Joule Thief 101

[sm0ky2]

Try to clean that up

[MileHigh]

So, despite the mathematics, and solid foundations in electronics theory,
despite all that has thus been discussed here,

you set in stone your ridiculous opinion, before even considering the experimentation?

Honestly, I could slice and dice your technical comments apart if I wanted to.  I think that you are just being a poseur throwing out some technical terms and making a tired worn-out resonance play.  Why don't you sketch out various "resonance" waveforms for Magluvin ahead of time so he knows what to expect and where to look for them?  Why don't you explain the multiple resonance instances that you alluded to and explain where they are and what and where each of the two resonating components are in each instance?  Or do you just want to sit back and watch the experimenters fiddle with pots and comment when you think someone has "struck" resonance?  Why should an experimenter have to "hunt" for resonance and find some "special delicate balance" if you have been pitching it all this time?  If I strike a bell it resonates.  What kind of "special" or "remarkable" results should Magluvin and others get if they "strike resonance?"  What are they supposed to see?

If Magluvin or anyone else succeeds in achieving something remarkable with a Joule Thief in "resonance" and explains what is actually taking place instead of just observing something, I will be happy to admit that I was wrong and acknowledge that something special is taking place due to the resonance.

On the other hand, if all that Magluvin or others can get is mushy wobbly scope traces that are difficult to explain and don't clearly show "resonance" and yield unremarkable power-in to LED-illumination-out results (or any other metric you want to define), what are YOU going to do?

MileHigh

[MileHigh]

Magluvin and other Joule Thief enthusiasts:

I reattached Magluvin's second scope shot to this posting.  Clearly a Joule Thief when running at a very low battery voltage does not operate in the standard switching mode as shown in his first scope capture.  I certainly can't explain what I am seeing.  I think that we have all seen similar scope captures of Joule Thieves running at very low battery voltages in the past - but most people have never seriously questioned what they were looking at.  Clearly something different is happening and it's happening at a much higher frequency than the normal operating frequency.

So do you just observe this, or, do you try to explain it and understand it completely?  If you can figure it out and explain it will you will get some satisfaction from that and in addition perhaps that new knowledge will give you more insight into trying to get a Joule Thief to resonate?

The good news is that you can slowly lower the supply voltage and observe how the waveforms change and that will help you a lot in determining what is taking place.

MileHigh

[Magluvin]

Magluvin:

Thanks for the scope shot.  You can see the how the Joule Thief transformer is an inverting transformer relative to the two test points.  If I assume that the winding is a 1:1 ratio then I can even see the battery voltage in the scope shot.  The LED voltage peaks at about 6.5 volts and the feedback volage peaks at about -5.3 volts.  So the battery voltage must be about 1.2 volts.

More importantly, this scope shot of a Joule Thief operating normally confirms everything I said about how it operates including the copy-paste articles.  There is no "resonance," there is just the energizing cycle time and then the discharge into the LED cycle time.  Add those two values together and invert it and you get the operating frequency for the Joule Thief.  The positive-feedback "snap ON" and "snap OFF" events are nice clean very-high-slew-rate events.  Finally you can see it is running at a nice "low" frequency of 4.2 kHz.

You can also see that potential issue that I raised on the other thread.  How bright is the LED at 6.5 volts?  (Actually the LED is probably over-driven at 6.5 volts and and you are getting less bang for your current buck at that operating point.  The LED would probably burn up with a constant 6.5 volts across it.)  How bright is the LED at 5 volts?  How bright is the LED at 3.8 volts?  Is this an issue?  Is the decaying voltage/current discharge curve though the LED resulting in a lot of "lost" inductor energy that is not really contributing to the brightness of the LED?  If so, how can you improve upon this?

MileHigh

No problem.

Well sure there is no resonance, yet. I dont think anyone has said that there is resonance going on in a normal JT.  That is why we are modding the thing to get it to do so. 

Mags

[Magluvin]

Try to clean that up

Will give it a go.  Need to increase the resistance a bit more, and slowly scan the pot and watch for what you are asking for. My 10k 10 turn broke internally. Have a 1k 10 turn that Ill put in series with a larger 100k pot and a 100k resistor and use the 100k as a course set and the 1k a a fine tune.

Thanks 

Mags