Low Frequency Pure Sine Wave Inverter Repair Help

[Keg]

I have a couple Inverters that a power surge or lightning took out during a storm.  The control-driver boards are what I'm having issues with repairing.  The H bridge is not a problem.  No schematics available making it more difficult.  Luckily, I've purchased a new control board to compare resistance and voltage readings.  I'd love to be able to repair the damaged boards for future backups.  New ones are not cheap!

They are a 24v, 6kw, 240v split phase and a 24v, 4kw, 120v. Both of the control boards are basically identical.  There is no noticeable physical damage on either. 

The main issue-question I'm having is with the supplies for the op amps.  The 24v comes into the board through a diode and then appears to go to a small pcb mount transformer.  Kinda odd that the op amps are supplied with +12.6v and -9.5v.  There is a 5v supply that only powers the 14011B Nand.  The 7805 is supplied with the 12.6v rail.

The +12.6v appears on the ER806 diode.  The -9.5v appears on the UF204 diode.  Shown in the one photo.  The +12.6v supplies the 7805 regulator.

I initially thought that the positive and negative rails were created by the tranformer controlled by the 3843B PWM controller and the IRF640N mosfet.  The output on the 3843B does go to the gate of the mosfet.  It's powered though by the 12.6v rail which would not be there on power up.

My my question right now is understanding how the positive and negative rails are created. 
Anyone out there could explain this?

Also, why would they do it this way with an uneven positive and negative rail?

Thanks, Ken

[Keg]

Just a heads up.  The voltages that I have on my initial post were taken from the new-working control board.

Ken

[Kim Christensen]

A lot of these PWM ICs like the 3843B bootstrap themselves as shown on page 16 of the datasheet.
Upon power on, Rstart will charge Cvcc until the voltage on the chip reaches a threshold which wakes the chip from it's slumber. Once the IC starts switching, it self powers itself from the Na winding via Dbias.

[Keg]

Sounds logical.  I'll check out that page on the datasheet and check the circuit.  I don't have the voltages on the one board.  Possibly it's that capacitor or the chip is bad.  If I locate the cap I could see if it's charging on power up.

Thanks, Ken

[Keg]

Great find Kim,

That circuit is almost identical other than a few more pins on the transformer.  Doing some calculations, it would be dropping 13v across Rstart which at 1ma would be around a 13k resistor.  I'll be able to trace out the circuit on the board and see why the IC isn't starting. 

Thanks again, Ken

[Harry_22]

Hi! At start point 5 volts should appear at Vref pin.

[Keg]

Thanks Harry,

I'm going to start testing it this afternoon.  I'll check for the 5v also. 

Ken