TPSM63606: Short between VIN, VOUT and PGND

Hi Jan,

I believe we need more information. Maybe you can send me schematics of the relevant PCB's. Also, for clarity can you draw a diagram of how each thing is connected at each step?

Thanks,
Andrew

Hi Andrew,

thanks for your quicj reply.

Here are the requested schematcis:

Schaltplan_lp00799-01.PDF

lp00803_outjob.PDF

And here is a descciption of how to reproduce the short:

Please let me know if you need further information.

Best Regards

Jan

PCB1 is lp00799

PCB2 is lp00803 but not all of it. This PCB is cut into to part. Only the part with the buttons and leds are relevant.

Hi Jan,

Do you think there's anything downstream of TPSM63606 that's giving you grief? Is this the only part that's giving you trouble. Do those steps reproduce the broken part 100% of the time?

Thanks,

Andrew

Hi Andrew,

there is nothing else connected to the TPSM63606 than the parts on PCB1. There is a CPU Module in the SODIMM socket but that all works fine as long as I do not unplug and replug the cable.

I do not know when exactly the TPSM63606 goes short. It may be when I unplug the ribbon cable or when I reconnect it.

I have had this issue with three out of five boards so far. I do not dare to test it on the last two boards because I need them for product development. As long as all is connected together it all works fine. I did a load test on the TPSM63606 and they deliver 6A with no problem. I only have trouble when I have to change the setup and unplug replug the ribbon cable.

Well, my initial question was what event would make the Fets in the outputstage of the TPSM63606 go short?

Any idea about that? What do I have to look for if I repeat the test with plugging and unplugging? What pins should I monitor?

Hi Jan,

Is there an over voltage event happening? That's definitely something that can blow out the FETS. I don't see any Voltage clamping mechanism (like a zener diode) at the input of the TPSM63606 or at PCB1 after the power supply is connected.

Thanks,
Andrew

Dear Andrew,

Under- and overvoltage protection is implemented by U24 (LTC4368) .

U24 does not open FETs Q1 and Q8 if ViinRaw is below 20V or above 28V and so the TPSM63606 do not see more than 28V.

I tested this by applying up to 64V to P1 (on a board where we removed the shorted TPSM63606) with a bench power supply and measured the voltage at Q1 Pin D and Q1 Pin S.

While 64V is at Pin D there are only about 0,2 V at Pin S.

When I turn the bench supply below 26.5 V  Q1 opens and the (remaing 4) TPSM63606 are still alive.

So I think it Overvoltage Protection works or do you think I need to add Zener diodes?

I also measured the voltage at C32 (VinFlt+3V3) with a 500 MHz scope and I do not see any voltage more then 24 Volt if I plug powersupply at P1 or ribbon cable at J5.

Hi Jan,

Can you share your layout for the TPSM63606 area?

Thanks,

Andrew

Dear Andrew,

attached you find the gerber files for PCB1.

lp00799.zip

Best Regards

Jan

Hi Andrew,

do you have any suggestions what do to make the design saver?

Best Regards

Jan

Hi Jan,

I'm just thinking out loud here. Is there any power on PCB 2 when it's connected back to PCB1? I'm just wondering if the VLDO_in pin is causing our device to be damaged.

Do you have any spare TPSM63606's to rework onto broken boards on to test anything. I'd like to see VIN/VOUT/SW/VCC voltages when the device pops and during your procedure.

By the way I checked the layout. It seems fine, although there are some 0 ohm/1 ohm resistors on the EVM that make it easy to disconnect various pins from other pins that would be nice if they existed. Mainly one for VLDO_in and CBOOT.

Thanks,

Andrew