Joule Thief

[Hoppy]

New video update here: http://youtu.be/SdXAjT2QPiQ
In this video I also show the dielectric absorption effect that can be very deceiving.

Forum Discussion at:  http://laserhacker.com/forum/index.php?topic=70.0

Updated schematic coming soon at: http://laserhacker.com/

Excellent video. I'ts all to easy to inadvertently put a heavy load on the battery and get this charging effect. Another way is to have a low battery in the meter. Its good policy to spend out on a good quality meter and not use cheap meters.

I'm currently running your crossover circuit with Gadget's mod and monitoring running voltage to 4 decimal places with a good quality Tektronix TX3 DMM. Its amazing how slowly the fourth digit is decrementing! This circuit will run for many hours without the second digit changing. I have a fairly bright LED as a load and am using an 'E'-core ferrite core for the transformer. I did think at one point that I had a free runner as my voltage on my nearly 'flat' NimH AA cell climbed for over an hour before stabilising. I realised after that I had previously accidentally shorted the battery and had not waited for the big 10,000 cap to stabilise before starting the run.

Another thing to watch is that this circuit can vary in frequency just by touching various points and by applying an earth. This has the effect of the battery appearing to charge when the frequency drops. Its not easy to realise that this is happening if the frequency is high and the reduced frequency does not cause the LED to visually flash or the frequency cannot be heard.

keep up the good work.

Regards
Hoppy

[gadgetmall]

Xee2 ? What are your suggestions at to cap values ? So far i have worn out battery.cap combinations . i had 10 of the big 10000uf 25volt caps so i used them . .


Others . I know what i see . I realize that thru the years i have learned the do's and donts . I also am fully aware of  cap and battery recovery and having new batteries in my meters which i do . I saw recovery before this mod not after .

i guess what i have posted will be conscrewed to death

I need anyone who is experimenting with Ls crossover to use the mods and see for them selves and post their results . I always wondered about Lasersabers joule ringer 1 and NEVER could get it to work and not many did . NOW . i see three different replications i made and they all work better than the LS1 with no hassle at all making them work . ..

And i also have 4 digits and the the last three are rising on all my circuits . If you are dropping you need a different cap between the second germanium . Note that any germanium diode will not work . it has to be one of the two I think it is 1n34a with one stripe . as the other older two strip will not work 1n60 . also note that the LIGHT you use cannot be just an led . It has to have some resistance in it like a 220 led bulb . There you will obtain the Balance in light and charge.

a good circuit will require good tools . My resistor and cap box are those tools for setting this circuit to a balance . once that happens then i can hunt for replacement parts for them .If you have "bright light you will not see this charging take place. forget light when you are setting this up . the light is just an indication that the circuit is running. This is why it is important to run on less that the led will take in volts alone .

[xee2]

Xee2 ? What are your suggestions at to cap values ? So far i have worn out battery.cap combinations . i had 10 of the big 10000uf 25volt caps so i used them . .

Any value mica, ceramic, or film capacitor will not have the bounce back effect. If you can get your circuit to work with a 1 uF ceramic capacitor that will eliminate any chance of capacitor bounce back. Also, electrolytic bounce back is usually less than 2 volts (typically about 0.5 volt) so if you get a voltage gain of 5 or 10 volts with your electrolytic capacitors that would eliminate the chance that it was just bounce back.

[NickZ]

  Gadget:
   I'll add the diode/cap, and let you know what I find. We are all in this together, to help each other out. At least I am, ok.   Great.

  The main thing to remember is that the energy is always coming in the form of a conversion process, and not from the Christmas Spirit, Easter bunny, etz... So, where is this additional energy to charge the battery, or caps coming from. That is the question.   Dielectric absorption.  Probably. 
This is not dependent of the grid leakage, or body capacitance.

  So, as most of us here have also made some home made "crystal cells", why not add those cells to the input cap, both in series, as well as in parallel,  and see what happens.
  I also think that using two 2n2222 or two MPSA06 instead of the higher power but less efficient 2n3055, or TIP3055 type of transistors, will see a higher gain from this type low output circuit.

[gadgetmall]

Ncikz . Agree . When you make the circuit  get a balance going first with the resistance in series with an led or in my case two leds in series and a 33kresistor . But also this resistance needs a balance of the capacitor in series with the  diode . Aim for low light and indicator. from there the circuit is in a jitter state and any external touching will either make it brighter on the pos or dimmer on the neg. you can leave your meter leads on it if you like  but i found that messes up the balance and makes it charge Less . So i keep them disconnected till reading ..this morning was a shocking result over 1.5 volts and the nicad was only setting ar 1.228 two days ago.Nicads don't charge that high but this one held it for a bit until the light got brighter and it lost balance. . i don't even know if a cap is needed yet . Might not need one .

Ok . I just checked the cold unit setting outside . Guess what .   it is 1.659 across the nicad and cap .. not emf . not staticvolts . Usable energy .. Power from itself.Or from the vacuum of space itself . OR i have rediscovered power from the aether as Moray did by time consuming experimenting , I bet he had a cap sub and resistor box too