Efficiency is the problem with those kinds of efforts.
I have found a better way... not what I wanted to do... but the animal in me got out.... ( and I was lazy and could not be bothered with jamming in an scr to shut it off on overload..... other wise it retries too fast using the temp contact without a latch)
Here is the relevant picture. ( click to see properly )
I got the razor blade out and removed the 393... shorted pins 7 to 8 and then made a jumper from 1 to 4.
This gets the inhibit pins of both the drivers and the main chip to turn off (/ or on depending how you want to say it I guess) ... unequivocally.... ie they no longer get to inhibit anymore.
Then use the current IFB input as designed. It then behaves with no problems.
I cant understand why they used the 393 in the first place, unless the IFB signal from their resistive sense was too small, but they needed to get the timing correct so that the IFB pin on the main ship got the message first, but that is not the case. They decided to inhibit in three places at the same time..... and failed miserably. Not noticed with EI transformers ( but for odd noises), but lethal for torroids..... now it is fine, and I can use the current o/load in the chip to do the stop... test.... stop ... test etc, then turn off needing hard start.
The current transformer ( CT) has plenty of output, and needs attenuation and a small filter to behave nicely.
This is using the board at the start of this thread.
I still prefer the temp TFB pin for on and off at this stage.
Now it seems I can try all I might, but cannot fool it into blowing the fets.... up to 2.5kw at least.
2kw ( 2 fets/leg) and it runs for quite a while before it starts to warm up much at all ( 10 mins), and thats with no other cooling... very useable now.
Will still make the new board when I get to it, just so I can go full power to 8kw or more.... we'll see when I get some time.
Will post some pics when I can get the camera to talk to the computer again...... something squiffy at the present time.
It does look like a very simple and cheap way of getting a very solid little inverter.... for pennies really.
This is the first draft of what the board will look like not using the pre-made board, but with the three isolated supplies, opto's and chip... still have to put on the CT and voltage feedback stuff..... should fit......
[ Specified attachment is not available ]
................oztules