TPS54335A: TPS54335A Failure

Hi Alex,

The failed part VIN pin is shorted to GND shows the failure is most likely caused by over voltage on VIN.

You said the 24V bus come up in 2us? the slew rate is 24V/2us=12V/us, it's really a very high slew rate for an input power supply, did you monitor the VIN to GND voltage by oscilloscope with full bandwidth and short probe ground?

B R
Andy

What's your load about this two part?

High output current or big output cap will result in a ring in the input(maybe more than 28V) which may will breakdown the IC.

I made a BIG mistake on my post.... The 24V Power comes up in 10 MILLISECONDS  and it is fully regulated.

U4 Sees more load driving another switcher. U2 sees a rather light load.  It drives a few Schmidt Triggers and some low current pull-ups   We do not utilize anything close to its 3 amp capability.

We used an EXISTING design inside the company that has been used for many years.  Below is 24V power Supply Stat-Up

So what's the bad rate about the TPS54335ADDAR?

U2 TPS54335ADDAR Failed twice with a Short From Vin to GND.
U4 TPS54335ADRCR has NEVER Failed. If they have the same die I think it Input Power Supply Blows out U2 is should blow out U4.
U2 Load is very SMALL on the order of 50 mA <= This may be my primary PROBLEM

U4 Load is greater, 150 mA

Both U2 & U4 are connected to the same INPUT bus (24V) with the same bypass cap 4.7uF
They both have TWO Large Output Caps 1206 X5R 22uF MURATA GRM31CR61E226KE15L
______________________________________________________________________________________
I have another low current failure of this chip as it may be oscillating (on another design). Removing 22uF caps may Solve my Problem.

Hi Alex,

Since it's easy to be repeated, i think this is a application issue. This damage may caused by over voltage on VIN pin. You said that those two device were sharing the same bypass cap, this is not recomended. We have to put the input cap near to the device VIN/GND pin. We had similar test before, the device would be easy damage if the input cap is far away from the device. You should capture the spike voltage directly on the VIN/GND pin of the device, not on the bypass cap, and there should have obvious spike on VIN pin to GND pin.

BRs,

Ryan   

Ryan,



Sorry for any confusion on my part.
I appreciate any ideas anyone can give me.
-----------------------------------------------------------------------------------------
Review:
U2 TPS54335ADDAR Failed twice with a Short From Vin to GND.
U4 TPS54335ADRCR has NEVER Failed.
----------------------------------------------------------------------------------------

U2 and U4 each have their own 4.7uF Cap
U2 has C93 it is 4.7uF Right Underneath connected with vias
U4 has C24 with the same 4.7uF connected on top via copper pour.

If U2 blows up twice I still scratch my head why U4 is always fine. Meaning if there were a huge spike on the 24V Regulated Supply I would think U2 AND U4 would blow. The 24V bus is regulated (flyback switcher)


The Schematic for both designs was copied from existing company designs with 1000's of hours on them. (primary reason for doing this was to avoid the problem I have here) (I performed the qualifying tests weeks ago... I beat up the power supplies and I had ZERO issues.)

I have no sure way to make U2 fail. Out of three beta tests in the field one came back with U2 shorted.

Steve

-------------------------
The other design that failed was oscillation due to wrong output capacitor. That problem was resolved and it functions properly.

We did try putting 200uF ceramic caps on the output of that chip and we could not get any oscillation. The 5V supply on this design drives 50 mA as the one I am having issues with on the main design.