TPS63031: regulator failure (in many professionally reflowed boards)

Here are the schematics and the layout:

The only real difference between my layout and the recommended one is that the signal to the FB pin is routed around the output capacitors, not in the bottom layer. Maybe another small difference is that the path from pin 9 to the bypass capacitor is not direct. But I would think that this might cause a bit more noise, but not a total failure of the regulator.

Thanks, Sivan

Hi Sivan:

Thanks for asking. I afraid that you may missing a resistance between VIN and VINA. May you have a quick check? If the issue exist, we may need do further analysis.

Thanks. Do you want me to measure the resistance or to add an external one to see if this fixes the problem?

The data sheet shows a resistor inside the device, and the application schematics don't show an external one.

Hi Sivan:

Oh, sorry. I should check with you first. You choose TPS63001 or TPS63031? In your schematic, it's TPS63001. That's why I recommend to add a resistance. For TPS63031, it's not needed.  

The boards were manufactured with a TPS63031. This is what my order shows, but I will double check the device markings tomorrow. 

Sorry for posting the schematics with the other part number. I guess I changed the design and decided to use TPS63031 and did not change the schematics, only the approved part on the manufacturer's web site (CircuitHub).

I can still add a resistor there to see if that helps. 

OK, I discovered two things. One is that the bypass capacitor for VIN is connected on the ground side to the ground plane, not directly to the GND pin. I did not notice the instruction in the data sheet to connect this capacitor to GND and then to connect GND to the ground plane (sadly, this does not appear in the "PCB guideline" section of the data sheet).

I rewired two boards by hand to have the required layout but this did not help revive them.

The other thing I noticed is that on some failed boards the internal 100Ohm resistor between VIN and VINA seems to be missing (maybe it was blown up internally?) but adding one externally did not help. 

On functional boars and also on some of the failed board in which I do see the 100 Ohm resistance, the voltage drop across it is about 40mV, so the current for the control circuitry seems to be around 400uA. This is more than the Iq shown in teh data sheet, which is 25-35uA.

Hi Sivan:

Thanks for your supplement. So it's confirmed the devices are all TPS63031? How you find the issue? Just connect the power supply and find no proper output voltage or have do some test already?

May you do the A-B-A test? It means exchange the IC between functional board and failed board to see the issue is follow the IC or board.(Careful soldering). And the recover them as initial combination, repeat the test. If the issue follow the board, you may need to check the connection and soldering. If it follow the IC, we could suspect the IC is damaged already.

For the Iq, it's quiments current, not control current. It means it only test without switching. And I'll recommend a application note to you to better understand it. http://www.ti.com/lit/an/slyt412/slyt412.pdf   <IQ: What it is, what it isn’t, and how to use it>

Yes, they are all marked CEF, so they are TPS63031. 

And yes, the non-functional ones just do not show any voltage at the output.

I can try to swap an IC from a good board and replace on a bad board to see if the problem is in the board or the IC, but only in about 2 weeks.

Can the IC be damaged by the way I connected the bypass cap to a large ground plane (that includes PGND), not to the GND pin? I would imagine that this might increase noise and reduce performance, but not that it can damange the IC or completely prevent operation.

Hi Sivan:

May I know is there any updating? Thank you!