Capacitors and humidity

[Jim Starbuck]

This might be a dumb question but here goes:
I realize capacitors are humidity sensitive when shipped and stored but should we observe any special considerations after they’ve been installed in our models?
The reason I ask is because last weekend I experienced my first capacitor failure. It was in the recently completed HH660 custom build no less.
It had a bank of four 16v 220uf China cheapies in it.
The bank of 4 had been tested at 16.55 volts as that was as high as my power supply would go for an hour and a half.
The decoder tested at 13.75 volt output which should have put them at 83% capacity if my math is correct. Too high I know now.
The engine had about two hours run time on my Unitrack test loop with no problems at all. I clipped the same Zephyr to my sceniced module to make a video and with engine was on the track, as soon as I touched the power button...CRACK!... a flash of fire and smoke shot out the back of the cab.
After a change of shorts and disassembling the engine for inspection, I found one of the 4 had burned. Damage to the cab interior was minimal but this could (and should) have been much worse.
I chalked the failure up to low voltage inferior grade caps as discussed in multiple threads elsewhere.
Fast forward to today when the replacement 25v caps arrived from Mouser with what seemed like much instruction about handling and storage with regard to humidity. This got me thinking...Could humidity have played a part in the failure?
I live in Eastern Iowa where summertime humidities can run from 50 up to 100 percent when it gets extreme.  60-80 percent is common. I don’t intentionally subject my engines to extreme conditions but rather they enjoy the same conditions I try to keep for myself.
I know it’s just speculation about the cap failure because I was admittedly running them at higher than the recommended voltages. I did test the spares I had from the same 16v batch and they failed at much lower voltages around 11-13 however led spectacularly. A mild TICK and the case discolored.
I won’t use any more 16v caps but back to the original question regarding humidity and could it affect caps regardless of using higher voltage components.
Any thoughts or experience from others on this?

Thanks,
Jim

[C855B]

Uh, yup... you confirmed the advice about cheap caps the hard way. Sorry about that, chief.

Modern tantalum capactors should not be sensitive to humidity. The humidity advice from Mouser (etc.) is relevant for automated assembly and SMD soldering ovens. Stray moisture can play havoc with printed circuit board placement epoxies and soldering pastes, the reactions of both in the oven, and the quality of the solder joints. Mouser and Digikey pack moisture monitoring dots in nearly all their SMD shipments for this reason, not just capacitors.

IOW, no problem with moisture in use. That is, unless 100% humidity means running trains in a flooded basement or worse, which very unfortunately is the case for a lot of Eastern Iowa residents recently. Hope you're far enough away for that not to have affected you and yours.

[Steveruger45]

Yep, what Mike said.
I live in south east Texas where it is horribly humid all the time. 
My layout is in an un air conditioned garage too.

None of my locos have had any caps blow up in service of either type but I do a burn in test on all of them before install. That said, I’ve had a good few failures on the cheapo caps but none on the quality poly tantalum caps on a burn in test.

I never used the 16v tantalum caps but I did use the 25v ones but have since changed out all the cheap 25v tantalum caps  for quality poly tantalum 20v caps. 

BTW max recommended applied voltage to a tantalum cap is about 50% of the caps rated voltage and 80% for a poly tantalum.

[Jim Starbuck]

That’s great to know guys. Thank you!
 I’ll chalk this one up as lesson learned and tuition paid.

My town is right on the Mississippi River but fortunately we have a flood protection system of levies and walls that work extremely well. Things are pretty much back to normal around here river-wise but it’s been rain and more rain. Enough that it’s extremely close to delaying planting some crops past the window of opportunity for the growing season.

Jim

[peteski]

The humiditi precautions often seen with SMD components are not becuase they are sensitive to moisture during thei operation in whatever product they are installed.
I thas to do with the soldering process.  The epoxy encapsulant used to make their cases can absorb moisture.  When the circuit boards being assembled (by robots) go through the soldering process, it involves going throuh ovens which melt he solder paste, soldering those components to the circuit board.

These ovens heat up everything (the SMD components and the circuit board) to the solder-melting temperature for a prolonged period of time. Since that temperature is much higher than water boiling temperature, the water absorbed by the resin quickly expands, turning into steam.  This can fracture the epoxy resin. Those hairline cracks can late allow moisture, oxygen,  or other chemicals to penetrate down to the actual electronic component, shortening its life.

When you get those SMD components packed in airtight bags, they usually include a business-card size moisture indicator and instructions on how to slowly bake those components if the indicator shows that they were exposed to moisture.  he slow baking process (before running them through soldering ovens) will slowly evacuate the moisture from the epoxy encapsulant, preventing cracking during soldering.

When you hand solder these, the process only lasts few seconds and only the leads are heated. Rest of the component never gets hot enough to cause stress cracks.  So I would say that humid air is not something you need to worry about.  But as others have said, the cheap inferior Chinese tantalum caps are quite prone to self-destruct.