Jb2R3Hq88Go (http://www.youtube.com/watch?feature=player_embedded&v=Jb2R3Hq88Go)

http://www.astarmathsandphysics.com/o_level_physics_notes/o_level_physics_notes_ions_in_a_flame.html (http://www.astarmathsandphysics.com/o_level_physics_notes/o_level_physics_notes_ions_in_a_flame.html)

similar effect

http://www.youtube.com/watch?feature=player_detailpage&v=WNl0sFViaxs (http://www.youtube.com/watch?feature=player_detailpage&v=WNl0sFViaxs)

http://www.sparkbangbuzz.com/flame-amp/flameamp.htm (http://www.sparkbangbuzz.com/flame-amp/flameamp.htm)


interesting "cristal" radio with a candle:


http://oddmix.com/tech/cr_flame_detector_radio.html

@wings: It is of course well known to use Peltier Elements as a power source (in combination with fire).

I did some experiments, see http://www.overunity.com/12607/thermoelectric-generator-candle-and-cold-water/#.UPU9BGckQfw .

Usually one needs a big temperature difference (e.g. 100° Celsius, better more) between hot and cold side to draw some useful energy from a Peltier Element.

The new (and for me interesting) feature of magpwr's circuit (and the discovery of the 2SK170 transistor) is the fact, that very small temperature differences (e.g. 10° Celsius) can make a LED glow with a Peltier Element.


Eventually it could be interesting to charge a small 1.2 Volt accumulator with a small temperature difference (without fire, just natural temperature differences e.g. between water and air, or air and ground). Not a world saving thing, but nice.

And in itself, it is interesting, that one can make a LED glow starting with about 50 mV. The big event (for me) is the discovery of the 2SK170 for the Joule Thief type circuits, bringing the Voltage down to 50 mV (which I never saw before magpwr's disclosure).

Greetings, Conrad

Hi Wings ,conradelektro,

Thanks for showing me the thin film peltier  i have done research on this thinfilm version around 1 or 2 years but didn't know where to buy those.Maybe at that time it was just at discovery stage.

Thin-film peltiers
http://www.nextreme.com/ (http://www.nextreme.com/)

I will not use buy those thin film peltier yet unless i'm focus on making circuit and everything else tiny. 

My next investigation  might or might not be "overunity" which remain to be seen/tested yet.This is what i have in mind base on an old discovery which i accidentally discovered around 4 or 5 years ago.

This is my long story short how i accidentally discoverd something of interest-
You see i love upgrading pc along with graphics card once every 2 years since i love playing games for max 2hours to keep my mind\cpu working at top speed.

I wanted to try a different kind of cooler available in market at that time.The cpu cooler i purchased uses a liquid/gas to cool the CPU. The gas used is Tetrafluoroethane-1,1,1,2 (R134A).
Please check the link to see how the cooling effect works.
http://www.bjorn3d.com/2005/09/asetek-vapochill-micro-ultra-low-noise-cooler/#.UPVnXGehfng (http://www.bjorn3d.com/2005/09/asetek-vapochill-micro-ultra-low-noise-cooler/#.UPVnXGehfng)

This is my discovery i took the  "VapoChill Micro" by hand of course.Since our hand tend to be shaky a little as long as we move.
This is what i noticed by merely moving the "VapoChill Micro" a little i noticed the base of cpu cooler getting really cold and also noticed "water condensation taking place quickly at the base".

Please take note i'm living in a tropical country where the temperature remains at around 30 degree celsius throughout the year ,+- 3c difference max.
I did not have those remote temperature monitor around 5 years back.But my guess is base on speed the water condensation taking place in matter of few seconds on gas based cpu cooler might be  around 10 or 15 degree celsius cooler than my room temperature.

That's all for my discovery part.
-------------

I have an idea which might generate "overunity" if it works.If we were to place peltier beneath the gas base cpu cooler(cold) and heatsink is place one side peltier to obtain hot temperature from my room(hot).

Place the (cpu cooler with peltier with heatsink under the peltier ) onto something that it will rock this unit to and fro or left to right easily.Since we now know that this gas base cooler is  really good at "losing heat" for whatever reason.
I'd believe we some of us have seen commercial unit sold as decor or entertainment visual unit  where it just move left to right maybe with 1 battery or solar.I have no name for it at the moment.

Since we know peltier could definately generate electricity by temperature difference generated especially from rocking motion(gas cpu cooler) against my room temperature 30c.
We take that electricity from peltier to power the rocking motion coil via the 25mV joule thief.As the unit moves left to right for example gas cpu cooler gets even cooler.

This is just theory to generate electricity on infinite basis with no further human  intervention.It really depend on peltier efficiency among other thing eg:design.

@magpwr and all OU by cooling fans:

Look at http://en.wikipedia.org/wiki/Heat_pipe (http://en.wikipedia.org/wiki/Heat_pipe) , it describes the principle behind "heat pipe cooling".

About the shaking: according to my understanding, the "shaking" of a "heat pipe cooler" is a one time option. The droplets which have formed at the top of the heat pipes are shaken off and drop to the bottom, where they evaporate and cause noticeable cooling. But once all droplets have fallen, it will take a long time till they form again (at least minutes). And it does not matter whether all the droplets fall at once (by shaking) or fall off one by one (cause by gravity once they become too big for adhesion) over a longer period.

So, let's say the air flow through the cooler is such that the droplets form within 1 minute. Now, we can shake off most of the dropplets at the end of each minute by hand. Or, we can let the thing be and the droplets fall slowly over the whole minute. In both cases the dissipated heat will be about the same.

I see no OU effect. The dissipated heat depends on the air flow (e.g. Litre per Minute) through the heat sink (laminates) and on the temperature and moisture of that air.

The advantage of a heat pipe cooler (pipes in a laminate, without compressor) is that it can be smaller than a conventional heat sink (laminates, without the heat pipes). This is practical, but not OU.

Let's say we move 100 Litre of air through a conventional cooler per minute. We have to move the same 100 Litre of air through the heat pipe cooler per minute, but the air will have to move through a smaller space, which means we need a bit more fan energy, because the air has to get a higher speed (in order to press the same 100 Litre per minute through a smaller hole). So, the heat pipe cooler even needs a little bit more energy for the fan than a conventional cooler. But roughly speaking, they need the same fan energy (at a given air temperature and moisture, and cooling down to the same temperature)

The next OU hurdle is the efficiency of Peltier Elements. The best Peltier Elements have an efficiency of 7%. So, the cooler would have to be at least 15 times OU.

Having ranted against OU, I am still convinced that a "heat pipe cooler" can be a big help when using a Peltier Element, because the cooler can be smaller, which is practical. And the size difference is big. And a "heat pipe cooler" can cool down to the melting point of the working fluid (a conventional air cooler only down to air temperature).

My humble low tech design: A platform (big or small, depending on the money you want to spend) floats on a lake or river. On the platform are many Peltier Elements arranged like ceramic tiles on a floor in a horizontal plane. The hot side is above and carries big heat catching laminates through which the wind pushes the rather warm air. The cold side is below and carries a heat sink (laminates) reaching deep into the water. Water in a lake is colder the deeper one goes down, water in a river might be cold because it comes from the mountains. This works day and night, as long as the air is warmer than the water, which is the case in the area where magpwr lives. It is not so in Austria, where the air is easily cooler than the water in winter and even in spring and autumn during a cool night. But one could flip around the platform when the air is cooler than the water. Today the air is -2° Centigrades and the water in the lake is about 8° Celsius (at least 4° at the bottom). So, the flipped around platform would work.

But, I think, going for photo voltaic on the equator is the better option. The best application for Peltier Elements is a Plutonium reactor giving off heat on Mars where the air temperature is minus 90° to minus 200° Celsius (or in space). Small problem, the Plutonium could kill you at least for 50.000 years. So, please, do not experiment with Plutonium.

Greetings, Conrad

Hi magpwr,

I have done some search on very low collector-emitter saturation voltage transistors and found the following type:
ZXT13N15DE6  Vcesat=5mV at Ic=100mA   and  45mV at Ic=1A  data sheet http://www.diodes.com/datasheets/ZXT13N15DE6.pdf (http://www.diodes.com/datasheets/ZXT13N15DE6.pdf)  and available at Mouser ( http://uk.mouser.com/Search/Refine.aspx?Keyword=zxt13n15 (http://uk.mouser.com/Search/Refine.aspx?Keyword=zxt13n15) )

Regarding the JFET types I wonder if you tested already the so-called switching or chopper type JFETs like J105, J106, J107, J108 or J109 ?  They have 3 Ohm to 18 Ohm drain source ON resistance, respectively, at zero gate source voltage.  What do you think? Their transconductance is not specified due to the switching purpose of operation but it is surely inherently pretty high.  Here are prices and data sheets: http://uk.mouser.com/Semiconductors/Discrete-Semiconductors/JFET/_/N-96ngcZscv7?Keyword=j10&FS=True (http://uk.mouser.com/Semiconductors/Discrete-Semiconductors/JFET/_/N-96ngcZscv7?Keyword=j10&FS=True)  Notice that in the Vertical Column Drain Current (Idss at Vgs=0) the currents are wrong (like 0.005mA) just study the data sheets.

Regarding your idea, it sounds very interesting and as a shaking solution you may wish to consider the vibro motor of cell phones: they are tiny electric motors with an small weight fixed to the rotor shaft in eccentric position. Unfortunataly they have very high RPM (several thousand) but maybe giving them much less supply voltage could help on this. Here is what I mean:
http://tinyrc.com/qfm/racing/dls/microspec.htm (http://tinyrc.com/qfm/racing/dls/microspec.htm)   but there are many types due to the variety of mobile or cell phones or pagers.

Thanks for all the great work.

rgds, Gyula

EDIT  magpwr: I have just read through your earlier posts and noticed you already mentioned the J series of JFETS are not good (if you really meant J105 to J109 as I wrote above) and you also hinted at using vibrating motors to replace Peltier cell.  So sorry for "reinventing the wheel" LOL  that was unintentional.

Hi gyulasun,

It seems the 4th generation transistor link you shown me (VCEO=15V; RSAT = 29mOhms ; IC= 5A) could compete with mosfet.

But those component are smd(Surface mount version).
For testing and creating prototype it's better to use transistor that could be added and removed easy from test board.

I have already tested all jfet series JXXX and all of those didn't work.I have mentioned it somewhere in my youtube channel title 25mv Joule thief.

I need to have time to work on my plans especially gas heatsink with peltier underneath, heatink under peltier as well.
If i move it by hand it should generate electricity.But i realised water could be created from condensation at the lower part of gas heatsink.
Unless peltier output is connected to joule thief to generate higher seeback which might prevent water formation from happening.Not tested yet.

Quote from: magpwr on January 16, 2013, 09:08:06 AM
Hi gyulasun,

It seems the 4th generation transistor link you shown me (VCEO=15V; RSAT = 29mOhms ; IC= 5A) could compete with mosfet.

But those component are smd(Surface mount version).
For testing and creating prototype it's better to use transistor that could be added and removed easy from test board.

...

Hi magpwr,

Thanks and I agree, it is not convenient to fiddle with SMD components.  For test purposes I often used SMD component prepared with color wires of 3-4 cm long soldered to the SMD so you actually deal with or plug into the testboard the end of short wires.

In the meantime I found similar very low saturation type, seems they are manufactured in SMD nowadays.
Vceo=20V  Rsat=18mOhm   Ic=7A       http://www.mouser.com/ds/2/115/ZXTN19020CFF-77912.pdf (http://www.mouser.com/ds/2/115/ZXTN19020CFF-77912.pdf)   

Good luck with your plans on gas heatsink with Peltier underneath.

rgds,  Gyula

Have you tried the J105 or TN0702 FETs for a low voltage JT (http://rustybolt.info/wordpress/?p=5155)?  The trick is to get it to start by itself.  I also put a 1.5V button cell in series with the gate to reduce the voltage needed to get it to conduct.  The gate draws zero current so the cell lasts almost forever.

You can get some good ideas from the schematic and information for a low voltage converter (.PDF) found here (http://www.ece.uvic.ca/~jbornema/Journals/064a-97ia-jmd.pdf).

Quote from: magpwr on January 04, 2013, 09:55:26 AM
Hi conradelektro,

Nice work man.This is how 1st started with <100mV Joule Thief using transistor 2SD1450 (0.13v..0.3v VCE(voltage drop between emitter and collector at 400mA.By referring to datasheet and looking at vce
for expected circuit current draw.

If you refer to my youtube channel  "sanjev21" "45mV Joule Thief" was my 1st low power joule thief  using  "2SD1450" Japanese transistor.But it was not self starting and i had to manually short emitter and collector of transistor to kick start oscillation.

My other experiment which i have not posted in youtube using the best NPN transistor i could find in digikey "ZTX1048A" which have the lowest Voltage drop across emitter&collector at mere(27mV...45mV) at 500mA.Little smaller than TO92 and could handle 4Amps or 20Amps pulse no "heat sink".
Using this ultra low drop transistor i could achieve oscillation as low as "30mV" but as usual i have to kickstart oscillation at higher voltage eg:60mV by temporary shorting emitter&collector.

I have also combine ZTX1048A transistor(To boost output current) and (2SK170 or 2SK364 or 3SK366 .But take note 2SK170 do have better lower starting voltage by 1..3mV)  Junction-FET(To self start osc at low voltage).Connecting 22ohms from coil to transistor and 1.8k coil to JFET Gate while connecting (JFET Source to transistor emitter,J-FET Drain to transistor Collector).Design for <100mV.

For this above transistor and JFET combined circuit i was able to charge 1000uf cap in matter of <5sec to reach around 2volts.

But the current drain for above transistor with JFET circuit  is  around 100mA .
I was using 50FARAD 2.5v Ultra capacitor(charged with 1AA battery for few seconds to reach 100mV) as power source.Since it's known ultra cap got very low internal resistance.

The above circuit will work very nice on 1.5V battery but need to tweak 22ohms resistor to higher value to maintain current draw at decent level eg:100mA.
Then again i realised it's pointless using JFET for 1.5v battery since we could use transistor which is more efficient at higher current.

Peltier only produce <5mA if I connected to multimeter if i use my hand as heat source(36.5c) against room temperature in my case(30c for a hot country) 6.5C difference.
The peltier i'm using do have internal resistance of around 3ohms(Out of the box /Not in use).
In theory i'd believe if we use lower internal resistance for peltier <2ohms we may get better current output but then again.I 'm not sure if multi meter could properly detect "current" at such low voltage produce from peltier 70mV..90mV.

Solar cell do work to power my JFET base joule thief in room lighting or led light.I'm not sure if using other power source like from crystal radio could work.

If we match "source and circuit impedance" then we could  further optimize on efficiency.


Lastly the circuit do work on 1 x 2SK170 as well with lower Circuit current draw at 1mA using peltier at around 60mV. I used 4 JFET  since i noticed there is improvement with output voltage especially with Led as load current draw increases to 4.2mA(Mentioned in youtube description).

Lastly if anyone wish to go for extreme low starting voltage at 22mV instead of 25mV got to try 2n3972 JFET (But very low current handling in itself output capacitor cannot be more than 100nf) and combine with 2SK170 x 3 or 4.

If anyone is a expert with SMD base tiny circuits there is UK base company which manufacture 2SK170 as LSK170  (TO92 package and SMD tiny version).
http://www.micross.com/packaged-parts-plastic-discretes.aspx (http://www.micross.com/packaged-parts-plastic-discretes.aspx)

Quote from: acmefixer on January 18, 2013, 08:45:29 AM
Have you tried the J105 or TN0702 FETs for a low voltage JT (http://rustybolt.info/wordpress/?p=5155)?  The trick is to get it to start by itself.  I also put a 1.5V button cell in series with the gate to reduce the voltage needed to get it to conduct.  The gate draws zero current so the cell lasts almost forever.

You can get some good ideas from the schematic and information for a low voltage converter (.PDF) found here (http://www.ece.uvic.ca/~jbornema/Journals/064a-97ia-jmd.pdf).

@acmefixer: How good and interesting magpwr's circuit with the 2SK170 transistor is (if one has a very low Voltage power supply), can be seen from the fact that neither the circuits discussed in your blog nor the circuit from the PDF document can reach down to 25 mV. The voltage converter in the PDF-document needs at least 0.3 Volt and the circuits in your blog seem to use ordinary 1.5 Volt batteries.

I carefully verified magpwr's claim. With 4 transistors 2SK170 the circuit consumes about 2.2 mA at 50 mV (110 µW); a white LED glows nicely, of course not super bright. And one can make a LED glow by touching the hot side of a Peltier Element with the hand (the cold side at room temperature).

In case one uses only one 2SK170, the power consumption at 50 mV is about 500 µA (25 µW, the LED glows less bright)

Now, show me a circuit and a transistor which can do better (less than 50 mV supply Voltage and less than 500 µA supply current to make a white LED glow, self starting). This might not be very useful, but it is an interesting experiment.

Greetings, Conrad

Hi acmefixer,

It's been some time since you provided suggestion on ZTX transistors in youtube.I did purchased ZTX1048A and ZTX pnp version.
"I will be revealing a "one button" programmable code lock using pic12f629 and  using ZTX transistors in H-bridge configuration using one 1 AA battery to power gear motor. in youtube"
Once 70rpm motor arrives.The standby power is <15uA.It's a unique  circuit using only 1xAA battery to operate microprocessor and power the latch or lock example lock for drawer.


I did waste money purchasing  just 2x J105 while waiting for (10 x 2SK170)for 2USD to be delivered since the spec for JFET although it looks like it was the best N-channel JFET with lower turn on resistance.

As usual i start test using ultra cap 50 Farad charged to around 100mv using 1AA battery for 1 to 3 seconds.

I placed the JFET into same toroid with 4:100 but oscillation did not happen <100mV.

From the link you provided it looks it will work around 300mv maybe.

The circuit does needs alot of components which may increase current draw.

@magpwr  I needed a low voltage source, so I started experimenting with a thermocouple.  I got this from an old water heater or furnace, so long ago I can't remember which.  I put it in the flame of the burner on my stove and it measured 42 millivolts.  I connected a 1 ohm resistor across it, and the DMM dropped 2 mV to 40 millivolts.  So it is capable of driving a very low resistance load.

The Joule Thief takes about 130 milliwatts and 88 milliamps at the normal 1.5V, and much more current at lower voltage if you want the LED to be lit to its full brightness of 20 milliamps.  So getting a Joule Thief to run at less than 1/2 volt is a very difficult and challenging project.  I think that it would be much, much better to put several peltiers or thermocouples in series to get at least a few hundred millivolts, so that the Joule Thief doesn't have to use power MOSFETs in parallel to get the power the LED needs.  I also think that working at 40 millivolts from the thermocouple is just a really big problem, since even the shortest leads of the components can have tens of millivolts drop at high currents.

You're going to have to find out this for yourself.  I know from my experiments at a half volt from a single solar cell.  Best of success, and good luck.

Hi Folks,

I found a Lin Tech "energy harvesting" product line for solar, Peltier/Thermoelectric  and Piezo electric voltage sources and they work as DC-DC converters.
Here is the full family:  http://parametric.linear.com/Energy_Harvesting   and this is an IC able to start and operate from already 20mV DC difference:  http://www.linear.com/product/LTC3108   it produces 4 different DC outputs between 2.35V to 5V.

Its working principle is that the minimum 20mV DC input feeds a step up transformer and an inner switch chops up the DC input via the primary of the transformer and the up-transformed pulses at the secondary of the transformer is tuned to resonance by outside capacitors and this AC is rectified and taken care of by the built-in circuits.  An interesting solution.  Digikey, Farnell or others have it.

rgds, Gyula

Hi  gyulasun,

I'm did research on LTC3108 previously.

But the key component looks like too tiny for me to work on.Besides it needs more discrete component eg:capacitor.

But price of component for my 25mV joule Thief likely below < 8USD and the efficiency could be comparable with LTC3108.

The efficiency is not tested yet.Base on input voltage x current and output voltage x current.

@gyulasun: Very interesting, thank you for drawing our attention to the LTC3108 and its family of DC-DC converters.

This shows what can be done professionally with a low Voltage source like a Peltier Element. The price for a stable output Voltage and the other niceties (PGOOD)  is of course a few mA used up by the many electronic components.

The high price of about Euro 10.--  probably means that the market demand is low.

In the data sheet one reads: "The LTC3108 can also be used to trickle charge a standard capacitor, supercapacitor or rechargeable battery, using
energy harvested from a Peltier or photovoltaic cell."

But exactly for this application magpwr's very simple circuit might be better, because one does not need all the high end features (PGOOD, noise reduction with various capacitors, three different stable output Voltages).

Once more I am thinking about building a very simple device with two Peltier elements that charges a small 1.2 V rechargeable battery whenever I light the ceramic stove in my living room. See the attached drawing, the "device" is intended to lean against the ceramic stove.

Greetings, Conrad

I made a little play thing which can be leaned against my ceramic stove and drives magpwr's circuit.

Greetings, Conrad

P.S.: I know there are units on sale which put out 2 Watt http://biolitestove.com/campstove/camp-overview/tech-specs/#sub (http://biolitestove.com/campstove/camp-overview/tech-specs/#sub) , my lean-to Peltier Gadget is just for fun.

Hi Folks,

In the data sheet of the 2SK170 you can read that this type is selected into 3 groups when manufactured and these groups are identified with different suffix in the type name.  The selection is based on the drain current,  I_DSS measured at zero gate-source voltage.

These are  2SK170GR  which has  I_DSS = 2.6~6.5mA   
                  2SK170BL  which has  I_DSS =  6~12mA
                  2SK170V    which has  I_DSS =10~20mA

I wonder which suffix you guys happened to purchase because obviously the 2SK170V has the smallest R_DSON  and the highest Y_fs forward transfer admittance. 

I remembered a Design Idea from the Electronic Design News Magazine, see this link: http://www.edn.com/design/power-management/4320557/JFET-based-dc-dc-converter-operates-from-300-mV-supply   
The writer of the article works (or worked) for Linear Technology and most probably he established the base circuit for the later creation of the LTC 3108 type DC-DC converter family.  The circuit shown in the Design Idea works from 300mV DC as a minimum input and includes several components, so it is rather complex compared to magpwr's circuit but the reason I mention it is that it may serve as a "food for thought" for anyone here wishing to experiment with these setups, showing certain solutions with discrete components.

@Conrad      Very nice build, thanks for showing your setup with the stove.

rgds, Gyula

Hi gyulasum,


This is a interesting finding which shows current handling capability 3 different version of  2SK170.

Just noticed the 10 pieces of (2SK170 BL  6~12mA version) which i purchased from www.futurlec.com (http://www.overunity.com/www.futurlec.com) for $0.20 ea .It's found through search "2SK170" for that site.

It's unknown if slightly higher current handling capability would have impact on "input voltage  detect sensitivity".

Unknown if the 2SK version is similar like those for transistor eg:BC547A,BC547B,BC547C which have different gains.

I noticed that base on datasheet BF862 and J105 does have better RDS/Lower on resistance than 2SK170.
But unfortunately  the startup voltage for BF862 and J105 is not really impressive at around 300mV.There maybe relationship between current capability and voltage\current detect for JFET base on current technology.

To increase current output\input at low voltage it would be more sensible to use transistor EG:ZTX1048A with 2SK170.

I have tested the J105 this afternoon which was disappointing which could not start at 0.1v but does start oscillation at 0.3.. 0.4v

@gyulasun: you are finding interesting circuits.

The two circuits shown in this publication http://www.edn.com/design/power-management/4320557/JFET-based-dc-dc-converter-operates-from-300-mV-supply derive the output Voltage from the primary (the coil with the few turns) at the Drain (Collector) of the transistor.

An then, according to the block diagram (you have to look up the data sheet), in the http://www.linear.com/product/LTC3108 the output Voltage is derived from the secondary (the coil with the many turns) at the base of the transistor (like in magpwr's circuit).

That might explain the low 20 mV supply Voltage limit in the LTC3108 (and the 25 V in magpwr's circuit) and the 300 mV in many other circuits.

I got the 2SK170 BL transistors from Reichelt (Germany) and Mouser (have a subsidiary in Germany, but usually deliver from the USA).

Greetings, Conrad

If you make the coil that it has a greater turns ratio, the JFET will start oscillating at a lower voltage.  Remember that at zero gate voltage, the JFET will conduct current, so it's a matter of getting the loop gain above unity with the supply voltage that you have.  Also I believe that with a supply at a few dozen millivolts you will not be able to get the output to a useful level unless you use two stages - one to get it started and another to switch high supply current.

Also, remember that "An hour in the library is worth ten in the lab".

See atttachment.

Quote from: magpwr on January 19, 2013, 09:08:29 PM
Hi gyulasum,


This is a interesting finding which shows current handling capability 3 different version of  2SK170.

Just noticed the 10 pieces of (2SK170 BL  6~12mA version) which i purchased from www.futurlec.com (http://www.overunity.com/www.futurlec.com) for $0.20 ea .It's found through search "2SK170" for that site.

It's unknown if slightly higher current handling capability would have impact on "input voltage  detect sensitivity".

Unknown if the 2SK version is similar like those for transistor eg:BC547A,BC547B,BC547C which have different gains.

I noticed that base on datasheet BF862 and J105 does have better RDS/Lower on resistance than 2SK170.
But unfortunately  the startup voltage for BF862 and J105 is not really impressive at around 300mV.There maybe relationship between current capability and voltage\current detect for JFET base on current technology.

To increase current output\input at low voltage it would be more sensible to use transistor EG:ZTX1048A with 2SK170.

I have tested the J105 this afternoon which was disappointing which could not start at 0.1v but does start oscillation at 0.3.. 0.4v

Hi Acmefixer,

The 2SK170 absolute lowest possible voltage still remains at 25mV for oscillation to start&continue.It's not about the coil turns ratio anymore.Already tested it.
At 24mV it's no guarantee.

My latest discovery J210 N-channel JFET will work very well with ZTX1048A transistor.But lowest startup voltage is at "35mV" for JFET to kickstart transistor (Collector connected to drain,emitter&source).If ZTX transistor kick started manually on it's own lowest supported voltage is 30mV for oscillation to continue.
Secondary coil to gate of "J210" must be 330KOhms and secondary coil to ZTX1048A must be 330 Ohms at 35mV input.If resistor goes 150k or below to the gate the JFET will not work at all unless it's around 0.3volts .One more thing the J210 series JFET doesn't do oscillation on it's own at  voltage <100mV it merely produces alot of spike provided using resistor 330kOhms to it's gate...to automatic kickstart transistor.
I was surprise base on above configuration the voltage could exceed 10volts output for 35mV input(10v/0.035V =285 x voltage boost) input at around 6mA current draw ,circuit output capacitor using 100nf.Current still remains sufficient for led to be lit.

2SK170 and ZTX transistor combination still produce the best result.

The lowest I've been able to get a JT to go is 80 millivolts, however I didn't really try to optimize the circuit. You may be able to do better.  I used a button cell to get the bias voltage I needed to get it to work at low V.  I used the TN0702, and you can see the circuit here (http://rustybolt.info/wordpress/?p=171).  But click on the image several times to get it to enlarge.  You may be able to get yours to do better by using the bias voltage method I used.

Quote from: conradelektro on January 20, 2013, 05:56:02 AM
@gyulasun: you are finding interesting circuits.

The two circuits shown in this publication http://www.edn.com/design/power-management/4320557/JFET-based-dc-dc-converter-operates-from-300-mV-supply (http://www.edn.com/design/power-management/4320557/JFET-based-dc-dc-converter-operates-from-300-mV-supply) derive the output Voltage from the primary (the coil with the few turns) at the Drain (Collector) of the transistor.

An then, according to the block diagram (you have to look up the data sheet), in the http://www.linear.com/product/LTC3108 (http://www.linear.com/product/LTC3108) the output Voltage is derived from the secondary (the coil with the many turns) at the base of the transistor (like in magpwr's circuit).

That might explain the low 20 mV supply Voltage limit in the LTC3108 (and the 25 V in magpwr's circuit) and the 300 mV in many other circuits.

I got the 2SK170 BL transistors from Reichelt (Germany) and Mouser (have a subsidiary in Germany, but usually deliver from the USA).

Greetings, Conrad

An experiment with two 30 mm x 30 mm Peltier Elements (inexpensive Peltier Elements):

The heavy brass mortar (big strong walled cup) stays at room temperature a long time. One can fill cold water, ice or snow into the plastic container placed on top of the brass mortar (which stands up side down on the table).

A rectangular hole at the bottom of the plastic container is covered by an aluminium flat bar.

The two 30 mm x 30 mm Peltier Elements are glued to the outside of the aluminium flat bar.

Greetings, Conrad

Quote from: acmefixer on January 20, 2013, 08:59:34 AM
If you make the coil that it has a greater turns ratio, the JFET will start oscillating at a lower voltage.  Remember that at zero gate voltage, the JFET will conduct current, so it's a matter of getting the loop gain above unity with the supply voltage that you have.  Also I believe that with a supply at a few dozen millivolts you will not be able to get the output to a useful level unless you use two stages - one to get it started and another to switch high supply current.

Also, remember that "An hour in the library is worth ten in the lab".

See atttachment.

Hi acmefixer,

I would be interested to read the patent you show a schematic from in your attachment. Could you give the patent number?

I have read your blog, seen your schematics with the button cells to create gate bias for the TN0702.  I believe you could include in the same circuit a JFET oscillator like magpwr has shown with the 2SK170 and the only purpose for this oscillator would be to produce the 2-3V DC bias voltage.  By using a third coil next to the needed two coils for this oscillator, you could galvanically isolate the output of this oscillator so the biasing of the TN0702 (or other MOS or JFETs) should not be a problem. And then you could run the whole circuit from as low as your 80mV DC without the bias cells, and without a significant input power increase.

rgds, Gyula

hi conradelektro,

Nice work.

I wonder if we use hot water against room temperature into a small pot.Maybe it could be turned into a small dinner's light on the restaurant table with green concept in mind.
It would make environmentalist happy since there is no candles to burn this time.Maybe like a led candle floating on hot water in small pot.

To Acmefixer,
Unfortunately mosfet is not the solution for input voltage below <100mV.Noticed you are using button cell in your circuit. I'm aware mosfet need to driven with higher voltage unlike transistor.
Please do try with Junction-FET which it's the only workable\self starting solution for voltage below <100mV.

Hi magpwr,

Here is quote from the 2SK170 data sheet:  High |Yfs|: |Yfs| = 22 mS (typ.) (VDS = 10 V, VGS = 0, IDSS = 3 mA)

and I think the key is: the IDSS is mentioned to be 3mA where the device has this high transfer admittance Yfs=22mS. I have not seen any JFET type for which as high Yfs as this would have been specified at such a low drain current as 3mA.
I also deduced the same conclusion from the data sheets of BF862 or J105 etc jfet types that they may be better candidates but your practical test for the J105 proved just the opposite, and I wonder if you have tested the BF862 too?  Because in the Design Idea in which the BF862 was shown to work from 300mV DC supply the schematic is different from your 2SK170 circuit so correct comparison can only be done by using the same circuits i.e. yours with the BF862 instead of the 2SK170. IF you have already done this, then sorry for this rant.

rgds, Gyula

Quote from: magpwr on January 19, 2013, 09:08:29 PM
Hi gyulasum,


This is a interesting finding which shows current handling capability 3 different version of  2SK170.

Just noticed the 10 pieces of (2SK170 BL  6~12mA version) which i purchased from www.futurlec.com (http://www.overunity.com/www.futurlec.com) for $0.20 ea .It's found through search "2SK170" for that site.

It's unknown if slightly higher current handling capability would have impact on "input voltage  detect sensitivity".

Unknown if the 2SK version is similar like those for transistor eg:BC547A,BC547B,BC547C which have different gains.

I noticed that base on datasheet BF862 and J105 does have better RDS/Lower on resistance than 2SK170.
But unfortunately  the startup voltage for BF862 and J105 is not really impressive at around 300mV.There maybe relationship between current capability and voltage\current detect for JFET base on current technology.

To increase current output\input at low voltage it would be more sensible to use transistor EG:ZTX1048A with 2SK170.

I have tested the J105 this afternoon which was disappointing which could not start at 0.1v but does start oscillation at 0.3.. 0.4v

I also believe it's a patent, at least it looks like ones I've seen before.  I believe it was posted in a forum, may even have been this one, and I saved it.  That's all I know.  The point I wanted to make was that in the patent schematic, they drive the gate with the same winding that is used for the output.

Thanks you for reading up on my schematics and blog.  I'm glad that someone has found it useful.  With a pot and/or resistors totaling 1 meg or more across it, the button cell's lifetime is in years, so there's no problem with it being used. I even used a 0.1 uF plastic capacitor connected to the gate, and just charged it up by touching a 1.5V cell to it.  It stayed charged for many hours, so I could use an even larger value and charge it up and it would stay charged for days or weeks, and still do the job of biasing the gate.

Today I have been experimenting with the J105 and I was able to get it to oscillate at 60 millivolts.  But I ran out of space on the core and couldn't wind any more wire on it..  I think if I wind another larger coil with several hundred turns, it should be able to go down to as low as 30 or less mV  You can read my blog about it (http://rustybolt.info/wordpress/?p=5881).

I just had an idea that will save me a lot of work.  But I have to try it to see how well it works.  More about it in an update to my blog.

Best of success in your endeavors.

Quote from: gyulasun on January 20, 2013, 06:09:50 PM
Hi acmefixer,

I would be interested to read the patent you show a schematic from in your attachment. Could you give the patent number?

I have read your blog, seen your schematics with the button cells to create gate bias for the TN0702.  I believe you could include in the same circuit a JFET oscillator like magpwr has shown with the 2SK170 and the only purpose for this oscillator would be to produce the 2-3V DC bias voltage.  By using a third coil next to the needed two coils for this oscillator, you could galvanically isolate the output of this oscillator so the biasing of the TN0702 (or other MOS or JFETs) should not be a problem. And then you could run the whole circuit from as low as your 80mV DC without the bias cells, and without a significant input power increase.

rgds, Gyula

This Youtube video (http://www.youtube.com/watch?v=L_ybrn2AMiQ) shows one experimenter's way of doing what you're trying to do.  It may be a seventy year old low tech way, but it does the job.

I'm still experimenting with four J105 JFETs in parallel and I'm getting the LED to light at about 80 millivolts.  See my blog.

Here's another experimenter's circuit (http://panteltje.com/pub/lighting_a_LED_with_a_candle_IMG_3604.GIF).  Lights a LED with a candle.  The circuit converts 30 to 40 millivolts.

Quote from: magpwr on January 20, 2013, 10:12:26 AM
Hi Acmefixer,

The 2SK170 absolute lowest possible voltage still remains at 25mV for oscillation to start&continue.It's not about the coil turns ratio anymore.Already tested it.
At 24mV it's no guarantee.

My latest discovery J210 N-channel JFET will work very well with ZTX1048A transistor.But lowest startup voltage is at "35mV" for JFET to kickstart transistor (Collector connected to drain,emitter&source).If ZTX transistor kick started manually on it's own lowest supported voltage is 30mV for oscillation to continue.
Secondary coil to gate of "J210" must be 330KOhms and secondary coil to ZTX1048A must be 330 Ohms at 35mV input.If resistor goes 150k or below to the gate the JFET will not work at all unless it's around 0.3volts .One more thing the J210 series JFET doesn't do oscillation on it's own at  voltage <100mV it merely produces alot of spike provided using resistor 330kOhms to it's gate...to automatic kickstart transistor.
I was surprise base on above configuration the voltage could exceed 10volts output for 35mV input(10v/0.035V =285 x voltage boost) input at around 6mA current draw ,circuit output capacitor using 100nf.Current still remains sufficient for led to be lit.

2SK170 and ZTX transistor combination still produce the best result.

Hi Acmefixer,

Interesting circuits links.Noticed one of the circuit is using 10xBF245C (JFET) but using candle which produce higher temperature to light led.

I have done testing on BF245 and many other JFET series which have posted in forum base on startup voltage but noticed the for current handling 2SK series JFET is still the better choice.

It's interesting searching for a interesting article on circuits which is good and yet hard to find.It's good i  posted this circuit to share so that someone may come up with better version or design and we improve it from there.

I'm curious as to why you choose 2SK170 over others.  Toshiba says it is low noise, designed for audio preamplifiers, which is far from being used as a switch.  There are JFETs that are designed for switching, which is what the Joule Thief is really doing.  What do they do that's better and makes you choose the 2SK170 over those?

Thanks.

Quote from: magpwr on January 22, 2013, 07:44:27 AM
Hi Acmefixer,

Interesting circuits links.Noticed one of the circuit is using 10xBF245C (JFET) but using candle which produce higher temperature to light led.

I have done testing on BF245 and many other JFET series which have posted in forum base on startup voltage but noticed the for current handling 2SK series JFET is still the better choice.

It's interesting searching for a interesting article on circuits which is good and yet hard to find.It's good i  posted this circuit to share so that someone may come up with better version or design and we improve it from there.

Hi Acmefixer,

It's a good question how i came to find 2SK170 to be the best.

There was this crazy experiment i was doing with same toroid used for my 45mV joule thief which i was using Japanese transistor 2SD1450 but it not the self starting version,as shown in youtube "sanjev21".

I did a lot of research and found that mosfet can never be used in joule thief,because it do not release immediately(in ms or ns) after switching on.
Althought mosfet is a "excellent switcher" if combine with microprocessor.

It was a long shot.There was youtube video whom posted about "self starting stirling engine" using peltier at the source of heat to spin motor to start engine.
It was using mosfet 2N7002 with inductor to positive.The startup voltage was laround 0.4v.

I experimented using "multisim a virtual electronic lab".The transformer in it i'm stuck with "TS_POWER_25_TO_1" (100% efficient) 25 turns:1.(The only transformer i could use for ultra low voltage joule thief)
Tried all the transistor,mosfet,fet,jfet.Then found JFET had the self starting capability.Then spent 3 days to find the best one then slowly lowered source voltage to 0.025v.

I have "9 version of ultra low voltage joule thief" virtual lab file but only 3 is worthy of mentioning as current handling for the rest is not good enough.

JFET standalone version as shown in my youtube video, better version is combining ZTX1047A or ZTX1048A with 2SK170(Since transistor got better current capability using JFET to auto kickstart transistor as well),then there is a version
which i used JFET with 2 resistor to control the current going to transistor.The problem with last version the startup voltage is little higher
but give me "constant current draw" all the way to 2v source.

I tested with 40 over key component or components in virtual environment to derive this ideal circuit by accident.Please ignore 330Ohms resistor to diode as i placed it to allow circuit to function if provided i'm using 1000uf cap at output for 25mV input.

Hi Folks,

In the meantime I have done some patent search on this low voltage DC-DC converter topic and I think I found good info.

First of all I found the patent that includes the schematic Acmefixer showed in Reply #94. It is US4734658 (Low voltage driven oscillator circuit), see here:  http://www.google.com/patents/US4734658 (http://www.google.com/patents/US4734658)   (By the way here is a very similar circuit shown much earlier: http://www.google.com/patents/US3681674 (http://www.google.com/patents/US3681674) ). Both are good reading for those wishing to get a deeper insight into very low voltage converters.

Here is another patent which includes what kind of features a FET should have ideally for such low voltage converters, see this: http://www.google.com/patents/US4015609 (http://www.google.com/patents/US4015609) 
So for a FET the drain-source ON resistance should approach zero, of course and the gate-source cut-off voltage (or threshold voltage for MOSFETs) should be also near to zero (so that a tiny mV amount of change in the gate-source voltage should cut the drain current off). In the patent you can see an attempt to improve BOTH characteristics of the off the shelf JFETs of that era.
To reduce the ON resistance, the inventor also paralleled some FETs. And to reduce the cut-off voltage, the inventor connected two such paralleled FET groups in series: the joined source pins of the paralleled group A is connected to the joined drain pins of the paralleled group B, while all the gates in groups A and B are joined to be a single gate point, the main drain pin is the drains from group A and the main source is the sources from group B. This way the resulting cut-off voltage is reduced below that of any such single JFET. (Of course the series connection of the groups increases the ON resistance BUT the advantage received from the lower cut-off voltage is more beneficial from the input DC supply voltage level point of view.)

So magpwr's finding the 2SK170 type and paralleling them was the correct first step to reduce the ON resistance, hence to be able to reduce the working supply voltage. And luckily, this type is characterised as having high transconductance (22mS) at 3mA drain current i.e. it has high enough gain to start oscillation at a low supply voltage and this transconductance is enhanced each time your paralleling the FETs. AND the cut-off gate-source voltage for this type is between  -0.2 to -1.5V only, the  -0.2V goes with the GR suffix and the  -1.5V goes with the V suffix, while the BR suffix is somewhere inbetween.

Now we know why the higher current switching type JFETs like J105 etc may not work  from a lower supply voltage than the 2SK170: because these JFET types inherently have -5 to -10V (or higher) gate source cut-off voltage ranges and this is a drawback in this respect. (The higher cutoff gate voltage inherently decreases transconductance.)

The following correct step was to turn to the ALD110800 and ALD110900 MOSFET types (from TinselKoala link) as these are advertised as zero-voltage-threshold MOSFETs. However their big drawback is the 104 kOHm drain-source ON resistance at the zero gate-source voltage (this huge drain-source resistance goes down to 500 Ohm when the gate-source voltage is increased to +4V from zero).  So it seems this zero voltage threshold feature cannot bring advantage in such low supply voltage converters. Of course these new MOSFET types can work very nicely in circuits where the high ON resistance is not neccessarily a drawback and can be matched.

Then I found  a patent which shows an interesting solution to reduce the JFETs relatively high ON resistance so that overall efficiency may be increased.  By the way, even if the JFETs are paralleled, they still can have rather high ON resistance, just consider the 2SK170BR, it surely have a drain-source ON resistance between 800 to 1600 Ohm at zero gate-source bias and if you parallel 4 of them, this range reduces to between  200 and 400 Ohm. So compare this to the primary coil's self impedance which is surely well under 1 Ohm at DC and probably represents a few times ten Ohm at the oscillator's frequency, it is a transformed impedance from the gate side of the fet, depends on the primary/secondary turns ratio too.)

While some circuits use bipolar transistors connected in parallel (as a second switch) with the jfet or MOSFET (this is discussed here: http://www.google.com/patents/US4322724 (http://www.google.com/patents/US4322724) ) but here in this patent a MOSFET is used in parallel with the jfet, see here Fig.3: http://www.google.com/patents/US7170762 (http://www.google.com/patents/US7170762) 

All in all, what I deduce from all these patents info, including the work of magpwr, acmefixer, conradelektro and scratchrobot,  I think you need to use at least 1:80 to 1:100 or even higher turns ratio, you may need to use resonance at the secondary coil of the transformer (see the tuning capacitors at the secondary coil: http://www.linear.com/product/LTC3108 (http://www.linear.com/product/LTC3108)   and you can combine either low saturation bipolar transistors or MOSFEts in parallel with the 2SK170 jfet as a second switch, controlling this latter together with the jfet from the up-transformed input voltage. If you have a logic level MOSFET that can fully conduct at say 1V gate-source voltage, and you have (say) 1:100 turns ratio, then 10mV across the primary coil will insure operation of the MOSFET in theory BUT to have 10mV voltage drop across the primary coil, the voltage drop across your jfet defines input sensitivity because the full input voltage gets divided between the primary coil and the jfet's ON resistance. So paralleling the 2SK170 jfets is still needed I am afraid to self-start the oscillator from as low supply voltage as possible. Can you all agree with this?

rgds,  Gyula

@gyulasun

You did a very useful research. I will try to understand your conclusions by going through the many references you posted. With my bigger toroids I can do the 1:100 reatio (e.g. 5:500) with the 0.22 mm enamelled wire. Thinner wire is just to awkward to do by hand.

Just a hint, here you can look up every patent you wish to see:

http://worldwide.espacenet.com/singleLineSearch?locale=en_EP (http://worldwide.espacenet.com/singleLineSearch?locale=en_EP)

http://worldwide.espacenet.com/advancedSearch?locale=en_EP (http://worldwide.espacenet.com/advancedSearch?locale=en_EP)

What I want to do, is to make a LED glow in the garden with a copper and an aluminium tube in the ground (as a basic galvanic battery). The copper and aluminium tube in the ground show a Voltage of about 400 mV, but the Voltage brakes down below 50 mV when connecting magpwr's circuit. May be at 10 mV, the tubes can supply enough current to reach the 100 µW necessary to make LED glow.

Magnesium and copper might work, but I do not like magnesium, too expensive and all sorts of side effects. I still have to do a test with a galvanised steel tube and a copper tube, but at the moment everything is frozen solid in the garden under 0.2 meters of snow.

That's why I play with snow and a Peltier Element (see the attached photos). The little saucer is tin, the big mortar (cup) underneath is brass.

Greetings, Conrad

Quote from: conradelektro on January 27, 2013, 05:41:32 AM
...
With my bigger toroids I can do the 1:100 reatio (e.g. 5:500) with the 0.22 mm enamelled wire. Thinner wire is just to awkward to do by hand.

...

Hi Conrad,

Thanks and may I suggest to make taps on the secondary coil at  1:50 and 1:75  turns ratio to make matching more flexible. Also, you could consider using ferrite pot cores or normal ferrite E cores instead of the toroid core shape, these latter cores are much easier to wind.

rgds, Gyula

hi conradelektro,

In theory maybe moving the rods closer to each other would reduce the resistance or internal- resistance(In the case of battery) for your garden battery.

It reminded me of a crazy experiment done around 15 years ago using zinc/carbon battery case after removing all the black stuff in battery.
I actually stuffed fertilizer in it which did not work well.Then i added little water with fertilizer to make it moist and throw mix back into battery case.
I then charged with around 7.5v to battery for 2min.I then placed the battery into a mechanical movement clock design to work on 1xAA battery.
The clock actually worked for few days then "fertilizer AA battery" died.

hi gyulasum,

Thanks for providing some interesting links.It does help us to promote creativity in each of us.I never knew people actually patent ultra low voltage voltage booster.
Good thing mine is unique and i will never patent my design especially for simple "electronics".It would be a waste of effort or time maybe money to patent.

---------


I would be planning to work on this project to compliment my 25mV joule thief.It's using radio wave to charge battery.I will be using it with ultra-caps.
http://www.youtube.com/watch?v=tjxldE8JsDw (http://www.youtube.com/watch?v=tjxldE8JsDw)

I was active in this forum where there was discussion on the "RF diode charger" and etc.
http://laserhacker.com/forum/index.php?topic=104.0 (http://www.laserhacker.com)

I noticed it's very difficult to find antique items in my country like 1N34A that i have to order online from ebay.
I had problems even finding small copper tubes normally used in Air-con smaller version for copper tube\toroid transformer to experiment on resonance as i mentioned earlier in this forum.

@gyulasum:

I got this core http://at.farnell.com/ferroxcube/etd49-25-16-3c90/ferrite-core-half-etd49-3c90/dp/3056417?Ntt=3056417 (http://at.farnell.com/ferroxcube/etd49-25-16-3c90/ferrite-core-half-etd49-3c90/dp/3056417?Ntt=3056417) (two halves, bobbin and clips). Would this be any good to wind the 5:500 coil onto? A bit big, but for a test size does not matter?

And I got this even bigger core: http://at.farnell.com/ferroxcube/etd59-31-22-3c90/ferrite-core-half-etd59-3c90/dp/3056430?Ntt=3056430 (http://at.farnell.com/ferroxcube/etd59-31-22-3c90/ferrite-core-half-etd59-3c90/dp/3056430?Ntt=3056430)

@magpwr:

I was not successful with the radio wave receivers (as a power source). But I might have done something wrong.

I tried the copper and aluminium tube very close together in the earth, no success. It works very well when the tubes are wrapped in paper soaked in salt water, but I do not want to throw a lot of salt into my garden. But I will try fertilizer around the tubes in spring.

I like my "magpwr circuit in Epoxy" (see the photo in my last post) very much (besides its odd shape). I used it submerge in water, no problem. I will buy some clear casting resin and make a few more for outdoor use.

Greetings, Conrad

Hi Conrad,

Yes, the smaller sized core ETD49/25/16 is already okay, no need for the higher sized one. And do make some taps on the secondary if you agree.  Your question on the core size: on a bigger core the coils will have higher inductance than on a smaller one (cross section area is higher), this means  the oscillator would operate at a lower frequency with the bigger core.  Your E cores are surely good from some Hz to at least 100 kHz or maybe higher (3C90 material) so you have a very wide range available, by tuning the secondary with some hundred pF or maybe 1nF capacitor in parallel with it.  On your smaller core the 500 turns will give L=N²*A_L=500*500*4.2uH=1050mH   i.e.  roughly 1 Henry inductance, this will resonate at about 22 kHz with a single 2SK170 JFET's gate-source input capacitance (nearly 50pF at 0.3V drain-source voltage from data sheet) if you do not use any other tuning capacitor. In case of 4 such FETs in parallel the frequency may go down to 11kHz or so.   (Of course these frequencies are rough estimates.)

Gyula

Hi magpwr,

I also disagree with patenting in general, however, design engineers with big ego tend to do so and boost with their patents, you know it  sound good for them...  LOL

Regarding the radio wave charging battery idea: problem is I think that even if you are lucky to collect a few Volts in a capacitor of some thousand microfarad (not in a supercap) it does need a certain time to collect it, and if you wish to charge ultra or super caps by radio waves, it may take days to have them filled up to a certain level unless you live near to a higher power transmitter. Crystal radio receivers have in general a high output impedance and a charging capacitor is definitely a low impedance device, albeit it demands an exponentially decreasing current from an initial high value when fully or nearly discharged. However there exists a so called step charge process when you insure the voltage difference between the charger output and the capacitor is always small and gradually increased in small steps so that no huge inrush current. This step charge method is also a long process though and the control circuit also needs some power to work from.
Regarding the Germanium diode rectifiers, you may replace them with syncronous MOSFET switches IF you already have an RF voltage of near 1 to 1.5V across a resonant LC tank circuit fed from an antenna. I mean syncronous rectifier as shown with the ALD MOSFEts in the Cutler.pdf in TinselKoala's link BUT use logic level MOSFEts instead of the ALD types, the logic level types can already open and conduct at 1V or 1.5V positive peak voltage on their gate with respect to their source.

It is unfortunate that in your country copper tubes are a problem to buy. On ebay here are some copper tubes with an outside diameter of

1/8 inch: http://www.ebay.com/sch/i.html?_trksid=p5197.m570.l1311.R6&_nkw=1%2F8+copper+tube&_sacat=0&_from=R40

3/8 inch http://www.ebay.com/sch/i.html?_trksid=p5197.m570.l1313&_nkw=copper+tube+3%2F8&_sacat=0&_from=R40
             http://www.ebay.com/itm/Copper-Coupling-For-3-8-O-D-Tubing-One-Coupling-/400298753167?pt=LH_DefaultDomain_0&hash=item5d33aa3c8f

1/2 inch http://www.ebay.com/sch/i.html?_trksid=p5197.m570.l1311.R9&_nkw=1%2F2+copper+tube&_sacat=0&_from=R40

Regards, Gyula

Quote from: magpwr on January 27, 2013, 06:56:44 AM
...
hi gyulasun,

Thanks for providing some interesting links.It does help us to promote creativity in each of us.I never knew people actually patent ultra low voltage voltage booster.
Good thing mine is unique and i will never patent my design especially for simple "electronics".It would be a waste of effort or time maybe money to patent.
---------
I would be planning to work on this project to compliment my 25mV joule thief.It's using radio wave to charge battery.I will be using it with ultra-caps.
http://www.youtube.com/watch?v=tjxldE8JsDw (http://www.youtube.com/watch?v=tjxldE8JsDw)

I was active in this forum where there was discussion on the "RF diode charger" and etc.
http://laserhacker.com/forum/index.php?topic=104.0 (http://www.laserhacker.com)

I noticed it's very difficult to find antique items in my country like 1N34A that i have to order online from ebay.
I had problems even finding small copper tubes normally used in Air-con smaller version for copper tube\toroid transformer to experiment on resonance as i mentioned earlier in this forum.

Hi gyulasum,

Thanks for providing links to copper pipes/tubes on ebay,this will be the last resort as the shipment cost base on weight of copper would be  high.
I will still continue with search for copper tubes in my tiny island-Singapore -50km(~35miles) wide.
There is potentially some businesses that sell copper tubes further away from my place.

It's frustating after buying toroids online base on specs as shown in this original video -double amazing joule thief 150leds on 1 AA battery,but simple copper tubes can't be obtained easily.
https://www.youtube.com/watch?v=blVqaiYv_WU (https://www.youtube.com/watch?v=blVqaiYv_WU)

I had already purchased  zero threshold ALD MOSFEts(QUAD) for culter radio noticed in document which author which mentioned produces 0.125volts at output.
Even it fails to provide sufficient power to 25mv joule thief maybe on "interval basis" from charged capacitor,it would be used as radio as plan B.

I agree: much easier to wind.  Much more expensive, too, if you can get one the right size.  I've found a few in power supplies, but they usually require a lot of work to get them ready for winding.  If you're lucky you might find one that has the feedback winding already on the core.  Maybe with 200 or more turns on it.  Then it's easy to wind another few turns on the core for the primary.  I think if I want to spend the money on a core, I should just buy one that's already wound and ready for service.

Regarding taps...  I made it a lot easier to wind the toroid by just winding 50 turns of a manageable length of wire on at a time, then bringing it out and connecting it to another manageable length of wire, and on and on until I had 200 turns.  That's how it ended up in this picture (click on it multiple times) (http://rustybolt.info/wordpress/?p=5881).

Quote from: gyulasun on January 27, 2013, 06:08:49 AM
Hi Conrad,

Thanks and may I suggest to make taps on the secondary coil at  1:50 and 1:75  turns ratio to make matching more flexible. Also, you could consider using ferrite pot cores or normal ferrite E cores instead of the toroid core shape, these latter cores are much easier to wind.

rgds, Gyula

@acmefixer: I looked at your blog http://rustybolt.info/wordpress/?p=5912 . According to my tests the 4:100 ratio on a toroid seems to be the best for the 2SK170 transistor (or for 4 of them in parallel). I even got the impression that 8:100 was slightly better (brighter LED at 50 mV, and for sure at 100 mV).

The coils are not the problem for even lower Voltages, it is the transistor. And one would have to go for a two stage circuit, if the power is there, which needs to be 10 mA in order to get 100 µW with 10 mV. About 100 µW is necessary to make a LED glow, starting with 1 mW one can claim that the LED shines (and one has to provide at least 2.2 Volt).

For me that concludes my work on this circuit for the moment. In Spring I will test several earth batteries with magpwr's circuit cast in resin in order to make the circuit weather and water proof. I want some glowing LEDs in my garden.

The earth batteries I have in mind (two different metals like copper and aluminium or galvanised steel) give at least 200 mV, the problem is Amperage. There are a few threads about earth batteries and I have to go through them to recall the ideas put forward. I do not want to use magnesium (too expensive and it corrodes pretty fast).

In case a reader knows a fairly good earth battery by heart, I would appreciate some hints. Magpwr's circuit opens up new possibilities because 50 mV would be enough (in case 2 mA can be drawn).

At the moment I have two 0.5 meter tubes (10 mm diameter) in the garden, straight down next to each other (copper and aluminium). They show 320 mV and 20 µA short circuit current. This is way too little power for a LED. I will try a copper pipe and a galvanised steel pipe (which I have yet to buy).

I will also try a good earth connection (2 meter long copper pipe straight down) and an antenna wire thrown over my wooden garden shed (or even wound around the shed at roof hight). In the Joule Thief thread there was some discussion about "ambient energy receivers", see http://www.overunity.com/6123/jule-thief/15975/#.UQqrGmckQfw, which I will test. Magpwr's circuit might do the trick to light one LED with this power source.

Greetings, Conrad

Hi Folks,

Some relatively cheap sources of ferrite pot cores or pre-wound coils:

http://www.surplussales.com/inductors/FerPotC/FerPotC-1.html 

http://www.electronicsurplus.com/LeftNav/Inductors/Ferrite_ToroidCores.cat 

Of course, ebay can also be a good source but you have to "hunt" for cheaper sellers.

Gyula

My earth battery Joule Thief is now working:

It needed the ALD110900 (instead of the 2SK170) to make the circuit run on a very simple earth battery. The battery consists of a copper tube next to an aluminium tube driven into the ground (both tubes only 0.5 meters long, diameter 10 mm, about 5 cm distance in the ground). I also put a 1 meter copper pipe and an 0.7 meter galvanised post into the ground, but that was an overkill.

The LED is of course not very bright, but one can see it clearly from at least 10 meters away (in the dark).

See the attached drawing and photos. Note, it does not work with the 2SK170 because it draws too much power, which can not be sustained by the simple earth battery. Even the long tubes are not sufficient.

I will now put this circuit into a casting resin to make it weather proof.

Greetings, Conrad

Deleted, mistake!

Hi conradelektro,

Interesting oscillator circuit using ALD mosfet.

I was curious about the efficiency for the ALD based circuit.
Using :200mV x 250uA (Found in uploaded image)

I used VoltsxI(Amps) =0.2v x 0.00025A =0.00005 watts or 50mW (Current consumption to light led)

I then try to apply for my 2SK170 circuit VxI =60mV x4.2mA=0.00000252 watts or 2.52mW(Current consumption to light led)

But i noticed my circuit current draw is higher for lower voltage requirement although efficiency wise the 2SK170 base circuit looks better.
Maybe it make sense why ALD mosfet is preferred choice for  radio.

Quote from: magpwr on February 06, 2013, 08:10:07 AM
Hi conradelektro,

Interesting oscillator circuit using ALD mosfet.

I was curious about the efficiency for the ALD based circuit.
Using :200mV x 250uA (Found in uploaded image)

I used VoltsxI(Amps) =0.2v x 0.00025A =0.00005 watts or 50mW (Current consumption to light led)

I then try to apply for my 2SK170 circuit VxI =60mV x4.2mA=0.00000252 watts or 2.52mW(Current consumption to light led)

But i noticed my circuit current draw is higher for lower voltage requirement although efficiency wise the 2SK170 base circuit looks better.
Maybe it make sense why ALD mosfet is preferred choice for  radio.

These are my calculations:

circuit with 2SK170: 60 mV and 4.2 mA --> 0.06 x 0.0042 = 0,000252 = 252µW (LED is fairly bright)

circuit with ALD110900: 200 mV and 250µA --> 0,2 x 0.000250 = 0,00005 = 50 µW (LED just glows)

My simple earth battery delivers about 1.2 mA short circuit current, which would be too little for the 2SK170 circuit, even at 60 mV.

I just measured the ALD110900 circuit at my earth battery again (0.5 meter copper and aluminium tube next to each other in the ground): 0.53 Volt and 540 µA --> 280 µW.

It just happens that the ALD110900 circuit is perfect for my simple earth battery because the earth battery delivers up to 0.9 Volt but very little current.

For other applications, like a Peltier Element, the 2SK170 circuit is better, because a Peltier Element delivers quite some current even while its Voltage stays low (below 100 mV). At 200 mV the 2SK170 circuit draws already too much power (about 14 mA).

Here my measurements of the two circuits:

2SK170 circuit:  50 mV -- 2.2 mA , 100 mV -- 6 mA , 200 mV -- 14 mA , 500 mV -- 34 mA (I would say, above 200 mV the circuit is not useful any more.)

ALD110900 circuit: 180 mV -- 200 µA , 200 mV -- 250 µA , 1 V -- 1.2 mA ,  1.5 V  -- 1.25 mA (I would say, between 200 mV and 1.5 V the circuit is useful in case one wants a very limited current draw. Of course, the LED is never very bright.)

Greetings, Conrad

hi everyone,

My experiment using copper tube with 7 toroid wrap around it was a failure using 2SK170 with toroid 4:100 using peltier as power source.

However if i try connecting (copper tubes toroid) in series with same(100:4) toroid at pri 4 turns powered by 1.5 battery instead of peltier it works.I was able to light 10mm led quite brightly if i tap power source from within the copper tubes using a straight 18AWG.The led only lit one way only.

Using 2n4401 or 2n22222 transistor instead of 2SK170 if i'm using 1.5v battery.

Latest finding-Please taken note the copper tube/ferrite combination only works well in "high frequency".

is JT can boost the voltage to 5V-12V with high output current w/c will charge the ultracapacitor quickly using  TEG(1.65V @ 16A)

Now J.L. Naudin does some experiments with this strange "transformer" (many toroids on a copper tube):

http://jnaudin.free.fr/dlenz/indexen.htm (http://jnaudin.free.fr/dlenz/indexen.htm)  (the video shows schematics of the test circuit)

http://jnaudin.free.fr/dlenz/DLE01en.htm (also shows the schematics and construction details)

My own tests just showed that a lot of current is necessary to light a few LEDs; more power than lighting the LEDs directly.

I do not understand what Naudin sees in his experiment? It is an interesting "transformer", but to light a 10 Watt lamp with 80 Watt input is not very efficient.

A dynamic system can easily show less power draw under load as long as the input exceeds the output by a huge factor. The impedance (dynamic resistance) of a system changes when the magnetic properties change. And putting a load on this "transformer" will change its magnetic properties.

But there might be something I do not understand, could someone explain the expectations people have in connection with this "transformer"?

Greetings, Conrad

Hi Conrad,

I just seen these heat pads that work off 5vdc, 600ma.

5x10cm
https://www.sparkfun.com/products/11288

5x15cm
https://www.sparkfun.com/products/11289

Video:
http://youtu.be/WdNAQIR5iSI?t=10s

Wonder if you could insulate/sandwich it with a high R value on one side of the peltier cell and pulse it with 5-12vdc to maintain the temp.

Well, back to shoveling, got about 33" of snow.

The shed failed, :o at least the furnace didn't get damaged.

Yikes DTF!    :o   Glad you are okay and hopefully still have power there.   It's times like this everyone could really benefit having a good independent-of-the-grid energy source. 

@DreamThinkBuild: very nice picture of the snow chaos in your garden, hope you will be able to clean up without too much trouble.

The heating elements are very nice and interesting, because they seem so sturdy and easy to use.

But I would not heat up (or cool down) a Peltier Element by help of electricity because the efficiency of a Peltier Element lies between 4% and 7%. So, one wants to use heat (or cold) which is just there for free like from the sun or the warmth from a combustion engine or the cold from the snow.

Greetings, Conrad

I made a mini version  :)

http://www.youtube.com/watch?v=wbfd0k8FzmE (http://www.youtube.com/watch?v=wbfd0k8FzmE)

Quote from: scratchrobot on February 11, 2013, 04:15:27 PM
I made a mini version  :)

http://www.youtube.com/watch?v=wbfd0k8FzmE (http://www.youtube.com/watch?v=wbfd0k8FzmE)

@scratchrobot: very nice build.

How long does it take till the cold side of the Peltier Element (which is not touching your body) becomes as warm as your body and the LED stops glowing?

The experiments I did (either with body warmth or with heat from an oven or with heat from candles) showed that cooling is the big problem. It is very difficult to keep the cold side of the Peltier Element cold (after a few minutes).

One would have to pump a lot of cold water passed the cold side to transport the heat away. In real technical solutions one sees very big heat exchangers.

Greetings, Conrad

Quote from: conradelektro on February 11, 2013, 04:39:16 PM
@scratchrobot: very nice build.

How long does it take till the cold side of the Peltier Element (which is not touching your body) becomes as warm as your body and the LED stops glowing?

The experiments I did (either with body warmth or with heat from an oven or with heat from candles) showed that cooling is the big problem. It is very difficult to keep the cold side of the Peltier Element cold (after a few minutes).

One would have to pump a lot of cold water passed the cold side to transport the heat away. In real technical solutions one sees very big heat exchangers.

Greetings, Conrad

It runs only for about a minute or so, i also tried the other side of the peltier against the window to cool it and get about the same runtime,
Maybe if i keep one side warm with my body and cool the other side against the window it keeps running  :)

Greetings

Hi scratchrobot,

Nice work.

That was the main intention for creating this circuit.
If you mounted small chipset heatsink which cover the peltier surface area and strap peltier around wrist using a sport band maybe with only small heatsink eg:used in chipset cooling.Then as we move(Our body generate even more heat) the small  heatsink sticking out of sport band will dissipate heat nicely and generate power on the go 24/7 (Day or night)with no mechanical wear and tear.I think this unit could last for many years or decades.
It's better to get peltier which have silicon seal at the side to prevent any moisture from entering unit which may degrade performance over time due to corrosion.

Another alternative to mount peltier unit in hat with few thin metal like copper plates in hat stretch in hat like flat finger over head and with rather flat heatsink sticking out at the back of hat or towards the back of hat.This will look more like military self charging unit,maybe in my vision.If  any electronics equipment on hat not in use.The unit will trickle charge batteries 24/7.

I have attached the image of heatsink on peltier which the 2 mounting part sticking out which could be easily removed with a Flat-nose pliers.

http://www.overunity.com/13175/25mv-joule-thief-powered-by-peltier-merely-using-our-body-heat-free-energy-247/dlattach/attach/119721/

I ran this up in Multisim 12 (gives a lot more control for transformers) and it refused to work.  Could you upload the Multisim 10 files for this, Magpwr, as it could be something I did or the new transformer is not like the old one (your transformer was in 10 but now in 12 has been completely removed).

Quote from: Dark Alchemist on September 20, 2013, 06:07:57 PM
http://www.overunity.com/13175/25mv-joule-thief-powered-by-peltier-merely-using-our-body-heat-free-energy-247/dlattach/attach/119721/ (http://www.overunity.com/13175/25mv-joule-thief-powered-by-peltier-merely-using-our-body-heat-free-energy-247/dlattach/attach/119721/)

I ran this up in Multisim 12 (gives a lot more control for transformers) and it refused to work.  Could you upload the Multisim 10 files for this, Magpwr, as it could be something I did or the new transformer is not like the old one (your transformer was in 10 but now in 12 has been completely removed).

hi,

I can only provide you the screen shot and the video.
Please observe the transformer used.It's the only one which works in this tool at the lowest possible voltage. :D

Best place to buy 2SK170 after searching high and low in internet is http://www.futurlec.com/Transistors/2SK170.shtml (http://www.futurlec.com/Transistors/2SK170.shtml)  $0.20 ea   (Please type 2SK170 in search for that site to purchase.It was not listed)
In ebay it priced much higher.

Well, that sucks because, as I said, that transformer has been removed from newer versions of Multisim.  If I can't simulate it myself I will not even attempt to make it so I was hoping you had the circuit I could see what Multisim 12 does with it but if you don't then I will just have to watch the video and move on to something else.

Thanks for your response.

2SK170 is now classified as obsolete so we need a replacement for that part.

Since I couldn't get your transformer in MS 12 due to it being removed by National I moved on from this project but not from the goal.

10 for $5.50 from China:
http://www.ebay.com/itm/10PCS-Transistor-TOSHIBA-TO-92-2SK170-BL-K170-BL-100-Genuine-and-New-/360683877467?pt=LH_DefaultDomain_0&hash=item53fa6f345b

I found some BC337-25s in a US Ebay store, 50 for $3.20, free shipping, so I'm ordering.

Quote from: TinselKoala on October 02, 2013, 02:36:44 AM
10 for $5.50 from China:
http://www.ebay.com/itm/10PCS-Transistor-TOSHIBA-TO-92-2SK170-BL-K170-BL-100-Genuine-and-New-/360683877467?pt=LH_DefaultDomain_0&hash=item53fa6f345b

I found some BC337-25s in a US Ebay store, 50 for $3.20, free shipping, so I'm ordering.

That last part tickles me but the first part is only for now since the main manufacturers have closed down those transistors which sucks.

Quote from: Legalizeshemp420 on October 02, 2013, 03:44:36 AM
That last part tickles me but the first part is only for now since the main manufacturers have closed down those transistors which sucks.

hi Legalizeshemp420,

If you need latest 2SK170 in SMD format SOT-23  .Do search for LSK170 it's the only company which continue the Toshiba J-FET.

Please don't bother or waste your time searching for better ones.I have already tested with over 20 version of J-FET.
As i have previously mention the 2nd best one is 2N3972 (cost lot more around $7 in ebay)which can start osc at 22mV but it's current\load handling is really lousy.

The circuit posted in youtube was my prototype.If you want to improve current handling please connect "low vce drop" 1 npn transistor in parallel with 1x J-JET 2SK170 or LSK170.

Best selection for NPN transistor is ZTX-1047 ,ZTX1048 these if shorted between collector and emitter manually can osc as low as 29mV or 30mV or go for Japanese transistor 2SD1450 45mV which was shown in my older youtube video.

This is the only lowest voltage Joule Thief which exist in youtube and cost less than <$10 or <$5 to make.
If you plan to run on 1XAA (1.5v and below)do change the only resistor to around 22k.

I only reveal the best to public.

The LSK170 datasheet shows a TO-92 and a SOT-23 package but very hard to find a distributor in my one Google search.

Never mind.  Could not get link to work.

Bill

Quote from: Pirate88179 on October 02, 2013, 09:21:35 PM
-line.cn/view_download.php?id=1238977&file=0111\lsk170_1179182.pdf (http://www.overunity.com/-line.cn/view_download.php?id=1238977&file=0111%5Clsk170_1179182.pdf)

Here you can get 5 for $2.03.

Bill

miscopied link?

Quote from: Legalizeshemp420 on October 02, 2013, 09:25:55 PM
miscopied link?

Not miscopied.  I did it three times and somehow, the link I was looking at and copied and pasted brought me back to OU.com.  I have no idea why.

I just googled the part number and it came up on page 2.  I hope that helps.  Every other page I looked at wanted to "give me a quote".

Bill

Quote from: Pirate88179 on October 02, 2013, 09:29:20 PM

Not miscopied.  I did it three times and somehow, the link I was looking at and copied and pasted brought me back to OU.com.  I have no idea why.

I just googled the part number and it came up on page 2.  I hope that helps.  Every other page I looked at wanted to "give me a quote".

Bill

Yep, that is what I saw was just quotes when I Googled.

http://www.ebay.com/sch/i.html?_nkw=2sk%20170&clk_rvr_id=527380270894&adpos=1o3&MT_ID=70&crlp=21242113238_2416792&device=c&geo_id=10232&keyword=2sk+170&crdt=0 (http://www.ebay.com/sch/i.html?_nkw=2sk%20170&clk_rvr_id=527380270894&adpos=1o3&MT_ID=70&crlp=21242113238_2416792&device=c&geo_id=10232&keyword=2sk+170&crdt=0)


Here is a link to ebay where you can get the 2SK170 which is now called the LSK170.  (I hope this link works)

Bill

***EDIT***  It works!  Mouser shows these as discontinued but, the ebay prices are not that bad it seems.

Quote from: Pirate88179 on October 02, 2013, 09:57:33 PM
http://www.ebay.com/sch/i.html?_nkw=2sk%20170&clk_rvr_id=527380270894&adpos=1o3&MT_ID=70&crlp=21242113238_2416792&device=c&geo_id=10232&keyword=2sk+170&crdt=0 (http://www.ebay.com/sch/i.html?_nkw=2sk%20170&clk_rvr_id=527380270894&adpos=1o3&MT_ID=70&crlp=21242113238_2416792&device=c&geo_id=10232&keyword=2sk+170&crdt=0)


Here is a link to ebay where you can get the 2SK170 which is now called the LSK170.  (I hope this link works)

Bill

***EDIT***  It works!  Mouser shows these as discontinued but, the ebay prices are not that bad it seems.

It would seem that SK170 has been discontinued no matter if 2 or L which is a shame.  Eventually it will be like trying to find a Germanium transistor.

For prototyping it is fine.  But, to make an ongoing device that can be improved over time and possibly marketed, it sucks.  I ran into the same thing trying to source those flash circuit inverter transformers.  I have over 100 of them now but...no where to get more except one place that wants $8.00 each in 200 quantities.

I am now working on making my own here in house.  I can do it for about $2.00 each I think.  We will see.

Bill

Quote from: TinselKoala on October 02, 2013, 02:36:44 AM
10 for $5.50 from China:
http://www.ebay.com/itm/10PCS-Transistor-TOSHIBA-TO-92-2SK170-BL-K170-BL-100-Genuine-and-New-/360683877467?pt=LH_DefaultDomain_0&hash=item53fa6f345b (http://www.ebay.com/itm/10PCS-Transistor-TOSHIBA-TO-92-2SK170-BL-K170-BL-100-Genuine-and-New-/360683877467?pt=LH_DefaultDomain_0&hash=item53fa6f345b)

I found some BC337-25s in a US Ebay store, 50 for $3.20, free shipping, so I'm ordering.

TK:

I must have missed your post because that is what I posted about the 2sk170's.  I should read more and post less perhaps.

Bill

Quote from: Pirate88179 on October 02, 2013, 10:19:13 PM
TK:

I must have missed your post because that is what I posted about the 2sk170's.  I should read more and post less perhaps.

Bill

That store says they can supply large quantities of the 2sk170 too. If that's true maybe some JT fanatic should order a couple hundred, just so that they are available for experimenters.

Quote from: Pirate88179 on October 02, 2013, 09:57:33 PM

Here is a link to ebay where you can get the 2SK170 which is now called the LSK170.  (I hope this link works)

2SK170 is also available from Amazon. But, I just checked and there are only 19 matched pairs left in stock.

[size=78%]
[/size]

Quote from: xee2 on October 03, 2013, 11:37:57 PM

2SK170 is also available from Amazon. But, I just checked and there are only 19 matched pairs left in stock.

[size=78%]
[/size]

Heyas, xee.  Love your work on YT just wish you had comments enabled and I have been watching them for about 2 years now.

Nice to know Amazon has them but they are dieing out quickly. :(

Why is it when something wonderful works it ceases to be made when nothing has taken its place?

@Magpwr
Can i get a schematic of this circuit by using another transistor that is available in the market (obv not 2sk170)

Quote from: omdano on January 11, 2014, 03:52:59 AM
@Magpwr
Can i get a schematic of this circuit by using another transistor that is available in the market (obv not 2sk170)

Linear Technology also sells complete energy harvesting integrated circuits that can work down to very low voltage and low power inputs.

Quote from: MarkE on January 11, 2014, 04:12:14 AM
Linear Technology also sells complete energy harvesting integrated circuits that can work down to very low voltage and low power inputs.

Sadly i dont have time to order from the internet  ... i only can build my own circuits in the meantime

Quote from: omdano on January 11, 2014, 05:35:39 AM
Sadly i dont have time to order from the internet  ... i only can build my own circuits in the meantime

hi omdano,

I'm sorry to say after testing more than 30 fets.The best fet is still 2SK170 and there are couple of successful replication of my ultra lowest voltage self oscillating joule Thief to validate my claims and it cost <$5 or <$3 to make.
If you say you don't have time to order fets from the internet-God bless you. ;D

I still do hold the personnel record lowest operating voltage 25mV joule thief on the cheap in youtube. :D :D :D
If you want to modify the circuit to run on 1.5v battery simply change the resistor from 1.8k to 22k.

The circuit diagram was already revealed in video on a piece of paper on the floor.It's free.

Oh sorry i forgot to say that my peltier produces about 100mV and i am going to use 2 to produce 0.2
i want a transistor for a 0.2 JT XD , can you help me with that one :D

thanks
God bless u 2 :D

Quote from: omdano on January 11, 2014, 07:34:50 AM
Oh sorry i forgot to say that my peltier produces about 100mV and i am going to use 2 to produce 0.2
i want a transistor for a 0.2 JT XD , can you help me with that one :D

thanks
God bless u 2 :D

hi omdano,
I do have older video 45mV (0.045volts) which was using Japanese transistor.Do refer to my youtube "sanjev21".I think it self run on 0.2volts,it's been a long time.That video is considered obsolete now.

Based on magpwr's work, the circuit on the attached drawing is the best I could do. It was already published in this thread way up. (As I said, it is magpwr's circuit, I just replicated.)

On the drawing I state the power consumption at several supply Voltages starting at 50 mV. Putting 500 mV into the circuit is already too much.

E.g. at 50 mV: if the Peltier element can not deliver 2.2 mA, it will not work (Voltage will drop below 50 mV).

If you provide a higher supply Voltage (e.g. 200 mV) the Voltage will drop till at least 2.2 mA are delivered.

The 2SK170 seems to be the only transistor easily available which can do that. Best to use 4 in parallel because the transistor has a resistance of ~200 Ohm (between drain and source) when open. Look up the specs of the 2SK170, it is an amazing transistor for very low Voltage and low power applications.

Your 200 mV (even the 100 mV from one of your Peltier elements) is plenty, as long as at least 2.2 mA are provided. I had to buy a 50 mm x 50 mm Peltier element, the smaller area Peltier elements delivered too little current when touched. (If you fire the Peltier element up with a flame and provide good cooling on the other side you get Amperes. The low temperature gradient between a human hand and ambient air needs a big Peltier area to give useful current.)

You might know that Peltier elements are not efficient, they only convert about 4% of the heat energy into electricity. Therefore you need a 200° temperature gradient, lots of heat and good cooling to get Ampere out of it. Whatever heat you can provide from human skin (best the warm hand) will result in very very little electricity. The 2SK170 is that very rare transistor which can still operate on the low Voltage (down to 50 mV) and the low Amperage (down to 2.2 mA).

On the photo you see a pan underneath my 50 mm x 50 mm Peltier element. Before I took the photo I put the pan outside in winter to cool it down to o°. This gave a temperature gradient of 30° (hand to 0° pan) for a minute and the LED was bright (and my hand was cold).

If you want to harvest heat from a human body you have to cover the whole body with Peltier elements and soon the person would shiver (if the outside is cold) or would heat up (if the outside is warmer).

Lately I bought this dynamo flashlight http://www.werbeartikel-discount.com/product_info.php/products_id/30608 and it beats everything I ever had in the field of alternative electricity production. An age old technology made useful with modern components. The human hand can provide a lot of mechanical energy without you even noticing it. I became a believer in the hand dynamo or the home trainer dynamo (which can deliver 40 Watt). Very simple technology which will not break easily.

Greetings, Conrad

Futurlec has them for 20 cents apiece, 16 dollars US for a hundred.

http://www.futurlec.com/Transistors/2SK170pr.shtml

Which of the sellers have the fastest delivery service ? i want to get the 2SK170 in 17 days , thats what i am worried about (BTW i am middle eastern living in Dubai) the problem is the Shipment time always .

Thanks for all the help  ;D

I don't know what customs delays are like where you are.  These are commonly available commodity transistors.   If they are not available on the Peninsula then you should be able to obtain them without difficulty and a few days postal delay from:  Europe, India, or Asia.  A roughly equivalent part is an NTE132.  Or you can order those from Radio Shack on the web.

Quote from: MarkE on January 12, 2014, 04:28:44 AM
I don't know what customs delays are like where you are.  These are commonly available commodity transistors.   If they are not available on the Peninsula then you should be able to obtain them without difficulty and a few days postal delay from:  Europe, India, or Asia.  A roughly equivalent part is an NTE132.  Or you can order those from Radio Shack on the web.

Nah we have lots of shops here , but i thought since it was discontinued .. meh ok thx

@Sanjev21

So..
http://planetbuddy.co.uk/joule-thief-very-very-low-input-0136-volt/
I saw you there , do you confirm that circuit ?
and since we are using BC547 cant we replace it with 2N2222 ? they are almost the same

Thanks

And one thing ... There are some elements that remain cold because of their high heat resistance that can be thermal pasted to the peltier cold side to ensure a cold side in all situations  , thats what i am trying to do in my project

Quote from: omdano on January 28, 2014, 02:27:05 PM
@Sanjev21

So..
http://planetbuddy.co.uk/joule-thief-very-very-low-input-0136-volt/ (http://planetbuddy.co.uk/joule-thief-very-very-low-input-0136-volt/)
I saw you there , do you confirm that circuit ?
and since we are using BC547 cant we replace it with 2N2222 ? they are almost the same

Thanks

hi omdano,

I'm unable to advise which is better unless you start experimenting to know the results.

There is no such thing as permanent hot or cold side base on your second question.

The "energy" for my experiment comes from the food we eat to keep us alive hence our body do generate heat 24 by 7 as long we are alive. :D :D :D

If you live in the wild with no access to grid power or solar power at night this device can help you find your way in the dark eg:miners(dark in the mines) ,soldiers and etc.

I 've got my 2sk170 , Thx magpwr for this wonderful circuit , thanks to u guys  too for helping

Quote from: omdano on February 23, 2014, 10:18:45 AM
I 've got my 2sk170 , Thx magpwr for this wonderful circuit , thanks to u guys  too for helping

hi omdano,

Don't forget to post a photo of your completed circuit  with led lit once you get it running. :)

Cheers.

I am going to make you more than that , Just wait for april .

Edit :

One question
if i had 2 peltiers , each producing 100mV , i want to put them in series to produce 200mV
won't the second peltier just suck up the voltage and turn it into heat ?

Quote from: omdano on February 23, 2014, 12:01:02 PM
I am going to make you more than that , Just wait for april .

Edit :

One question
if i had 2 peltiers , each producing 100mV , i want to put them in series to produce 200mV
won't the second peltier just suck up the voltage and turn it into heat ?

My opinion:

It is fine to put Peltier elements in series (in order to have higher Voltage). But the current can not be more than the current from the weaker element. The elements will give the same current if they have the same temperature and the same temperature gradient (between the hot and cold side) and if they have the same size and make.

If one puts Peltier elements in parallel (in order to have more current) they should have roughly the same temperature and the same temperature gradient  (between the hot and cold side), also same size and same make.

Simply said: if you use more than one Peltier element (either parallel or in series) make sure they have the same conditions on their hot and on their cold side (of course the hot side is hotter than the cold side). This might not be easy to achieve. A smaller Peltier element will heat up faster than a bigger element. Different brands will have a different heat conduction.

If you want to keep it simple, use just one element. Peltier elements are available in different sizes (it soon becomes expensive, 50 x 50 millimetre elements are already expensive). I found that 30 x 30 mm or smaller Peltier elements are pretty useless at low temperature differences.

Peltier elements become usefull if a temperature difference of 200° (centigrades) is available, which implies very good cooling on the cold side. Getting the heat on the hot side is fairly easy (open flame), but the cooling on the cold side is difficult (in practical terms). Usually some water cooling is arranged.

Peltier elements are ususally used as sensors (providing very little power) and not as a generator (providing useful power).

Peltier elements are in general not a very useful power generator technology on earth. But in space the sing, because the cold side is simply exposed to space (near absolut zero) and on the hot side is usually a nuclear reaction providing heat for years. You do not want a nuclear reaction near you and you do not have absolut zero near you.

Greetings, Conrad

Yes, you should be able to put Peltiers in series, just like any other DC source.

(oops, just saw Conrad's post, same thing)

So... did anybody try to Oscilloscope the output of this circuit ? (with any voltage input - please mention)

Well, I don't have a 2sk170 but I've certainly posted plenty of JT scope traces. And LTseung has a thread that is mostly discussion of scopetraces from his JTs.

Quote from: TinselKoala on February 24, 2014, 03:34:03 PM
Well, I don't have a 2sk170 but I've certainly posted plenty of JT scope traces. And LTseung has a thread that is mostly discussion of scopetraces from his JTs.

Yeah because .. i would love write an equation for the JT circuits inputs and outputs
to make things simpler .

Good luck!

Actually the JT is a type of blocking oscillator and these have had equations written for them in plenty.

Here you go, calculate away to your heart's content.

http://www3.alcatel-lucent.com/bstj/vol38-1959/articles/bstj38-3-785.pdf (http://www3.alcatel-lucent.com/bstj/vol38-1959/articles/bstj38-3-785.pdf)

Here's a good overview of transistor blocking oscillator operation, but without the equations:

http://mysite.du.edu/~etuttle/electron/elect37.htm (http://mysite.du.edu/~etuttle/electron/elect37.htm)

So the JT is really a blocking oscillator with components chosen so that it will operate at low input voltages.

is there an easier way for winding the 100 windings ? XD

Are you winding toroids? Yes, there is.

http://www.youtube.com/watch?v=ipfYPbNl_rc

Well ..  so..  How did u measure the voltage in the video ?
People usually use Oscilloscope but you just simply used a multimeter
Where did u connect the multimeter ?
(the output circuit voltage)

Quote from: omdano on March 01, 2014, 01:52:47 AM
Well ..  so..  How did u measure the voltage in the video ?
People usually use Oscilloscope but you just simply used a multimeter
Where did u connect the multimeter ?
(the output circuit voltage)

Did I link to the wrong video? Sorry, the video I intended to link to was just showing how to use a spindle bobbin to wind the toroid easily.

What video did you actually watch? There was no multimeter shown measuring anything in the video I "thought" I linked to -- and which still appears to be the one in the link.

If you want to see some measurements on a scope, why didn't you say so? You asked about how to make it easier to wind 100 turns on a toroid, I thought.

Here are some scope measurements of  JTs.
http://www.youtube.com/watch?v=6zFOhTU-bt0 (http://www.youtube.com/watch?v=6zFOhTU-bt0)
(For some reason I say "50 milli Ohms" when I meant to say "50 milliAMPs" a couple of times when talking about the Current trace on the scope. Sorry about the confusion...)
http://www.youtube.com/watch?v=nqJPi35cRrE (http://www.youtube.com/watch?v=nqJPi35cRrE)
http://www.youtube.com/watch?v=uLEyv2jbWrU (http://www.youtube.com/watch?v=uLEyv2jbWrU)
http://www.youtube.com/watch?v=9MxIXESkS3I (http://www.youtube.com/watch?v=9MxIXESkS3I)
http://www.youtube.com/watch?v=KWDfrzBIxoQ (http://www.youtube.com/watch?v=KWDfrzBIxoQ)

Quote from: omdano on February 24, 2014, 01:38:23 PM
So... did anybody try to Oscilloscope the output of this circuit ? (with any voltage input - please mention)

http://www.youtube.com/watch?v=dR3wQf7xqNM (http://www.youtube.com/watch?v=dR3wQf7xqNM)

lol
@TinselKoala
i should have tagged magpwr

i was talking about his main video in the OP

Quote from: omdano on March 01, 2014, 01:52:47 AM
Well ..  so..  How did u measure the voltage in the video ?
People usually use Oscilloscope but you just simply used a multimeter
Where did u connect the multimeter ?
(the output circuit voltage)

hi omdano,

There is no need for Oscilloscope since we are working with DC.

You will need to the "fix" hot or side for the peltier(which effect which wire from peltier produce + or -) since you are going to connect to 100uf or 1000uf capacitor at the input of joule thief(+ -).

The peltier (+ -) output is connected to multimeter and joule thief output is connected to multimeter in video.This is just for demonstration purpose.

This joule thief circuit can be completed in few hours since 100 winding for toroid will take some time."Need alot of patience" :D

That's all the advise i can give."Good Luck"

In my video, using the spindle bobbin I put 20 turns on a toroid in a couple of minutes, so you can do 100 turns in ten or fifteen minutes. How long is that?

Hello magpwr

i have built the whole circuit and it didnt work with a peltier ..
i tied to test it with a AA battery and it worked.. what do you think might be the problem in my case ?

(i didnt put the 100uf cause it said optional)

Edit : it worked but the led didnt look like it was producing any light , i will try to change the gate resistance

@ omdano


The Peltier may not produce enough current. Can you try circuit with a power supply? It probably takes a few mA to light the LED at 0.2 volts.

I added another peltier to produce more current so now its working fine :p thx

Quote from: omdano on March 11, 2014, 06:46:04 AM
I added another peltier to produce more current so now its working fine :p thx

:)  Congratulations. Can you post a video so others can see how you did it?

TinselKoala did some tests with the 2SK170 Joule Thief circuit, see http://www.youtube.com/watch?v=OgndINJHnd0 (http://www.youtube.com/watch?v=OgndINJHnd0) and http://www.youtube.com/watch?v=35IMWNp2akg (http://www.youtube.com/watch?v=35IMWNp2akg)

I put the circuit together again and did some scope measurements:

- the power supply was set to about 100 mV (I used only one 2SK170, if one puts 2 or 4 in parallel the supply Voltage can go down to 50 mV)

- note that the 2SK170 has a high resistance when switched through (at least 100 Ohm), the primary winding is therefore never pulled completly to GND

- if one measures over the "LED and the diode" (scope CH 2, at the base resitor) one sees that the LED is switched off and on

- if one measures only over the LED (scope CH 2a) a "ripple current" seems to show up, but this is an artifact because the diode is blocking the scope probe from going more positive

- the supply Voltage should not be higher than 500 mV, the current through the transistor (drain to source) would be too much

- the circuit swings at about 30 KHz, but a scope probe (indicated as CH 2 in the circuit) at the base resistor changes the frequency to about 25 KHz

- if the Toroid has less inductance the frequency will be higher but the supply Voltage has to be higher too (may be 200 mV)

Greetings, Conrad

Yes, exactly. That 2a trace is the one. You say this is an artifact caused by the diode, but surely the scope probe is seeing the same thing the LED is seeing on that side of the diode. It looks to me like the LED must be just barely flickering, around its fwd voltage point, dimming and brightening in time to the "ripples" on that signal. I haven't had the time to explore it with a photoresistor to see what the light from the LED is actually doing yet, I got distracted by another project.

You've motivated me to rewind that tiny ferrite bead, though, to see if I can get a lower operating voltage. I think I can get 100 turns onto it, but the wire keeps breaking (it's scrounged wire from a TV yoke).

Since the video, I've put a 1n5711 Schottky diode in place of the 1n34a. It got me a few more millivolts lower but not much. Down at the bottom of this page there is some interesting info on detector diodes.
http://www.techlib.com/electronics/crystal.html

Finally I have found time to assemble suggested circuit from this thread
https://www.youtube.com/watch?v=ozQwneBTRB0

Start up operation voltage 24-25mv with coil from Coilcraft 1:20
Vin ~ 0.076V -> Vout ~ 3.12V
Iin   ~ 1.6mA -> Iout  ~ 0.14ma
Possible to start from 14mv with 1:50, but Vout will be only ~1.4V
Resistance of secondary coil in 1:50 twice bigger than in 1:20
By tuning resistor it is possible to get higher output voltage in addition around 100-200mv, which is enough to compensate hysteresis from second circuit

Second circuit contain Voltage detector BD4930. Due to internal hysteresis Vdet*0.05, LED blinks with 100% power only at V=3.12V. And Tantalum capacitors 470uF and 4.7uF
and I have to replace diode (from graphic card, protective diodes), otherwise I lose 0.7V on it.

What do you think about these transistors, some of them complete analogue of SK170, another have twice bigger conductivity
CPH6904-TL-E     http://de.rs-online.com/web/p/transistoren-jfet/7919403/ (http://de.rs-online.com/web/p/transistoren-jfet/7919403/)
2SK2394-7-TB-E http://de.rs-online.com/web/p/transistoren-jfet/7925155/ (http://de.rs-online.com/web/p/transistoren-jfet/7925155/)
2SK3557-7-TB-E http://de.rs-online.com/web/p/transistoren-jfet/7925167/ (http://de.rs-online.com/web/p/transistoren-jfet/7925167/)
CPH5905G-TL-E  http://de.rs-online.com/web/p/transistoren-jfet/7925278/ (http://de.rs-online.com/web/p/transistoren-jfet/7925278/)
TF252-4-TL-H      http://de.rs-online.com/web/p/transistoren-jfet/7919594/ (http://de.rs-online.com/web/p/transistoren-jfet/7919594/)

Can I study with this brightness of light?

Hi
I am making a thermoelectric flashlight which works completely from the heat of our hand.I am using a peltier module(TEC1-12706) which is cooled by water.So,can anyone suggest an efficient,working circuit for me please.

hi everyone,

I have just release one of my sidetracked project and the latest prototype in my youtube channel "sanjev21".
This the first and the only video in youtube release base on DIY project using Body Heat to power up a spinner using a peltier module.

This is also the highest efficiency energy harvesting device which outperform the commercial design. :D
Body Heat to Kinetic -Prototype v2.0 Energy Harvesting device
(https://www.youtube.com/watch?v=qDjvQSrOV6c)
http://www.youtube.com/watch?v=qDjvQSrOV6c (http://www.youtube.com/watch?v=qDjvQSrOV6c)

Please read below to understand what is happening.
-----------------------------------------

After few years of current topic idle it seem certain people are charging high price around $35...$39 base on a prototype design which i have released to public freely merely for replication purpose.
That was just a ver 1.0 design and transformer was merely shrunk using Coilcraft 253P "1:20" SMD transformer.

I have not even bothered to release the improved version 2.0 yet which was already completed few years ago.
It was also a test for me to see if anyone whom is competent enough to improve my released design further.But none seen so far but worst commercialization of my circuit have occured.

http://www.ebay.com/itm/70mV-Joule-Thief-powered-by-peltier-TEG-with-body-heat-green-LED-/222180828198

http://www.ebay.ie/itm/25mV-Joule-Thief-harvesting-powered-by-peltier-TEG-with-body-heat-green-LED-/221935843924

http://www.ebay.com.my/itm/221946116905

You should know who you are.

My intention was free but someone here ruined it for everyone whom simply enjoy circuit building.

QuoteSo,can anyone suggest an efficient,working circuit for me please

Ann Makosinski (http://www.cbc.ca/news/technology/body-heat-powered-flashlight-takes-teen-to-google-science-fair-1.1317745) did use (https://www.googlesciencefair.com/en/projects/ahJzfnNjaWVuY2VmYWlyLTIwMTJyRAsSC1Byb2plY3RTaXRlIjNhaEp6Zm5OamFXVnVZMlZtWVdseUxUSXdNVEp5RUFzU0IxQnliMnBsWTNRWXA2ZVVBZ3cM) power converter integrated circuit, LTC3108 (http://cds.linear.com/docs/en/datasheet/3108fc.pdf) The circuit contained FET's that would oscillate at voltages as low as 20mV. When used with a recommended   transformer, the IC would provide well over 2.5 volts AC. Such a research here is both reinvention of wheel, both off-topic with respect to overunity.

Quote from: Zephir on April 16, 2017, 02:07:44 PM
Ann Makosinski (http://www.cbc.ca/news/technology/body-heat-powered-flashlight-takes-teen-to-google-science-fair-1.1317745) did use (https://www.googlesciencefair.com/en/projects/ahJzfnNjaWVuY2VmYWlyLTIwMTJyRAsSC1Byb2plY3RTaXRlIjNhaEp6Zm5OamFXVnVZMlZtWVdseUxUSXdNVEp5RUFzU0IxQnliMnBsWTNRWXA2ZVVBZ3cM) power converter integrated circuit, LTC3108 (http://cds.linear.com/docs/en/datasheet/3108fc.pdf) The circuit contained FET's that would oscillate at voltages as low as 20mV. When used with a recommended   transformer, the IC would provide well over 2.5 volts AC. Such a research here is both reinvention of wheel, both off-topic with respect to overunity.

hi Zephir,

I am using only discrete components and i did mention it produce much higher current output compared to a current commercial unit in my youtube video description.
At 6°C difference i am getting 210uA short circuit current at circuit output.I was trying to prove that my unit is much better than the tested commerical unit which got certain plus points but not in current output.

One of the commercial unit merely deliver 15uA compared that against mine 210uA.There nothing to talk about voltage output around 4volts which is the easy part.Think about it.

Energy given by body as heat is still considered "free energy" where we take "wasted radiated heat" given off from our body 24/7 (If alive) and put it to work.
It's a proof of concept but not necessary efficient.

I am a one man team.

Nothing much to add.

Side story
---------------------------------------------
Electrical power for the New Horizons mission to Pluto is furnished by a single radioisotope thermoelectric generator (RTG), which transforms the heat from the natural radioactive decay of plutonium dioxide into electricity

QuoteAt 6°C difference i am getting 210uA short circuit current at circuit output

This is still too low - Ann Makosinski (a grade 10 student) got for 50 mV DC from the Peltiers about 6 mA at 5 Volts AC which was sufficient to light the LED. The result was a bright light at 5 degree Celcius of Peltier differential. Yours 210 uA is still at least ten-times worser result. On the web you can find Joule-Thief circuits (https://www.youtube.com/results?search_query=joule+thief+very+low) with silicone (or preferably germanium) transistor in common base configuration, which is capable to oscillate at ~ 30 mV only.

I feel sorry that your design has been stolen - but I'm afraid, it will not work anyway so effectively like the professionally designed integrated circuit mentioned above. Environmental or waste energy recycling isn't overunity, which this forum is all about. Your device still consumes an energy and given the amount of electricity produced the crank dynamo powered battery would be way cheaper solution. Try to replace the lithium battery in ultracheapo crank dynamo flashlight (http://www.banggood.com/3-LED-Dynamo-Wind-Up-Hand-pressing-Crank-Flashlight-p-955183.html) with supercapacitor - and I can assure you, you'll be pleasantly surprised with result.

Quote from: Zephir on April 16, 2017, 07:51:39 PM
This is still too low - Ann Makosinski (a grade 10 student) got for 50 mV DC from the Peltiers about 6 mA at 5 Volts AC which was sufficient to light the LED. The result was a bright light at 5 degree Celcius of Peltier differential. Yours 210 uA is still at least ten-times worser result. On the web you can find Joule-Thief circuits (https://www.youtube.com/results?search_query=joule+thief+very+low) with silicone (or preferably germanium) transistor in common base configuration, which is capable to oscillate at ~ 30 mV only.

I feel sorry that your design has been stolen - but I'm afraid, it will not work anyway so effectively like the professionally designed integrated circuit mentioned above. Environmental or waste energy recycling isn't overunity, which this forum is all about. Your device still consumes an energy and given the amount of electricity produced the crank dynamo powered battery would be way cheaper solution. Try to replace the lithium battery in ultracheapo crank dynamo flashlight (http://www.banggood.com/3-LED-Dynamo-Wind-Up-Hand-pressing-Crank-Flashlight-p-955183.html) with supercapacitor - and I can assure you, you'll be pleasantly surprised with result.

hi Zephir,

Since you wanted to debate with me i shall prove 2 things since you wanted to compare me with the little girl.

Firstly the video of mine was released for members in overunity  to replicate on 20th Dec 2012 -Ultra lowest 25mV self starting Joule Thief ver1.0-Free Energy.
-Using single peltier module

The girl video was released on 30 Apr 2013 - The Hollow (Thermoelectric) Flashlight - Google Science Fair
-Using 3 peltier module.


The commerical version "allow short burst in mA" for RF transmission purpose.

I do challenge you or if you can get anyone to try using LTC3108 to "immediately power up the display spinner" at 6°C difference and not in burst after few seconds charging capacitor. ;D ;D ;D

Since i already mentioned mine is better than the commerical unit in output.But Commercial unit got more feature like "supervisory circuit built in".

Yes peltier is not a efficient way to generate electrcity only this i agree from the start.

Yes my device is able to self oscillate at 15mV but i am interested in more power output.

------------------------------
Think about it if every kids use the factory design components then where is the learning curve.No basic foundation is that what you wanted for the future. :D

The following schematic is of an energy harvester which works at only 20mV.
I traced the schematic out from the PCB and measured the component values.

And here's another schematic which I traced out from the circuit board:

Quote from: ZeroDivide on January 27, 2019, 12:49:51 AM
And here's another schematic which I traced out from the circuit board:

hi Zerodivide,
I am curious how do you obtain the part number from the J-fet smd.Is there any site you are using to obtain the j-fet part SMD number.