Hi All.

This is the second rev of this board. First rev didn't have this problem whatsoever. After months of use and abuse.

The second rev didn't change this circuit. So far, 3 out of 5 boards that i have brought up had one or more of these capacitors shorted out.

The pmic is: TPS65931211RWERQ1
The caps are: CL03A225MP3CRNC
So far the shorted caps have been on pins 2 and 3, VOUT_LDOVINT and VOUT_LDOVRTC respectively. These are 10V rated 2.2uF caps and the 1.8V LDO's are for internal loading only. These caps are the only thing connected to each pin.

My thoughts are:
Damage from assembly: While possible, I wonder why it's (so far) these 2 particular ref des.
Bridging: i didn't notice any. And when i pulled the caps off and cleaned them, they still measured about 2-8 ohm of resistance.
Bad caps: but then why so far on the three bad boards are they on the same pin. this cap is used in about 25 other places on the board.
PMIC damaging caps: Somehow the pmic internal LDO is damaging the caps. this seems unlikely, but possible.

I will be checking more samples today.
Including testing the caps before turning on for the first time.

 

Please google MLCC flex cracks. Prepare to be amazed.
Suffice to say that at a company I worked for at the turn of the millennium, these caused a recall costing tons of $.

 

I was unable to determine what current frows in C39 & C39. I do see that TI calls out a 0603 while you used a 0201.
If this was in the switching part of the IC, I would worry about the current carrying ability of a C0201.
I don't know. Just a thought.

 

I was unable to determine what current frows in C39 & C39. I do see that TI calls out a 0603 while you used a 0201.
If this was in the switching part of the IC, I would worry about the current carrying ability of a C0201.
I don't know. Just a thought.
View attachment 342807

interesting. These are just LDO, no switching. I’ll look more into this though. Thank you for that info.

 

Please google MLCC flex cracks. Prepare to be amazed.
Suffice to say that at a company I worked for at the turn of the millennium, these caused a recall costing tons of $.

Thanks for the insight! I’ll dive into this and see if it could apply

 

https://ntrs.nasa.gov/api/citations/20190001592/downloads/20190001592.pdf
Cracking Problems in Low-Voltage Chip Ceramic Capacitors

Cracking remains the major reason of failures in multilayer ceramic capacitors (MLCCs) used in space electronics. Due to a tight quality control of space-grade components, the probability that as manufactured capacitors have cracks is relatively low, and cracking is often occurs during assembly, handling and the following testing of the systems. Majority of capacitors with cracks are revealed during the integration and testing period, but although extremely rarely, defective parts remain undetected and result in failures during the mission. Manual soldering and rework that are often used during low volume production of circuit boards for space aggravate this situation. Although failures of MLCCs are often attributed to the post-manufacturing stresses, in many cases they are due to a combination of certain deviations in the manufacturing processes that result in hidden defects in the parts and excessive stresses during assembly and use. This report gives an overview of design, manufacturing and testing processes of MLCCs focusing on elements related to cracking problems. The existing and new screening and qualification procedures and techniques are briefly described and assessed by their effectiveness in revealing cracks. The capability of different test methods to simulate stresses resulting in cracking, mechanisms of failures in capacitors with cracks, and possible methods of selecting capacitors the most robust to manual soldering stresses are discussed.

 

Quite a few years back, thru an error, I did succeed in creating a high power oscillator from a three terminal 12 volt regulator IC. I did not fail any capacitors because I switched the power off quickly because of the wrong voltage.
My point being that an inadequate bypass capacitor installation can certainly provoke a high power oscillation. AND that might lead to an internal overheating failure of the low voltage rated capacitors.
SO I am suggesting an evaluation of the mandated 0.1 mfd bypass caps as the source of the failures.