Kapanadze Cousin - DALLY FREE ENERGY

[Dog-One]

The other thing that helped me get over a huge hurdle was the number of
turns on the primary for the core I was using.  My core could not handle
having only 12 turns per leg.  When I used 16 turns per leg, everything
worked fine for the frequency range I needed.  Pretty certain I was
saturating the core.  When your core saturates, snubbers will do nothing
to prevent the MOSFETs from heating up.  Basically you are dumping
gobs of current over a very short interval, especially if you have hefty
filter caps on the input power source.  That current turns directly into
heat at the point of highest resistance--your MOSFETs.

[AlienGrey]

The other thing that helped me get over a huge hurdle was the number of
turns on the primary for the core I was using.  My core could not handle
having only 12 turns per leg.  When I used 16 turns per leg, everything
worked fine for the frequency range I needed.  Pretty certain I was
saturating the core.  When your core saturates, snubbers will do nothing
to prevent the MOSFETs from heating up.  Basically you are dumping
gobs of current over a very short interval, especially if you have hefty
filter caps on the input power source.  That current turns directly into
heat at the point of highest resistance--your MOSFETs.

is this a yoke or a toroid ? also are you using a cheap low-frequency ferrite from a 50 or 60 hz system as that could produce the effect you're having also you don't say what voltage you're using so your information could be misleading others who are perhaps using the correct components for the job, also did you bi-filar wind the drive windings ? and are you recycling the fly back spikes into the supply ? all this information could be useful to others who are struggling. for instance if using the correct toroid might only need 8 turns to achieve correct working at 12 volts drive using a bifilar winding.

[Dog-One]

is this a yoke or a toroid ?

I used a 3 inch high-perm nanocrystaline toroid.  Expensive motha.

Once adjusted properly, I have no trouble pushing 500 watts of power through it.
MOSFETs run at room temperature.

also you don't say what voltage you're using so your information could be misleading others who are perhaps using the correct components for the job

48 volts per leg.  24 volts is easier to get working.  More volts, more turns.

As far as "using the correct components", I'll leave that to the reader to discern.

, also did you bi-filar wind the drive windings ? and are you recycling the fly back spikes into the supply ? all this information could be useful to others who are struggling. for instance if using the correct toroid might only need 8 turns to achieve correct working at 12 volts drive using a bifilar winding.

I used the resistive snubber setup that snags the back EMF, collects it in a cap
and dumps back to high value resistors.  Done this way, the resistors get hot
when you're doing something wrong.  They only get slightly warm when power
is actually transferring to the load as intended.  Therefore they make it super
simple to tell what's happening before even attaching a scope.  If you turn the
push/pull on with the inductor coil disconnected, these snubber resistors get hot
quick, because there is no load, everything is reflected energy.  On the other
hand, when you have the inductor coil connected, tuned and heating a metal
screwdriver almost orange, the snubbers run cool, because the power from the
push/pull is going to the load and not reflecting back.

Attached is the setup I assembled a couple of years ago and still use for
all sorts of projects requiring a good push/pull setup.  Works pretty well also
for a single leg pulser as long as you don't need more than 50% duty cycle.

If I was going to rework this circuit, I'd use 15 volt drive via pluggable DC-2-DC
converter for the gate and the SG3525 chip.  Other than that, this is battle
tested driver that meets all my expectations.

[AlienGrey]

I used a 3 inch high-perm nanocrystaline toroid.  Expensive motha.

Once adjusted properly, I have no trouble pushing 500 watts of power through it.
MOSFETs run at room temperature.

48 volts per leg.  24 volts is easier to get working.  More volts, more turns.

As far as "using the correct components", I'll leave that to the reader to discern.

I used the resistive snubber setup that snags the back EMF, collects it in a cap
and dumps back to high value resistors.  Done this way, the resistors get hot
when you're doing something wrong.  They only get slightly warm when power
is actually transferring to the load as intended.  Therefore they make it super
simple to tell what's happening before even attaching a scope.  If you turn the
push/pull on with the inductor coil disconnected, these snubber resistors get hot
quick, because there is no load, everything is reflected energy.  On the other
hand, when you have the inductor coil connected, tuned and heating a metal
screwdriver almost orange, the snubbers run cool, because the power from the
push/pull is going to the load and not reflecting back.

Attached is the setup I assembled a couple of years ago and still use for
all sorts of projects requiring a good push/pull setup.  Works pretty well also
for a single leg pulser as long as you don't need more than 50% duty cycle.

If I was going to rework this circuit, I'd use 15 volt drive via pluggable DC-2-DC
converter for the gate and the SG3525 chip.  Other than that, this is battle
tested driver that meets all my expectations.

What on earth are you building it does the job with a simple Signal generator fed into the inductor, I'm pretty sure a 2in by 1in torrid will suffice you don't need to drive a kw into the inverter to drive the input the idea is the energy comes from the ground and the collected pull down static. plus some magic that's supposed to happen if the input into the inductor is too much the insulation will melt and the thing will catch fire, been there tried that, with the cheap cable.

[NickZ]

   So I bought some more wire to finish my new yoke coil windings
  The Free motha. Well, almost free.
  So I'll be testing the reversed windings style, like Stalkers uses.
That is, where the two yoke secondary coils, are reversed from the 12 and 12 turn primary coil.
  Yep, looking for the magic...

  Itsu:
   I've made some close up scope shots of just the two fet's drain ringing signals, expanded,  to study the frequency on each one of the two ringing part of their wave form. I'll post them a little later, so we can match up the snubber values for them.
   I'll test the new yoke without the snubbers on, first, cause my FETs are junk, anyway. According to Hoppy, And they might all be bad, but there might be a couple fets that will work, for now. Anyway, the new FETs and scope are coming soon.
Hopefully the new yoke will work more balanced, and the drains ringing signal will look more similar and flater, and having less than the 100-120v peaks that I showed previously. Then the snubbers should be able to drop those nasty peaks.