1
\$\begingroup\$

We have a board built in the thousands, and several have (apparently) short-circuited a plated through hole (PTH) to an adjacent pour (see photo, note the other thermal damage is probably due to thermal conduction of via internal trace running to the PTH).

The PTH to pour clearance is 7mil, and the PCB manufacturer and assembler is big name CM!

Any ideas about possible causes or lines of inquiry to pursue?

---followups

  • This problem has occured in ~0.05% percent of the builds.
  • the damaged PTH is where the short is suspected to an INTERNAL pour.
  • I had assumed that 100% of these PCBs were electrically tested. perhaps I should not assume.
  • the applied signals are from a <50V battery which is rated for ~20A, but if we assume ~3kv / mil dielectric withstand we are nowhere near a breakdown even with some voltage ringing given our 7mil clearance.
  • Failures are found during sub assembly testing in a pretty organized / clean facility, and occurs upon connection of power, not during use.

enter image description here

\$\endgroup\$
  • 1
    \$\begingroup\$ You are going to have to be more specific. What are the input voltages for example. Where are the cables connected. \$\endgroup\$ – user94729 Oct 10 '18 at 5:50
  • 1
    \$\begingroup\$ Inspect one that has the short and didn't burn yet. Also, why doesn't the production testing check for it? \$\endgroup\$ – PlasmaHH Oct 10 '18 at 6:48
  • 1
    \$\begingroup\$ Which PTH is the one that's supposedly shorted to the pour? The undamaged one on the right-hand side has thermal reliefs, suggesting it is supposed to be connected to the pour. Same with the one at the top right. \$\endgroup\$ – DerStrom8 Oct 10 '18 at 10:32
  • \$\begingroup\$ if it is a PCB issue: how about showing as a board that's not burnt as well? How many layers? board file / Gerbers, or at least an extract of it around the trouble area? Given the damage, it must be a pretty solid connection, we should be able to see something. BUT it might not be the PCB: I have encountered a similar issue once, in my case, it was caused by too much solder shunting two pins of a connector between the board and the connector body, nothing to do with PCB fab. \$\endgroup\$ – MAB Oct 10 '18 at 17:11
  • \$\begingroup\$ also 0.05% is 1 in 2,000 - how many faulty boards have you actually found? If the fault is easy to find during sub-assembly testing, is it worth remediating? \$\endgroup\$ – MAB Oct 10 '18 at 17:15
1
\$\begingroup\$

On such wide question I can only give research questions.

  1. Validate if the board layout is correct.

    • Check if the continuous voltage present provides enough clearance and creepage.
    • Check for any surges or spikes, the re-validate the the board design.
    • Check for current rating and track/via within normal conditions.
    • Check for current ratings in abnormal conditions. (add fuses is necessary)
  2. Validate the Assembly is correct.

    • Are there lots of tiny solder balls after soldering? Improve reflow/solder process.
    • Is there manual soldering used? Too hot, too much, leftover flux of bad kind?
  3. I can't see if there used to be a component where the board has vanished, what if this component is abnormal or is used incorrectly?

  4. Is the operating environment clean (pollution degree) and how is moisture?

\$\endgroup\$
0
\$\begingroup\$

There could be registration error in the soldermask, exposing part of the outer pour. Then solder from the positive terminal could bridge to it.

Do you have evidence pointing to an short with an inner-layer pour rather than the outer layer?

\$\endgroup\$
0
\$\begingroup\$

I had assumed that 100% of these PCBs were electrically tested. perhaps I should not assume.

Even if they were electrically tested it was probablly at a relatively low voltage and before soldering.

The PTH to pour clearance is 7mil

If the clearance is directly from the hole to the pour (and not from a pad surrounding the hole to the pour) you should be aware that the specified hole sizes in PCBs are finished hole size, the size of a drill hole and hence the outside of the plating cylinder will be larger.

Personally I would back off the clearance for pours, I understand sometimes things get tight but using your minimum clearance everywhere seems to be just asking for trouble.

\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.