Lasersaber strikes again. A joule thief king ?

[d3x0r]

Was considering that maybe 'tesla torch' isn't horribly far off... "Ka4ep torch" err "kacher torch" might be better name
(simplified,  start resistor omitted)

second image:
My modifications end up drawing more power... but mine won't run without a cap(not sure where to put the other end; to power rail works; I have a couple other ideas to try) on the top end of the secondary (inducter to base)... and adding a diode on the primary ends up looping that energy around the coil (red circle); and doesn't affect normal run, but if a diode is additionally added from the base to ground(green), then it reduces power used....  instead of a diode the voltage is high enough to add a couple more LEDs...



Edit:
Wound a new coil that's 4 turns and lots of turns... 30uH and 27mH... on tiny ferrite... connected low inductance power->collector; and now drive a larger load at lower voltage (0.7V-1V 30-40mA); 10 5V white LED High and 10 Low.
top is full sinwave, biased a little towards the low side; no resistor.

[d3x0r]

Figured out where to put the cap... between the top of the coil and the load; reduces power consumption slightly
Now base diode doesn't help under-voltage as much; only goes to a square wave of like -5V but does help power consumption slightly.
2n2222 transistor
Running frequency is about 100Khz; and reverse calculating capacitance from frequency and inductance yields about 93pF... which I assume is the internal capacitance of the coil... it's a nice sine wave.  Added a series capacitor to the load; too small of a capacitance fails...
Ground connection to the load can either be the low side of the cap or a real ground, doesn't affect current draw much.
Does not require startup; but I'm using a hand generator which generates slight spikes in voltage which tickles it enough I guess... I did have a 1M resistor to tickle the base but whether it's connected or not doesn't matter.


Circuit running under 2V generates +/-50V... at 2V is about 40mA consumption at 0.6V goes down to about 10-20mA.


https://www.youtube.com/my_videos_annotate?v=2D5HuxmlyjY


Edit:
Was able to add a slight recovery to power from the load... it measures that there should be a voltage change; but not very much... 50mV (x10 probe, 500mV across 100ohm resistor... 5mA + and - approximately... but it's only for a very short time... )  I think some of it is lost being in-phase with the power used... it does add a hair of brightness at no more current usage....

[d3x0r]

So for a while I've been trying to figure why there's a dead space and then a sudden burst of oscillations. (on LS's device)
'ferroresonance'.  once it fires once, with the right load and various other conditions there's a burst of activity where the low voltage on collector side actually re-fires the transistor again with a slight ringdown, until another long pause before the gate voltage goes high again.  These tiny pot cores do have a very tiny gap in the center post... since if you torque a screw that doesn't go very far out cracks them.... ferrite being non malleable and very brittle.


I was targeting a higher brightness; which leads to a continuous oscillation; which leads to higher current draw. 


@TK
I was playing with a electro-smog harvester also; I thought I was close, but ended up I was just tuning to a sub-harmonic of the square wave I was testing with... most of the harvester has to be inductance? ... like if I calculate 1200000Hz (1200Khz) 2nF I get 8.7uH; which is like 8 turns of wire at about 3.5 inches... which picks up my frequency from the function generator, but gets not enough power for 1 LED... but at 33pF (533uH) how do I overcome the capacitance that's in the coil to get down to 33pF?  I mean the capacitance in just the wire exceeds that when using many feed?
I found this calculator http://hamwaves.com/antennas/inductance.html  ; but the design capacitance yields a negative number?  -9568.867751635291 
[strike]for a 5mmx50mm coil of 50 turns using 0.015mm wire ... hmm maybe I should use reasonable values [/strike]
for a 50mmx20.2mm coil of 100 turns of 32g (0.202mm) wire yields -0.785pf at 1.2Mhz frquency; (401.3uH)... hmm guess the induction is too high at that point...




I guess the capacitance should be as small as possible to get the most voltage out of it in resonance; which is what my caps between the coil and load end up doing... without the cap at the top I can't drive as high of a voltage load.
I remember from way way back when my dad made me a crystal radio; no power and it made the headset work pretty well.
Although; I don't get why the capacitance of the wire itself disappears.


--------
So then headsets; 2000ohm high impedance are sold; several look like the headphones my dad had (from military; he used to work on radar in alaska but anyway)... so not much current must be used; since (being generous) to get 12V... that's 6mA... would be curious how much current your smog harvester is actually harvesting.


some guy Evaluating 'airnergy' device from RCA that charges a battery from wifi harvesting.
https://www.youtube.com/watch?v=P8s3Xjeg0sk  (take 8 years to charge your battery)


another product http://www.powercastco.com/products/powerharvester-receivers/  power receivers for 900Mhz (phones) 23 dBm(200mW, hmm dBm relates to a log of W somehow), 4.3 V 100 mA (max)  ; (recommended usage to receive radio power and power remote sensors which in turn generate a radio signal; recycling the power)


-----
I guess falloff is (power out=100 kW/(4·pi·r2) ) for a 100kW source... so if I needed (2.7V@30mA) 0.081W, I could be 3939m from the source...
so if there was a 5MW source 500m away one could get 1.59W (5000000/(4*3.14*500*500)).  I'm not sure that would be enough for Ruslan's device.
(according to http://electronics.stackexchange.com/questions/4664/stealing-energy-from-radio-towers-or-power-lines discussion)
---
I built a resonant tank that's good for 60Hz; but I don't know how to get power into it since it's based on a pot core.

[TinselKoala]

Big caps take long time to charge... and to discharge. I've taken to putting "bleeders" on my big caps in JT experiments so I don't have to wait so long for them to get down to interesting voltage levels. A momentary contact pushbutton and a 1R or 10R in series, across the cap.

I've seen periods where the big caps will rise in voltage while powering the LED load. This is dielectric recovery, not energy from the circuit going back into the cap, and it happens most strongly just after a heavy discharge -- like by using the bleeder circuit -- and going back to normal discharge loads. The LEDs will shine brightly and the voltmeter on the cap will show a steady increase in voltage for a while. This can even be repeated several times with the supercaps I'm using, 10F @2.7V.

As far as tuning goes... it is theoretically possible to tune to a given frequency with arbitrary combos of caps and inductor values that satisfy the resonant tank formula, but I have found that there is a "balance" of sorts between capacitance and inductance that seems to work best. I don't know how exactly this optimum balance can be calculated. Maybe it is "pF ~= nH" or something like that. In any case I think the capacitors are more lossy than the inductors we usually play with, so a larger cap will waste more of the power in the tank than a larger inductor will. But the larger inductor will lose more by RF radiation than the large cap will.

The Ruslan thing: He has shown videos where he is actually inside the transmitter building of the FM radio station. I think he is probably a lot closer than 500m when he does his demonstrations. And there is much less fall-off in the near field, it goes more as 1/r rather than 1/r².  I have also just last night completed a wireless power transmitter that could easily be concealed in a coat pocket with the transmitting loop around waist or shoulders, and could be powered by a small LiPo battery. This apparatus will light up the Electrosmog Harvester or other co-resonant circuits out in the middle of a desert far from any radio stations or other sources of power... as long as I don't let you search my clothing.

Getting power into and out of the pot core coil itself isn't as hard as it might seem... after all the core is connected to other circuitry. And suitable receiving antennae can be concealed relatively easily.

http://www.youtube.com/watch?v=BQi4jz2puio



If you have any JFETs handy... try the ConradElektro/magpwr 2sk170 circuit!

[d3x0r]

Getting power into and out of the pot core coil itself isn't as hard as it might seem... after all the core is connected to other circuitry. And suitable receiving antennae can be concealed relatively easily.

http://www.youtube.com/watch?v=BQi4jz2puio



If you have any JFETs handy... try the ConradElektro/magpwr 2sk170 circuit!

But at 60Hz a suitable receiving antenna ?  quarter wavelength is 4,100,000 Feet :/ 


Ya I saw that mosfet version; was thinking of playing with that. ( http://www.overunity.com/13175/25mv-joule-thief-powered-by-peltier-merely-using-our-body-heat-free-energy-247/msg410419/#msg410419 )  I don't have any JFET..I think... just power mosfet; but I think that just changing the core winding ratios can overcome that... most of the fets I chose for low gate capacitance