Joule Thief Experiments Reboot

[SkyWatcher]

Hi all, I have been experimenting as usual. There was a young man who posted some intriguing you tube videos in the past.

He later removed those videos.

A video showed him powering 100 5mm white leds using a speaker wire pancake coil wired as a joule thief.

The output of all the leds was very impressive, full brightness and the efficiency was spectacular.

This motivated me to continue experimenting with joule thief type circuits.

Where a feedback coil is used, which allows the circuit to pulse itself and when the circuit is tuned,
allows for nice performance and efficiency.

I have been testing different coils and came upon a nice combination of parts.

The latest promising results, I'm using at least 100 feet, though it might be more, Ethernet network cable, it has eight 24awg. insulated,
solid copper wires inside that we can use.

I wired them all in 2 wire pairs, so we now have effectively 4 wires.

One of the wire pairs is used for primary coil, another is used for the feedback coil.

The other 2 wire pairs are placed in series and used as a secondary coil.

The secondary coil option seems to help get the voltages higher and into some sweet frequency spots.

The joule thief circuit is not stock, it is tweaked for more efficiency and is running an older 6 watt Eco-smart led bulb.

The secondary coil has an ultra-fast diode attached to one output end, to rectify the primaries magnetic collapsing field,
into a 100uF-450volt electrolytic capacitor.

The output of the capacitor is connected to the 6 watt Eco-smart warm white led bulb and the light output is quite impressive,
considering the power input.

Input is 30 volts - .0009 amps or 1 milliamp or 30 milliwatts input.

Using a boost converter for powering oscillator.

The light output, lights up my computer desk and immediate surrounding area, to perform all normal tasks
that a 1-2 watt efficient led bulb from Ikea, etc. would provide.

I'm very happy at these results and somewhat surprised, even though I'm well aware others have done similar things,
such as the other mans youtube videos.

I would suggest other people re-examine these joule thief circuits, especially using secondary coils and letting it loose,
with the high voltage it wants to produce.

Has anyone else, in the past or present, observed interestingly good results with a joule thief circuit, as I am observing?

Your comments are welcome.

[dragon]

This was posted by Mr. Rakarskiy some months ago on telegram, a joule thief type hybrid of sorts. I built this some time ago and it has some interesting results.

https://overunity-electricity.blogsp...nation-of.html

https://t.me/c/2083072553/1373

[SkyWatcher]

Hi dragon, thanks for the reply. Yes, i was noticing some videos from 'overunity-electricity' on youtube, very interesting things they are sharing.
And along the same lines as these joule thief experiments.

Here are a couple pics of the setup.

The computer desk area is being illuminated by 2 gutted led bulbs in parallel, rated at 5.5 watts-450 lumens.

I would estimate the circuit is putting out around 100-150 total lumens from the led bulbs, again for around 30 milliwatts input.

Comments and others experiments welcome.
peace love light

[gyula]

Hi SkyWatcher,

Nice and interesting setup. Your description sounds to imply that the circuit converts about 30 mW DC input power to a certain light output what you estimate
around 100-150 lumen. I know you have been experimenting with circuits for years that can drive LED bulbs as efficiently as possible, the higher the better of course.

Maybe you are not interested in some further measurements I would like to ask, if so, no problem.

For instance
1) could you check the 1mA current consumption from the 30V DC output of the zk-j5x boost converter with a separate Ampermeter? i.e. with a DMM or an analog meter.

2) could you check the DC voltage across the two gutted and paralleled led bulbs? i.e. what is the DC voltage level across the 100 uF capacitor?

3) could you check the DC current of the two paralleled led bulbs fed from the 100 uF capacitor?

I am simply curious, I know you did not claim any extraordinary results, just reported your findings and experiences.

Thanks for sharing,
Gyula

[SkyWatcher]

Hi gyula, thanks for replying and thanks for mentioning that, i didn't think to make those measurements, although i should have.

I set my DMM to the 200 milliamp current setting and it shows .7 milliamps or 700 microamps at 30 volt output of converter.

The DMM voltage across a single gutted led bulb in this check is 53.60 volts, same voltage across 100 uF capacitor connected to output of secondary coil.

Having difficulty reading the output current with the DMM.

The input checks out though and I'm happy i made these checks.

And yes gyula, I've experimented enough with these type of circuits powering led bulbs and this setup is the best light output I've ever seen, for such low input.

At least 100 Lumens or better, though it could be a little less or a little more, though still a lot of light for hardly any input, 21 miiliwatts input.

You can see in one of the pictures i posted, the led light shining brightly off the step down converter.

It's not a boost converter, it's a step down converter. I need to get a higher voltage, proper boost converter for further experiments.

I'm using a 31 volt power supply from an HP device i had previously.

All comments welcome, in light of these latest input measurements.

peace love light

[gyula]

Hi Tyson,

Thank you for checking the input current and the 100 uF capacitor voltage.
Regarding the LED current reading, the difficulty could be explained by the triangle-like waveform shape what the DMM may not like (the secondary coil pumps charge current when the primary's magnetic field collapses and cap voltage surely increases to higher than 53.6 volts, the LED is lit till the cap is discharged to a certain voltage level below 53.6 volts and then the next pulse comes from the secondary to repeat the process, hence no pure DC current can flow into the LED from the capacitor). An oscilloscope would be needed to measure the current across a shunt resistor.

A possible estimation to learn about the LED current would be to compare its brightness to the other similar type gutted LED you have, driven by a relatively clean
DC voltage source and measure its DC current when the brightnesses match by the eye. IF you wish to do it, of course.

IF you have no DC source in the 50 - 60 volts range. maybe you have a Variac driven from the 120V mains and a FWB rectification with a puffer cap could serve.
Here there would be still be triangle current wave shapes but at a much lower amplitude to cause trouble like the pulsed method may do the secondary works with.

I think the less than 1 mA average input current is a correct value, your circuit is a kind of a blocking oscillator which works with low duty cycle input current by default.
(Input peak current is surely much higher than 1 mA.)

Why I am curious on the LED current is that the primary coil receives a given peak current (Ip) from the 30 volts source, the accepted formula for a coil's instanteneously
stored magnetic energy in Joule is (L*Ip*Ip)/2 and this energy goes into the 100 uF capacitor via the secondary (and the diode). L is the primary coil inductance.

The stored energy in the 100 uF is 0.1436 Joule considering 53.6 volts. (C*V*V)/2 This would mean 143.6 mW in every second availabe But the LED cannot use this fully simply because
the capacitor is surely not discharged fully to consume all that.

Thanks for your kind efforts.
Gyula

[SkyWatcher]

Hi gyula, thanks for the reply and your interest in this project.

I will go the route of comparing the brightness, using a stable DC adjustable boost converter.

I will make the stable DC boost converter, using a joule thief again, with a ferrite toroid.

It will be the standard joule thief with diode from transistor collector to emitter negative.

Guess I'll use a Large capacity electrolytic after the diode output for more stable output.

Do you think just a large uF electrolytic will be enough for smoothing?

Do you think the light output seen in the pictures, compared to the input, is rather extraordinary,
or have you seen standard circuits surpass this easily, because i have not.

peace love light

[gyula]

Hi Tyson,

Well, the light output is surely very good and no I have not seen circuits surpassing that with as low input power as yours.

Re on large uF cap for smoothing: yes it can help filtering and clean the rectified pulse which would still be saw tooth like but with much smaller
amplitude riding on the DC average output voltage across the cap. Hopefully your DMM can measure it.

You may think of measuring the input current into the newly built joule thief when the brightnesses match, this is also a possibility. Hopefully the two gutted LEDs are close enough to each other
in their V-I characteristics, try to swap them to see.

Gyula

[SkyWatcher]

Hi gyula, thanks for the reply.

Sorry for the confusion, I'm only using one of the gutted led bulbs now and placed scotch tape over the top, to diffuse the light better.

Though only using one led bulb hasn't changed anything, it's just a more intense light output (blinding actually) with only one bulb, which is why the use of diffusing material.

Yes, the input current fed to the standard joule thief, might be a good way to see efficiency differences between the two setups.

I can already foresee, the standard joule thief using noticeably more input power to achieve the same results, though we shall see.

Maybe I can initially check the efficiency of the standard joule thief test circuit, using a resistor load at capacitor output, to see power out verse power in.

This way, we can account for the power loss of the standard joule thief and subtract that inefficiency from the results.

Let me know what you think about this?

peace love light

[gyula]

Hi Tyson,

Ok on your using one gutted LED bulb presently, you wrote this in your replies above, no confusion. The 53.6 V was across the single bulb.
The more intense light output you refer to above came about by the removal of the 2nd bulb, the discharge current from the 100 uF cap was roughly halved, hence more 'juice' remained for the single bulb. OF course I do not mean the high brightness is due to the use of a single bulb, it comes from your expertise.

Checking the input current into a standard joule thief and using a resistor load across its capacitor output to estimate power in versus power out is
good for that particular circuit. It is sure what you can also foresee would be justified: I think efficiency would be anywhere in range from say 60 % to 90 %.

Such test is not conclusive though for my original question concerning LED current in your high brightness circuit.

I understand a pulsed LED current is best evaluated by the voltage drop it causes across a shunt resistor with an oscilloscope. In the lack of a scope the 2nd 6W ECO bulb could be driven by pure DC voltage via a resistor to reach the same brightness as your circuit produces and the DC current of the bulb could be measured.
Here I assume the DC voltage across this 2nd LED would be close to 53 V at the matching brightnesses. If the matching brightness involves a different DC voltage level across the 2nd bulb in the DC test, it means the measured 53 V may not be correct due to the presence of some pulse across the 100 uF which fooled the voltmeter (I do not think this was the case).

If you cannot do this pure DC power test with the 2nd bulb by any reason (when the brightnesses match), no problem I understand.

Thanks for your kind efforts,
Gyula

[SkyWatcher]

Hi gyula, thanks for the reply. I found another gutted led bulb that lights up to good brightness, with only 24 volt input, direct from the step down converter,
no joule thief circuit involved.

I borrowed the step down converter from the current test project, to get a ballpark comparison.

The odd thing is, at 24.10 volts DC output, the converter shows no visible amperage.

So I assume it's below 1 milliamp, similar to the test project input power, and yet, the leds are blinding bright,
painful to look at and probably similar brightness or better light output compared to the test project.

Now I'm confused, I wonder if it's something to do with this step down converter, as I'm using a large 71 volt - 5600 uF electrolytic capacitor in parallel with the
converters output, for this test and the tests with the project under discussion.

I wonder if the smaller capacitor, that is built into the converters output section, is interacting with the large capacitor in some odd way, to cause such low input power,
or this is normal behavior when under powering leds in general.

Either way, this test using the direct DC output of the converter, is also giving quite a lot of light output from only 24 milliwatts or less, yet again.

Maybe this is all normal behavior, and I'm underestimating what led's can really do with straight DC, when under voltaged.

Though 24 milliwatts or less for painfully blinding light and quite a bit of very usable light, seems extraordinary.

Let me know your thoughts gyula.

As this is odd to me, or leds are just that amazing these days and one only needs to sip power to get amazing light.

Though again, all my past experiments have never had this good of light from leds for such low power, i must be missing something obvious.

peace love light

[gyula]

Hi Tyson,

I would suggest a simple test with the step down converter: could you load its 24 volts output directly with a normal resistor of say a few kOhms? like 1k or 2.2k?
And measure the current via this resistor, Ohms law predicts 24 mA when the resistor is 1 kOhm, etc.

Use the same Ampermeter of course. Use the 5600 uF cap across the converter output too. And then remove it and see if the 24 mA via the resistor changes?
Do you happen to have another Ampermeter to double check the DC current?
Will continue later.

[SkyWatcher]

Hi gyula, i redid measurements and think i had a measurement error, that stinks, I'll post the new results later

[gyula]

Hi Tyson,

No problem, instruments may often trick us.

In the meantime I think I found the posts by the member at the old overunity.com forum you mentioned in your 1st post above who deleted his intriguing videos.

The forum has been suspended due Stefan's illness but someone made it a readable archive but without all the attached files.

So here is the post to which that member attached his schematic https://overunityarchives.com/index....html#msg489068 but no schematic.
I believe this is the guy you referred to, right?

Thanks,
Gyula

[SkyWatcher]

Hi gyula, no I don't think that is the same experimenter. His videos start out with saying, "uh oh, it's super critical" and he shows high voltage arcs and things.

I'm still going to try and replicate his results. He used speaker wire pancake coils and showed that a smaller one and larger one did not have the same resonance effects.

So probably need a good flux capacitor coil setup, meaning a coil that has some capacitance to possibly see those results.

peace love light