UCC28782: ucc28782-converter failures-2

Hello Robin, 

I'm sorry to hear about these unexplained failures, and the inability to restart.  It is good that considerable prior operation has proven the general circuit design is sound. 

Despite all of your replacements and open-loop testing, the new controller apparently still detects some problem that prevents it from fully powering up.  Perhaps some other component was also previously damaged not not detected and is still in place.  Or, maybe a new problem has been introduced in the final stage of rebuild. 

In order to provide better help, we need to clarify some vague or broad meanings in some of your descriptions above, especially lacking a schematic for reference. 
"...PGND cap arced..."   Does this refer to the specific PGND pin on the UCC28782, or to a general "primary-gnd" net name on the board?  What is the purpose of the cap; where else is it connected? Which end of the cap was being probed and where was the GND-reference of the HV probe connected?  Was the HV probe being actively applied to the "PGND cap" while the converter was running, or was it hooked up prior to power up? What other nodes are close to the probe point the may have been inadvertently bridged during (static or active) probing?

"... still won't restart."  What specifically happens during the restart attempts?  Nothing; no signal changes, no activity at all?  VDD goes up and down between thresholds but no PWML pulses?  PWML pulses, but quits after a few and VDD always cycles?  Partial series of power pulses and Vout rises part way up but pulses stop before regulation could be achieved?  The debug process to follow depends on which of these symptoms is true.   

This really does need a debug as if it were a brand new board, because it is not behaving as expected per the proven schematic design.  Something on the real board is not matching the schematic.  Possibly an open pin from a poor solder joint, or a hidden solder bridge between pads, or an undetected failed part that "looks" good.  Verify that the controller was installed in the correct orientation, not rotated. 

 Assuming that there is only one problem blocking full start-up, we start at the last point where the normal sequence succeeded and postulate what could be preventing execution of the next step in the start-up sequence.  With sufficient information about the system conditions and signal levels or states, it needn't be an exhausting debug effort.

Regards,
Ulrich

Hello Ulrich:

Once you are in the scene,, we feel we will get to the bottom of this very soon...LOL

Let me digest what you wrote & give appropriate information.\thnx much

r