ceramic cap crack

Please describe the location and appearance of the crack

reply »

We also encountered the crack case in multi layer Ceramic Cap,0603 after touch up. Most failure were on the terminals. Previously, we used 0805 instead 0603.that time, failure were very few.

reply »

check your de-panelling process if you are using break-out tabs or V scoring.

reply »

A couple of things:

Does the cap break if you try to remove it with hot air? If not, than you are breaking the caps do to mechanical stress. We have had similar issues. I don't think your board is suffering from mechanical stress from depaneling or other processes (unless they are all grouped together, perpendicual to the score lines, and near the edge of the board). If they are, then check your depaneling process. If the cap is cracked, I'm surprised that your ICT even reads any capacitance, much less up to 350pf. It sounds like you are using a cap with an X5R dielectric. These have a high aging factor (7% per decade) and wide tolerance. These are the scavangers of the capacitor world, bottom feeders, and there is a reason they are so cheap...because they are. Go with X7R, heck, Y5V would even be better. I have been forced to use X5R in production and we have tolerance issues with all three. I believe two are Kemet and one is Panasonic so it is not an issue with the mfg, but with the dielectric of the part.

Try this experiment: Gentley heat each side of the cap with an iron, but do not remove it. Put it back on the ICT and check the measurement. If its an X5R dielectric, the value of the cap will increase to nominal and stay there for another 6-12 months before it deages again (it deages logarithmically starting at 1000 hours after the deaging process). What you have just done is deaged the capacitor by realigning the crytalline structure within the dielectric and this is the recommended method from the mfg, if you can believe it. Tell your purchasing department to stay away from X5R, unless quality is not a concern.

reply »