Kapanadze Cousin - DALLY FREE ENERGY

[Dog-One]

Be aware you can not use the TC4420 in combination with the lossless clamb, they haven't got the ENABLE pin.

I asked Verpies about that.  The enable is simply a safeguard.  It isn't
mandatory.  Just make sure your push/pull driver has some dead time.

What this means is you can simplify Verpies circuit somewhat and use
it as a retrofit for pretty much any push/pull driver design.  The only
potential hang-up you may run into is getting all the windings to fit
on your core/yoke.

[itsu]

Nick,

i have taken a snickers as well, feeling much better now too.

Those gate signals looking good now, you have enough dead time inbetween the pulses, so the heating up cannot come from the MOSFETs overlapping their firing time.

Make sure you have installed MOSFETs that test out ok from verpies his MOSFET test diagram.


As Hoppy also said, you should not run with the yoke connected without snubbers, but to calculate the snubber values we need to run it first, so kind of a problem here.

If you really think/see something is wrong with the yoke (unbalans) then fix that first.
Then you could run with a lower drains voltage (12V), or use a low ohm (2 to 4 Ohm) power resistor in the 24V yoke lead or 2 12V automotiv bulbs in series so when there
is a high current draw, it will cause automatic decrease of voltage and so protect your MOSFETs.

For calculating a RC snubber setup, we need the drain voltage signals (spike / ringing) to calculate the ringing frequency, then follow this procedure to find out the correct
RC values for each MOSFET (between drain / source):  https://www.maximintegrated.com/en/app-notes/index.mvp/id/3835
So when we know the ringing frequency (111Khz in te last test according to the drain voltages, but i doubt they are valid in your present setup as i expect some ringing in the
10Mhz range) you should use some capacitors across the drain / source to try to half that frequency, etc. see the 5 steps in the linked procedure.

Like Dog-One mentioned, i also skipped the enable line protection on the lossless clamp diagram (Circled H (D5, R5 and R6) and circled I (D7, R7 and R8)) and the
optional helper Schottky diodes D1 and D2.
R11 and R12 are current sensor resistors, very handy, but not really needed too when it works OK.
When using TC4420's drivers, R1 and R2 can be substituted for a single resistor as the output pins on a TC4420 are connected together (pin 6 and 7)
Same for R3 and R4.

Anyway, take small steps so we don't overlook anything.


Itsu

[Hoppy]

As Hoppy also said, you should not run with the yoke connected without snubbers, but to calculate the snubber values we need to run it first, so kind of a problem here

Itsu

This is why we have stressed in a few posts that Nick must first attempt to limit the supply current before experimenting and for calculating snubber values, otherwise he is inviting even more smoke! Currently, as I understand from Nick's posts, he driving the yoke without current limit or suitable snubbers!!

[AlienGrey]

This is why we have stressed in a few posts that Nick must first attempt to limit the supply current before experimenting and for calculating snubber values, otherwise, he is inviting even more smoke! Currently, as I understand from Nick's posts, he driving the yoke without current limit or suitable snubbers!!

Ha Ha! Nicks new wave form is just PERFECT   TO LIMIT SUPPLY IS EASY use a switchmode chip one of the SG3524 series regulator devices will work just fine or he can buy one (but problem) it will have an earth link, why not get a 50w 12volt or 2 in his case in series any fault and the bulb will light simple.

regards Allen (AG)

[AlienGrey]

    Itsu:
    Below are my current gate signals, showing a lower duty cycle.
  volts/div set at 10, and sec/div is at 20us.
  Let me know if you need anything else to figure the snubber values, and I'll always do as you suggest.

  TinselKoala: 
  Yes, my gates are getting 15v from the UC4420 drivers.
  My PSU is 24v, and the UC4420 drivers are getting 15v through the 15v voltage regulator, and to the gates.
  My TL494 circuit is getting 12v through the 12v regulators and filters.

     I've ALWAYS had hot running fets, always, no matter what I do. So, I think that the unbalanced problem as well as the overheating issues may be due to the yoke.  However, I'll install some snubbers, first. Even though the Stalker diagram which I'm following does not use any snubbers.  After that, I'll change to the new opposite windings yoke, and give that a try.

   apecore: Can you please upload Verpies lossless clamp diagram for me, as I don't have it anymore.
Looking for it on the site, can be a pain in the butt.
   
   Thanks, guys.

brilliant display! like it !
PS the hunt for zero point goes on !