Hey Luc,

Here are few pictures of the other different drivers I have built. One with many heatsinks has 24 HV diodes to protect mosfets but I think they distort signal too much.. so I do not use it. Yet I built one more of a simple version with better hexfets IRFPS40N60KPBF 600V 40Amp 0.011 Rds and also got couple of cpu cooling heatsinks with fans to assure heating would not be of any issue. This last one seems to look and run sweet although with  IRF640s I still had better results but they are too delicate for my needs and were blowing up. Tested some 50W feed for 10 min and all was cold to the touch. Will see how it goes.

Minde

I just wanted to throw in my two cents... it's nice to see that some people out there actually take an intelligent interest in Tesla's incredibly advanced (not to mention almost martyrlike devotion to humankind).  People who are not trying to string me along so I will buy their new age power crystals or invest my life savings in their vaporware scam.  Power to the People!  Stop sucking on the dinosaur tit!  Anyway, I think Tesla once said that electrical engineering that doesn't incorporate resonance in its design is simply a waste of time.  His wireless system would have eliminated the almost criminal power losses we get using wires that have resistance.  "overunity" or "perpetual motion device" are terms I hate to see, because almost every breakthrough in science was originally shouted down as being against the "laws of nature", which apparently only the established lecturers and university profs have the right to define.  Not that amateur scientists are somehow morally superior to the old guard... but... once established and entrenched, change in the system of beliefs must appear to be a terrifying attempt at revolution, to uproot their whole world that is based on the old system of beliefs.  When people close their minds and refuse to see evidence that contradicts their secular religion (as I think of it), then all innovators become modern day heretics.  such is the human condition (stepping off soapbox). 

Hi lads, my first post here, I read the first 5 pages of replies and didn't see my question, so here goes.

I'm just beginning to get into resonance with coils and was wondering, having seen this video - Coil Resonance Tutorial 2 http://www.youtube.com/watch?v=MgAF8MxS7Rg ... Once the right capacitor has been found for a particular coil, will that be enough by itself to ensure the most efficient circuit design, or would I need to include a frequency generator circuit in every Joule thief I make?

Very good videos Luc, if you keep making them, people will keep watching them.

Hi Spirality, thanks for your interest and positive comments.

I have found that there is an ideal capacitance for a particular coil at a certain frequency for minimum power in and max out.
Once you have found this and make an identical duplicate, then it should be the same. However if one of the three parameters change, then yes you will have to re-tune by one of these three. Usually frequency re-tune would be the first choice but if that is fixed then you can wind your inductor longer so you can tune by removing wire a small piece at a time. Capacitor tuning is possible but more expensive.

All the best in your experiments

Luc

 The circuit is just a capacitor and coil in series connected to the output of a signal generator.
Luc
 

Hi gixo,

Nice job,  however I think you considered the power dissipation in the 1.4 Ohm resistor as THE input power,  this 2.64 W is only PART of the actual input power.  The current via this resistor is I=1.935 V/1.4 Ohm=1.38 Amper.  So the input power draw from the 12V is P=12*1.38=16.56 W, and if you deduce the resistor dissipation of 2.64 W you get 16.56-2.64=13.92 W.  Now if you relate this to the 8.42 W dissipated in the output resistor, the efficiency is 8.42/13.96= 0.6    i.e. 60%

(Notice that in the calculation I accepted your measured voltage numbers, in fact the voltage drop across the 1.4 Ohm should be checked with an oscilloscope or a true rms AC meter for getting the real amplitude whether the voltage drop has any AC wave (if you use capacitor filtering at the input across both legs of the 1.4 Ohm with respect to the negative pole of the 12V source, making a low pass RC filter, then the measured voltage drop across this filtered resistor as a DC voltage value is acceptable.)

rgds,  Gyula

Hi gixo,

I wonder what you mean on the "two different input results"? Based on your data I figured only one input power and you accepted it (13.96W).

The moment you connect a resistor in series with your 12V DC supply which would feed your original circuit you have created a voltage divider and this divider splits the DC supply up to two voltage parts: one of these is the voltage drop across your series resistor (1.935V) and the other is the rest voltage: 12V-1.935=10.065V actually this is your voltage source from which your circuit operates. And you may check the input power by this 10V times the 1.38A current (this flows through both the series resistor and your circuit) and you get 13.8W (very close to the 13.96W, the tiny difference comes from omitting some decimals).

I understand that your output voltage can go up as high as even 500V across the output electrolytic capacitor(s) BUT this is the UNLOADED output voltage!
SO if you (say) measure the unloaded output voltage to be 550V you DO NOT have any current YOU consume from the output, now if you take the 293.3/10210= .0287A current and relate this to the 550V then you fool yourself because this .0287A flows IN your 10.21kOhm load resistor when you connect it to the output BUT then your 550V will NOT hold: it will go down to 293.3V as you measured, ok? 

The basic requirement for a successful output-input feedback you would like to achieve is that the output should be higher than the input.  A few years ago a member (poynt99) uploaded a possible block diagram for testing such looping, see here to download:
www.overunity.com/downloads/sa/downfile/id/248/   Study the last two pages, Appendix1  (you have to keep in mind that  LED devices also have their own forward voltage drops and also consume some power when they are lit). I show this you as a principle diagram, it should be modified for your circuits of course.  The 8 series LEDs across the input DC source serves also as a voltage stabilizer for that circuit involved in the document.  The bottom line is that you have to use some means for stabilizing the output voltage from your circuit when attempting to loop it back to the input, to prevent a possible run-away situation when the output indeed exceeds input.

rgds, Gyula

The reason for these results is now understood, thank you Gyulasun.

What is more confusing is how this device naturally resonates energy through the coils, yet does not perform past 100% return as does the H-Bridge controller tuned to resonate as is shown within the first 10 pages of this discussion.

In the future I'll try the feedback circuit you've shared as a standard measure of performance.

Would a voltage regulator chip as present in many 12V battery charger diagrams be able to improve the design, rather than using load bulbs to reduce the voltage?

The circuit is just a capacitor and coil in series connected to the output of a signal generator.
Luc

Hi just found this article, I'm not sure this statement is true as such, my confusion is the video the guy Luc is it ? talks about a narrow pulse gives him the best results, but then he has the same feed going into a CD4013 'D-type this device divides by two because it just triggers on the one edge so you would end up with a EW 50/50 waveform, if you want to correct this, sling the 4013 in the bin or rewire it for CLK reset mode with steering diodes or just use a device like the IR2111, this will give you a full bridge drive to your other two IR 2103 chips  or faster version, however you can  only get what is available locally with this logic the same applies to the power that is used to harvest it . So not much chance of any real power here. but it could be useful for battery charging.