Kapanadze Cousin - DALLY FREE ENERGY

[magpwr]

But none of them explain the asymmetry between the falling and rising edge.  See the green and red arrows below:
The data sheet lists the input capacitance at 5.2nF
With 1.8Ω gate resistor the RC time constant is 9.4ns so the 99.3% rise & fall time should be 47ns.
Because this capacitance is the same during the rise and fall times, they should be the same, but they are not.

Note, that the scopeshot below was taken at 20μs/div which is 426 times slower than the calculated rise time!
Itsu's gate signal has a short rise time as evidenced by the short fall time of his drain voltage.  It is the rise time of his drain voltage which is the problem.  This drain rise time should be even shorter than the fall time because he has a BAT42 speed-up diode in parallel with his gate resistor ...but the opposite happens.

This asymmetry should not happen with a resistive load ...unless his snubbers make the load asymmetrical and not so resistive.

hi Verpies,

If the curve is detectable in the us/div setting it may also means the supply capacitance isn't high enough for quick recovery coupled with center toroid may make it worst.

10000uf 50volts  was implemented in my latest youtube posted previously to the small toroid 1 turn 44uH (+) to toroid center tap.
But it was not mentioned since the main focus was on PWM circuit.

There is notable difference between 1000uf and 10000uf capacitor related to output.

[verpies]

Hoppy,
If the 2*V_CC on the drain is (2*24V) 48V, then a 54V TVS across it would normally be alright was my thinking, it will/should cut off anything above 48V (spikes).

That theory is correct but I think Hoppy was worrying about the practical behavior of the TVS diodes.
e.g. what about their breakdown voltage tolerances or junction capacitances ?

[Hoppy]

Hoppy,

If the 2*VCC on the drain is (2*24V) 48V, then a 54V TVS across it would normally be alright was my thinking, it will/should cut off anything above 48V (spikes).


I have 0.1uF decoupling caps directly across the drivers + / -  connections, also a 47uF tantalium cap for power boost is across it.

Regards Itsu

Itsu,

Is the 54V a VWM - Rated stand-off voltage, VBR - breakdown voltage, or VC - clamping voltage? The clamping voltage is typically 30 - 40% above VBR, so if the VBR of your diode is 54V, it would need at least 70V to clamp a spike but could still conduct at below 54V, as the maximum operating voltage determines the VWM which should be around 15% lower than VBR which is about 46V. So given that the max operating voltage could be greater than 48V, the TVS diodes may be conducting to some extent, assuming the 54V is the VBR rating.

Tha BAT 46's have a VRRM of 100V, so I think these are safe.

[verpies]

If the curve is detectable in the us/div setting it may also means the supply capacitance isn't high enough for quick recovery coupled with center toroid may make it worst.

As far as I know, the choke is before the power supply bypass cap and his MOSFETs were not driving anything inductive when that asymmetric scopeshot was taken.

Anyway, a scope probe placed on that cap will quickly discover any supply voltage sagging.

[itsu]

Hoppy,  verpies,

the 54V TVS specs were posted here:

http://overunity.com/12736/kapanadze-cousin-dally-free-energy/msg474113/#msg474113

I notice that they worked fine when directly across the drain / source of each MOSFET, but yesterday i had moved them directly across the primary coils,
but then the setup started to oscillate when more then 8V on the drains, so i have removed them now.


Itsu