PCBA contamination identification

[scoreb]

I have recently gotten some field returns that have high impendence shorts under some of the .1uf 1206 caps. The assemblies pass our end of line test but fail in the field under high humidity. I suspect contamination left over from the assembly process is the reason for failure, but need to prove it.
Does anyone know of a lab that can identify the conaminants found on a PCBA? Plenty of places can tell me how much contamination there is, but not what it is. If it's salt from the atmospheric conditions, that's one thing. If it's bromine left over from the manufacturing process, that's another.

Thanks,
Scott

 

[melone]

How high of a short? Could it be your caps? 1206 is a pretty big device to be shorting. Do you have any type of coating that covers your circuit? What are the typical atmospheric conditions that your circuit is subject too?

 

[Lewish]

I had problem similar to this about 10 years ago. With over 4000 smt parts on a board, it was a real show stopper.

The contamination turned out to be the adhesive used to hold the parts in place before reflow. The adhesive was being degraded by the board wash solution, but didn't show up for a while. Changed adhesives and the problem went away.

And in my case, it was also the caps that were the problem. At operating frequencies the contamination made them look like low value resistors.

Unfortunately, I don't know of any lab to suggest.

 

[IRstuff]

Most failure analysis labs have the capability to perform various analyses such as x-ray emission or Auger.

If the boards are not coated, there may be other things such as solder whiskers and the like. TTFN

 

[scoreb]

The shorts range in value from 60 ohms to >200k. The boards are not coated and have components on one side only, no adhesive.
The failures occur after subjecting the boards to 90F, 90% RH for 24 hrs. The boards are ramped from abient to 90, 90 in 30 minutes to avoid condensation.
In all cases removing the suspect capacitor, which reads in spec in ckt and out, eliminates the failure. Hot air is used to remove the component.
Alpha Metals WS609 solder paste is used.

 

[IRstuff]

90/90 is generally one of the promoters for solder or tin whiskers. Maybe there's a bad batch of tin plate on the leads. TTFN

 

[designchain]

Evaluation Engineering has a list of F/A labs. Check it out at

http://www.evaluationengineering.com/cgi-bin/bg/bgcat.asp?Category=9&CategoryName=Test+Labs/Services

Do you know what kind of wash (if any) is used after assembly? Do you have the rest of the reel that these caps were taken from (i.e. unused caps)? The lab could use those in the analysis.

 

[jbartos]

Suggestion: Visit

http://www.finishing.com/118/03.html

etc. for more info

 

[CarlPugh]

My client built some meters with zero problems.
You could breathe on the PCB and the meters would peg.
The problem was that water soluble flux was used and the PCB was not cleaned properly.
After proper PCB cleaning the meters were OK.
Other fluxes have the same problem, although probably not to this extent.
Heating the PCB to remove the capacitor will dry out the flux and the board will appear OK.
Robert A Pease (author of Troubleshooting Analog Circuits)has written in some of his articles about breating on PCB to increase leakage currents.
Good Luck