Kapanadze Cousin - DALLY FREE ENERGY

[verpies]

This is my plan on how to recover the BEMF and use it to power PWM circuit and drive Silicon Carbide mosfet which needs around 22volts at the gates."Akula version of active recovery circuit"

How are the two TL494s synchronized?
Even if the TL494s were synchronized, I noticed that according to the dot convention you are not generating a bidirectional flux in the core and the voltage induced across secondary will not be symmetrical AC.
The asymmetrical number of turns of the windings driven by the power MOSFETs will further deepen this asymmetry.  Was an asymmetrical unipolar secondary waveform your intention?

Note, that the HV spike generated when the current sourcing to both primary windings is interrupted, is generated by the primary leakage inductance, which is the inductance remaining in the primary windings after the secondary winding is shorted.

This means that the HV spike that is responsible for destroying the MOSFETs originates mainly at the primary windings.
Thus, any additional winding that wants to recover the energy stored in this primary leakage inductance, should mirror the winding that is responsible for it or the recovery will not be complete due to leakage flux bypassing the recovery winding.

If the additional recovery winding will have so many turns that it will generate much higher voltages than the V_CC then it will just form an always-loaded secondary.  Your goal should be to have current flow in the recovery winding only above a certain clamping voltage ...if you want to have another load that will be active below this clamping voltage.

[magpwr]

How are the two TL494s synchronized?
Even if the TL494s were synchronized, I noticed that according to the dot convention you are not generating a bidirectional flux in the core and the voltage induced across secondary will not be symmetrical AC.
The asymmetrical number of turns of the windings driven by the power MOSFETs will further deepen this asymmetry.  Was an asymmetrical unipolar secondary waveform your intention?

Note, that the HV spike generated when the current sourcing to both primary windings is interrupted, is generated by the primary leakage inductance, which is the inductance remaining in the primary windings after the secondary winding is shorted.

This means that the HV spike that is responsible for destroying the MOSFETs originates mainly at the primary windings.
Thus, any additional winding that wants to recover the energy stored in this primary leakage inductance, should mirror the winding that is responsible for it or the recovery will not be complete due to leakage flux bypassing the recovery winding.

If the additional recovery winding will have so many turns that it will generate much higher voltages than the V_CC then it will just form an always-loaded secondary.  Your goal should be to have current flow in the recovery winding only above a certain clamping voltage ...if you want to have another load that will be active below this voltage.

hi verpies,

It's late over here way past my bed time.

I have posted in the other topic can't recall Akula light.

The top i/c is sync via positive spike via the positive rail.It produce interrupter like pulse.
I know no such design in the internet and no one figured it out how the both i/c sync together.

At 7.50volts in out would be 10.50volts across led which acts like zener diode.
Even though it shows 0.00Amps in variable PSU but i am sure it's just below 10mA.

Time to sleep.I need energy for tomorrow.

[Jeg]

Hi guys
Hoppy, thanks for reminding me. I'll try the resistance in series to see what is going on.

Verpies
My scopeshot was without tvs. What is the best format to download pictures?

My method so far has proven to be superior and more efficient with only one inconvenience (the extra winding, which can be thinner BTW).

Yes. Lossless clamp is superior. I tried to wind one before some days but I didn't use speaker wire as Itsu and there wasn't enough space for me on the yoke. What is the less gauge of wire that we can use here so to be close to the 2.5mm2 in terms of current capacity?

[verpies]

I know no such design in the internet and no one figured it out how the both i/c sync together.

It's simple

1) Disable one TL494's internal oscillator by:
     a) Removing the capacitor connected to pin5
     b) Removing the resistor connected to pin6
     c) Connecting pin6 to pin14
2) Connect pin5 of the 1st TL494 to pin5 of the 2^nd TL494.


P.S.
I have to ask: Why are you using only one output of the TL494, which has the pins 13 & 14 connected together ?

[verpies]

My scopeshot was without tvs.

That explains the 240V.  Let's see a scopeshot with the TVS diodes installed.

What is the best format to download pictures?

PNG

What is the less gauge of wire that we can use here so to be close to the 2.5mm2 in terms of current capacity?

The answer is the ratio of the dead-times periods (when both MOSFETs are OFF) to the entire period of the cycle.  e.g: if that ratio is 1/5 then the cross sectional area of the wire can be 1/5, too.