MLCC capacitance increases after re-soldering

[Feynman]

Hi there,

On a board of mine I have a MLCC that is part of a filter (trivial RC-Filter). The bandwidth of this filter is in-circuit tested. All boards of a specific batch fail this test (the filter response is too fast). After a little debugging, everything indicated that the value of the capacitor might be too low. So I de-soldered the capacitor and measured it. The measured capacitance was fine and after re-soldering the very capacitor I de-soldered an measured, the board passes the in-circuit test. So the de- and re-soldering process increased the capacitance of the MLCC (by about 8%). At least that's the only explanation that makes sense to me right now.

I'm totally aware that MLCCs can lose capacitance with age and that this process can be reset by heating up the MLCC, e. g. re-soldering it. But the last re-flow process by my contract manufacturer was only a couple of weeks ago. So I don't expect any significant aging effects. Also other batches of the same board don't show this phenomenon.

The MLCC is a general purpose 0603 surface mount capacitor with X7R dielectric and 10% tolerance. The nominal voltage is 5x de-rated (25V capacitor in a 5V circuit).

Any ideas on what's going on here?

Greetings,
Feynman

[2N3055]

My takeaway from this is that your design is marginal, if a 8% change can make it fail.
While using MLCC with 10% tolerance and knowing that MLCC will change with age..
Not to mention X7R tempco that will go more than 8% over temperature ranges your device will normally see.

Manufacturers specify that MLCC won't be out of +-10% within first 1000 Hours...
Thats only 5  weeks...

[RoGeorge]

Could be, heating rejuvenates MLCC, you may want to read the pdf above this:
https://www.eevblog.com/forum/beginners/10uf-cap-measures-6-2-7uf-normal/msg3716068/#msg3716068

[reboots]

What's the capacitance value? C0G/NP0 dielectric would be a much better choice for stability, but it's expensive and bulky at higher values. (0.018uF is the highest Digi-Key carries in 0603, at $0.37/ea.)

[Feynman]

Nominal capacitance is 100 nF. The "failing" part doesn't bother me much, since this is not a critical parameter. In fact I already increased the tolerance of the test. As 2N3055 correctly pointed out, my tolerances where too tight, anyway.
I'm just interested in what's going on, since all previous batches where fine, even with the tight tolerance. But looks like I've underestimated the aging process and now I just got a "bad", but within specification batch.

[ANTALIFE]

Sounds like you are not using the right dielectric for the job, for example X5R & X7R capacitance is both voltage and temperature dependent
https://www.maximintegrated.com/en/design/technical-documents/tutorials/5/5527.html

If you want the capacitance to stay as close as possible to nominal over bias voltage and/or temperature, then (as others have said) NP0/C0G is your best bet

[Siwastaja]

Testing is unfruitful because even if the unit passes the test, capacitance degradation goes on and unit functionality changes on the field. So you need to come up with actual worst-case capacitance and decide if this is acceptable. Losing some 20-30% of capacitance due to aging is completely normal. If you use a physically large part with low DC bias, that effect might be just another 20-30%, then add 20% for tolerance, and your 100nF might be anything from 40nF to say 150nF. If it's a large-capacitance-in-small-package ran at more than single-digit % of voltage rating, so that DC bias effect is significant, make that from 15nF to 150nF.

Of course sometimes a simple RC filter can be acceptable even if the time constant drops to one quarter, in such cases I use X7R MLCCs as well. An example would be filtering 1MHz PWM to give some DC setpoint with reaction time in seconds.

[wraper]

So the de- and re-soldering process increased the capacitance of the MLCC (by about 8%).

Sounds about right for X7R. The most major capacitance drop happens right after the heating and slows down with time.