Successful TPU-ECD replication !

[BEP]

@ronotte

Another HAM! That figures 

Not real active myself until the cold winter. I still do some PSK to Europe on 10 meters when the waves are friendly (Only 5 Watts on a Q5RV!).

Good I'll have an idea what suggestions aren't needed then 

[EMdevices]

Hi ronotte

Nice photos and work bench/equipment, etc..   I love it !!!

Thanks for getting us excited again about simple experiments and pulsing coils.

My comments are this:  

This appears to be the same setup otto was talking about earlier.   It's a DC2DC converter, magnetic field is built up, current is cut, and the spikes happen through the light bulb. 

Now, you have included a longer multiturn cable through the center of your caduceous coil.   And you're seeing some pulse trains after the main one.   What I would do in your case, is try to figure out if those are due to RETRIGERING of the mosfets, or if they are truly reverberations in the coils.   So,  take a composite scope shot, showing the lightbulb terminal waveform along with perhaps the GATE voltage on the Mosfets.    You see, the high power magnetic fields could be INDUCING a voltage in your triggering circuit.

Anyway, that's what I do to separate the CIRCUIT dynamics, from the COIL physics.  

EM

[ronotte]

Hi all,

@EM

Yes I agree with you: I know that the building of so big spikes does induce via capacitive way & Miller effect a replica on Mosfet's Gate so this evening I'll try to verify using the second scope channel the relations existing with the input signal on the Gate. Please take note that the wire lenght connection between the Mosfet and the Caduceus are very short.

I anticipate that, as seen many times before, the signal on Gate is always very noisy...but.....if you succed to put the mosfet itself into avalanche mode than the signal on Drain is clean and DON'T replicate in any way the noise on gate! : just what I see now.

I'll report soon.

Roberto

[ronotte]

Hi all,

well it seems that I?ve understood the reasons behind the artifacts on my previous pictures. Let?s go with order:

For first thing I?ve reproduced the same conditions :

Finput= 146 KHZ, fdo input duty cycle=50%, ??..on pic 52 the bottom trace is the Mosfet?s Drain and the screen bottom is the zero line; the upper trace is the fdo taken just on Gate. You can see that the input signal amplitude is 6V and in the instant of input trailing edge (when the output goes to +520V), on the gate there is a quick damped high freq oscillations mostly under the switching threshold?.so don?t care  (that oscillation is probably due to Miller effect). Following the oscillation there is about an input peak of 6V that seems to correlate with the 120V peak on the Drain and then there?s a small input peak (just before the next input pulse) of about 400 millivolt, again which seems to correlate with the about 50V peak output.  All said input artifacts should not produce the output peaks ?as they are well under the threshold. So in my opinion the 2 output peaks are without an answer?until now. Then I controlled the generator settings to see if input amplitude & offset had any effect on fdo at Gate. OK it has been my fault, that waveform must be taken putting the scope ground clip directly on mosfet?s Source pin?.magically most of the trash disappear (adjusted also the generator settings: Vinput= +4V, offset=1.33V)!

Then I tried to reconstruct the situation of pic 47 setting the duty cycle to 70% (off) and found best point with Finput = 55KHz and I=0.7A. Situation is now on pic 54. Here the input scale is 2V/div and you can see that all the trash is gone together the damped oscillations on Pic 47, is left only almost a single peak of 50V probably as an effect due to input damped oscillations (See pic 54,55).

I do confirm as well that connecting the load in coil derivation without the mosfet?s drain connections gives no output.

Roberto

[ronotte]

Hi all,

the thing is getting interesting! after examining more deeply the pictures posted and the 'cleaned' waveform, it seems to me that there is still an answering waiting for the two pulses that appear so well in pics 44 and 52.

Actually there are two possibilities:

1 - That posted by BEP e.g. the pulses are an effect explained by TDR  (Time Domain Reflectometry)...so they are just a reflection due perhaps tho impedance mismatching between Caduceus coil and load.

2 - That published by JLN laboratories (http://jlnlabs.imars.com/spgen/index.htm) in their experiment with Soliton creation with Caduceus coil . Please find attached the observed waveform and do note that they are ...just the same!!! except my output amplitude is much higher due to: a more powerful Caduceus (24 nodes) and the presence of an inner collector coil....and there is an higher time between the main pulse and the 1st Soliton pulse, but this could be explained with the different phisical construction of my coil and also the different frequencies (much higher) now used..

Analyzing the first point, I've tried to vary the wire lenght in order to see if the 'reflected' pulses accordingly vary their position on  scope x-axis (time). Result is that the pulses position remain unchanged!....so no time domain reflections.

Analyzing the second point, the first big answer is: the pulses appear only using as input a squarewave with 50% duty-cycle!   the shape, the time interrelation seems equal to JLN report & attached pic!

Stefan, you have worked lot of time ago with JLN TEP project & Caduceus coil experiment...what do you think now? 

Another big difference with TEP project is that I do obtain a big output  (and COP near 1): you can see it on the lamp pics and the current showed on PS current dial....and there's no exciting coil like in TEP, again the amplitude (of the main 520V peak and the Solitons) are much higher (than in TEP) and the Caduceus  does stay cool.  I've tried to vary the phisical coil configuration but found that the two pulses always stay there!....I wonder what will happen feeding the Caduceus with a pre-mix of 2 or 3 frequencies!

Roberto