Steven Mark`s TPU

[tosky]

@rensseak
You can reference to the datasheet. of the MOSFET from the manufacturer. The datasheet tells the (minimum , typical, maximum) of the G-S cap. This cap. is not wanted and the manufacturer always want to keep it small.
It is not a cap designed on purpose just a side product of the manufacturing process. So no one will use a side product for a reliable circuit to do some function. It's value also not guarantee.
When there is a Capacitor(MOSFET), an Inductor(wire), Energy (Magnetic field). Oscillation condition established. Only need to trigger it once, it will begin to oscillate. But the frequency and amplitude will change according to the conditions.

[ronotte]

Hi all,

Otto, in your case I fear you have damaged (opened) one or more of the enameled wire coils: that's the reason why I did all the coils (except the feedback coils) using high isolated electrical wire.

EM, I can't disconnect now the Mosfets as they are soldered inside the TPU and thanks for the Negative Resistor circuit. Anyway two thinghs may clear the scenario about auto-oscillation:

1 - While doing this tests I forgot to say that I'm using a 'pulse transformer approach' to drive the power Mosfets so NO NEED TO PUT RESISTOR TO AVOID STRAY INDUCTION FIELDS on the Mosfet's Gate (distance between pulse-transformer - Mosfet's Gate < 10mm.

2 - Actually the oscillations does trigger while the 3 function generators are indeed phisically connected to pulse transformers (located inside the TPU)  BUT SWITCHED OFF!!!.  So no way that I know to capture stray field as generators output impedance is low = 50 Ohm (the same happens while generator is ON but the output digitally set to 0.01V) and I used mostly shielded cables. The pulse-transformers themselves are of course very low sensitive to stray fields as they are built.

It does remain the fact that the depicted scenario is well stable & repetitive and does deliver power. I'm also still convinced that in some way the mosfet, inside the TPU are working but, as the current required (at 12V) is much less that required with standard pulsing on the same frequency, may be that they run in a different modality. A waveform confrontation in different circuit points could be of help in understanding.

Roberto

[Gustav22]

Hi Roberto,

regarding the following:

...I can't disconnect now the Mosfets as they are soldered inside the TPU ...

and this

... A waveform confrontation in different circuit points could be of help in understanding...

and also
BEP's idea that the amount of copper of the CCs should be adjusted/reduced to move the auto-resonating-frequency closer to 35 kHz.
edit after BEP spotting my mistake:
BEP's idea that the number of turns of the CCs should be adjusted to move the auto-resonating-frequency closer to 35 kHz.

All these tests and verifications are very important but all would require that you interfere with the unit.
I think you should NOT disassemble or open the actual unit you presently have, if you can avoid it.
I would rather recommend to make a replication (or even two) and then play/change/adjust the parameters/specs on those clones.

I volunteer to help you winding them.

[ronotte]

@Gustav,

THANK, of course at least for now

I WILL NOT DISASSEMBLE MY UNIT

I've a rich scheduled & organized tests to do so I'll go-on with:

- Discover the trend of output on lamp by executing a complete F1 sweep (range 2 - 70KHz) for different PS voltage (range 5 - 45V)

- the same but using F1+F2

-the same but using F1+F2+F3

At the end I should have a complete useful (I hope) scenario.

If anybody like to suggest other meaningfull test I'll be happy to discuss.

Roberto

[BEP]

BEP's idea that the amount of copper of the CCs should be adjusted/reduced to move the auto-resonating-frequency closer to 35 kHz.

Please be aware with this type of inductive coupling many of the concepts are REVERSED, or should I say perpendicular
To increase the frequency of the collector you must increase the turns of the collector while maintaining the number of turns of the control.
You may also increase the number of control turns while maintaining the collector turns.
Increasing the collector mass only will slow the current and the resonant point.

So my sugggestion for the next one is either add collector turns OR add control turns to INCREASE the resonant point.

@ronotte

Please supply the specs or manufacturer/part number of your MOSFET gate transformer. I will take mine apart as it is not duplicating your results.

Also, are your MOSFETS still without a heat sink? I'm wondering about density of charges on the backplane of the MOSFET.
IFa field is rotating in your TPU - would not that field be able to trigger the MOSFET as it passed? - simply because the FET part is made to be effected by fields