UCC28951-Q1: PINS Vdd & Grnd KEEPS GETTING DESTROYED

I AM ADDING A PHOTO OF WHAT THE IC LOOKS LIKE.

ED

Hi Eddie,

I have contacted one of colleagues to answer your questions.

Regards,
Teng

Hi Eddie

The first thing I would say is that the UCC28951-Q1 and its related parts are in widespread use for high power DC/DC converter applications similar to yours so I am certain that the device itself is not faulty and there is not need for you to abandon it and look for other solutions.

However, I realise that simply saying that doesn't fix your problem.
I had a look at your schematic and it looks ok to me.
I have also in the past and as a test applied up to 35V from VCC to GND on this IC and the part survived ! (Abs Max is 20V) and I have never seen burn damage like you have seen.

So - I would suspect that there is something wrong with the VCC rail going to the device.
I'd suggest the following

Disconnect everything from the bias source you are using for the 12V VCC rail. Then monitor the output voltage of the source as it turns on to see if there is an overshoot or undershoot. Use a 'scope and look at the startup event on timebases from us to seconds
Then.
Connect one of the blank PCBs and repeat the measurement but this time look at the voltage on pin 23 (VCC) and 24 (GND) of the PCB traces.

Then, solder in one of the UCC28951-Q1 devices onto the blank PCB and repeat the process.

If it all looks good - then go back to your product's PCB and remove the controller. Apply 12V to the bias rail and check that none of the other pins around the device are being pulled to 12V - just in case there is a short circuit somewhere.

If the problem is still there then can you arrange to send two unused devices to me ?

Regards
Colin

got it, will let you know the results. thank you for your response.

ONE OTHER QUESTION

 THIS LAST IC TESTED WAS WITH NO COMPONENTS INSTALLED BUT WITH Vref Pin #1 FLOATING. I WAS ABLE TO SLOWLY INCREASE THE VOLTAGE TO PIN 23 & 24 TO A MAXIMUM OF 10V THEN SMALL ARCING, SMOKE PUFF THEN DOA.

 WHAT I DID NOT EXPLAIN IS THAT I APPLIED VOLTAGE TO THE IC WITH A VARIABLE VOLTAGE SUPPLY. STARTED AT 2V THEN SLOWLY INCREASED THE VOLTAGE WHILE MONITORING Vref TO Grnd. CAPACITORS WERE PLACED ON (Vcc/Grnd) AND (Vref/Grnd) TO STABILIZE ANY RIPPLE/FLUCTUATION IN VOLTAGE.  NO THER PINS HAD ANY COMPONENTS CONNECTED AND THIS WAS PERFORED ON A BREADBOARD WITH A VARIABLE SUPPLY.

 AS THE VOLTAGE AT Vdd/Grnd WAS SLOWLY INCREASED AND THE INUT VOLTAGE WAS INCREASED TO ABOUT 7.43V AND THEN THE Vref JUMPED UP TO 6.8V. THEN AS IT Vdd WAS INCREASED TO ABOUT 7.4V Vref WAS AT 7+V. AT 10V THE IC STOPPED WORKING. THE CURRENT JUMPED UP FROM 80Ma to over 140Ma. This is way too much current even at 80mA.

THIS IS A CONCERN/PROBLEM. Vref SHOULD NOT GO ABOVE 5V+-1.2%.

Question -- Why Is This This Happening. Please Explain.

IF THE VOLTAGE AT Vref IS THIS HIGH, DOESN’T THIS MEAN THAT INTERNALLY THE IC IS ALSO AT THE SAME HIGH VOLTAGE AND THIS WOULD ALSO MEAN THAT MORE CURRENT IS REQUIRED?

ED

Hi Eddie

None of this makes any sense. You should be able to put 12V from VCC to GND and the IC should work correctly. VREF should be at 5V. The only thing that bears any resemblance to correct operation is that the VREF becomes active, although at the wrong voltage, when VCC passes 7.43V which is in the UVLO range of the IC. I have never seen one of these ICs produce smoke.

Please contact your distributor and make arrangements to have some of your devices (some used and some unused) back to TI for failure analysis.

I would also like if you can send two unused devices back to me so that I can do an application test here in the lab.

Please also send me details of the full marking on the devices - top surface and bottom surfaces 

Regards
Colin