4
\$\begingroup\$

Recently been calibrating my solder reflow oven to work with a higher temperature solder paste. Had a few boards come out that were slightly over done.

At 250 degrees Celsius I noticed that the silk screen was starting to yellow and the electrolytic capacitors looked like they were bulging?!

Has anyone else noticed this before?

The capacitors seem to be fine, I've used the capacitance setting on my DMM to check, they all seem to have remained in tolerance. The circuit they are part of also seems to be operating perfectly normal.

Does anyone know what's happening here?

Sorry the pictures aren't any better, I've not got a very good camera:

enter image description here

Here's what they normally look like:

enter image description here

If you look closely you can see that the over cooked ones have a slightly curved top. I've since reduced the reflow temperature and reflow hold time on my oven and they don't come out like this anymore.

Here's the datasheet for the caps, unfortunately it doesn't really mention the maximum reflow temperature.

EDIT: The caps are from Farnell and should be legitimate. They were flat prior to reflow and rounded post reflow, they had not had power connected when I noticed this problem. The oven was at 250 Degrees for 120 seconds during the reflow. The temperature is monitored using a K type thermocouple (connected to a MAX6675 and Arduino) and has been verified against my multimeter's temperature probe and was found to maintain temperature to within +/-1 degree. The capacitors are connected in the right polarity according to my original circuit design that was mocked up and tested on breadboard. This phenomena has not occurred with any of the other boards or prototypes, the polarity has always remained the same between devices, it seems to just be the ones that were over heated.

\$\endgroup\$
3
  • 1
    \$\begingroup\$ Consider them cooked if you care about long term reliability. Regardless of why or what, caps should not visibly deform during reflow. [IMHO] \$\endgroup\$
    – Russell McMahon
    Feb 18, 2012 at 13:36
  • \$\begingroup\$ @RussellMcMahon Yeah, that's my thinking too. I'll just keep these as test boards now. I'm just curious as to what exactly is going on here, I had no idea this kind of thing even happens. \$\endgroup\$
    – Jim
    Feb 18, 2012 at 13:47
  • \$\begingroup\$ See this directly related paper: Robustness of Surface Mount Aluminum Electrolytic Capacitors When Subjected to Lead Free Reflow - the important conclusion is that the case distortion did NOT affect capacitor life. \$\endgroup\$
    – Russell McMahon
    Mar 1, 2012 at 7:22

2 Answers 2

2
\$\begingroup\$

See this directly related paper:
Robustness of Surface Mount Aluminum Electrolytic Capacitors When Subjected to Lead Free Reflow

The important and (to me) surprising conclusion is that the case distortion does NOT affect capacitor life. YMMV.

Reflow Sensitivity Results

  • Time to deformation strongly dependent on capacitor volume
  • Smallest and largest more susceptible to deformation

  • Moderate volume capacitors
    – 100 – 500 mm3
    – More robust

  • Some capacitors experienced deformation before the 40 second hold time defined in J-STD-020C

Conclusions

  • Case distortion mainly dependent on the volume of the capacitor

          – Small capacitors under 100 mm3 and   
          - large capacitors over 1000 mm3 susceptible to case distortion
          – Medium capacitors, less susceptible  
    
  • Case distortion did not influence capacitor life

    – No evidence of latency in any sample

  • Extended temperature or extended lifetime capacitors are more robust

\$\endgroup\$
1
  • \$\begingroup\$ Amazing,really interesting paper, cheers Russell \$\endgroup\$
    – Jim
    Mar 1, 2012 at 14:49
2
\$\begingroup\$

Those caps with the rounded tops are bad. They may act fine and appear to be within spec for now, but they are already damaged. There is no telling what they might do when. You should replace them now.

However, replacing them only solves half the problem. You also need to find out how this happened and then what to do about it. 250 degC is only 482 degF, which should be within the soldering profile of these caps. Keep in mind that temperature is only one component of the soldering profile with time also being very important. 250 degC or more for a few seconds is probably fine and expected. 250 degC for 30 minutes is way beyond the spec. If the datasheet does not specify a soldering profile, then talk to the manufacturer or use different caps that are properly specified.

A few other points

  • Are you sure these caps are genuine? Yes, counterfit caps do exist, especially in the far east. You may think of this as a gentleman's game, but there are many many dirtbags in far corners of the world that see it as a quick opportunity and are quite happy to perpetrate deliberate fraud. To them it's just business. Make sure you know exactly who you are buying these caps from, and who they buy them from. You can trust major distributors like Mouser, Digikey, Newark/Farnell, Avnet, etc. However, unless you have a personal relationship and have spent significant time there personally, you can't trust Wong's House of Discount Parts, probably called something very different when doing business in North America.

  • Are you sure the temperature really was what you think it was? Check the oven with independent measurements.

  • Are you sure the damage was thermal? Overvoltage or reverse voltage can also cause this symptom. Note that reverse voltage could be from backwards installation. Are you really sure the silkscreen is correct? Are the caps actually installed according to the polarity on shown on the silkscreen or the drawing?

\$\endgroup\$
1
  • \$\begingroup\$ I think I will have to get in touch with Farnell and see if they can get some more information from the manufacturer on the limits of the reflow temperature profile for the caps. To answer some of your questions I have provided an edit to my question including this information. Thx \$\endgroup\$
    – Jim
    Feb 18, 2012 at 15:18

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

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