Kapanadze Cousin - DALLY FREE ENERGY

[Jeg]

Yes, and the C3 and C4 capacitors need not be rated 160V.  Mere 50V would be enough and lower voltage caps have better ESR, too.


I agree with Itsu about the the MOSFETs' wire layout.  Also, I hope you are using this schematic to connect the W1 windings to C3, C4, D3, D4, GND and V_CC.  The EMI chokes L1, L2, L3 before the filter caps are a very good idea, too.
I am pretty sure that if you had copied  Itsu's transistor and driver connections layout, then you'd be rid of these spikes.

Thanks Itsu, Verpies
I'll try to improve layout and make the changes. What do you think about using insulated core mesh wire for driving gates? Grounded only at one side. 

[itsu]

Thanks Itsu, Verpies
I'll try to improve layout and make the changes. What do you think about using insulated core mesh wire for driving gates? Grounded only at one side.

Jeg,

no idea what insulated core mesh wire is, but i use some RG316 coax cable between the TL494 output to the MOSFET driver input but grounded at both sides.
I don't think its generally a good idea to ground only 1 side, you might create some resonance points that way.

Again, my MOSFET drivers are directly soldered with their (2) ground lugs (bend backwards) to the MOSFET source leg to minimize the loop there.

Itsu

[itsu]

You are not wrong and neither is Itsu.
It should be 48V in reference to ground.
Yes, and the C3 and C4 capacitors need not be rated 160V.  Mere 50V would be enough and lower voltage caps have better ESR, too.


I agree with Itsu about the the MOSFETs' wire layout.  Also, I hope you are using this schematic to connect the W1 windings to C3, C4, D3, D4, GND and V_CC.  The EMI chokes L1, L2, L3 before the filter caps are a very good idea, too.
I am pretty sure that if you had copied  Itsu's transistor and driver connections layout, then you'd be rid of these spikes.

Confirmed, its a straight 24V DC across the + / - of the electrolytic cap, and a 48V square wave across the + of the electrolytic cap and ground.

Itsu

[Jeg]

Jeg,

no idea what insulated core mesh wire is, but i use some RG316 coax cable between the TL494 output to the MOSFET driver input but grounded at both sides.
I don't think its generally a good idea to ground only 1 side, you might create some resonance points that way.

Again, my MOSFET drivers are directly soldered with their (2) ground lugs (bend backwards) to the MOSFET source leg to minimize the loop there.

Itsu

Insulated core mesh wire =  Coax cable (Sorry for my English)
The distance between 494 and my drivers is close to nothing. I'll use some coax between IR2110 outputs up to my mosfet gates which are long wires. I thought that by soldering both ends of ground shield, a ground loop might be appeared. I'll test it anyway.

You did a good work on minimizing this loop. It is a pity i can't proceed this way with my current experimental layout. But it is always a good idea for my next push pull build.

Thanks

[verpies]

Insulated core mesh wire =  Coax cable (Sorry for my English)
The distance between 494 and my drivers is close to nothing. I'll use some coax between IR2110 outputs up to my mosfet gates which are long wires. I thought that by soldering both ends of ground shield, a ground loop might be appeared. I'll test it anyway.

You really should read this article before you start.

There is a special section 12.10.1 there about sending signals down coaxial cables, but I recommend reading the other sections, too.

Whatever you do, your first order of the day should be modifying the circuit layout and its cabling to stop these spikes from appearing at your gates.