I have faced two capacitor failures in a product that I had to call back from field.
The product had exposure to moisture and I removed all components only to conclude that caps failed, below are two different types of failures
Ceramic caps were found to have a lower resistance between the terminals than what they should have. When tested with multimeter, their resistance settled to a very low value. The device was powered by CR2032 which went dead as a result
These small sized caps are known to fail with soldering related heat stress over time. Hand soldering is a given and there can be no guarantee that the soldering is done properly. The stress is not visibly seen and failure does not happen soon enough to be caught so I am looking for a way to deal with moisture while the device is in production floor.
Can waterproofing by potting avoid this failure?
Electrolytic caps also showed lower resistance between the terminals. These were "VENT" branded caps. New caps dont have this problem, these are leaded parts and quite big in size. Its surprising that improper soldering could damage them. Again, I feel its the moisture that led to the failure, though I plan to change to a better brand of cap than this, given that this brand has failures reported by many on the net.
Can waterproofing by potting avoid this failure?
wonderfuliot:
...I have faced two capacitor failures in a product that I had to call back from field....
....These small sized caps are known to fail with soldering related heat stress over time. Hand soldering is a given and there can be no guarantee that the soldering is done properly....
Then can you specify a minimum length of lead when soldering, so that the hot bit is further away from the sensitive part of the capacitor?
Of course potting can help to eliminate water ingress to the already damaged capacitor, but isn't it better to avoid the damage first? If you have no control over the quality of the 'product' that you are producing (selling?) then look at your method of production too.
wonderfuliot:
...Its surprising that improper soldering could damage them...
It's no surprise to me, and probably to most people who know how a capacitor is constructed. Again, take steps to avoid the damage rather than secondary actions such as potting.
Hi,
Are the ceramic caps surface mount type?
If so then they do fail more often than leaded through hole types.
This is due to their construction, condensation/corrosion will have an effect.
With electrolytic caps, brand name does have make a difference, also ESR.
If your electro's are being used in a high frequency SMPS type circuit, you may need to specify LOW ESR capacitors.
How close you run the DC voltage on them, compared to the rated voltage can also have an effect.
You don't have to deal with capacitor failures, you have to deal with moisture ingress - this is the cause and
needs fixing. Any water-damaged electronics should be disposed of responsibly, electrolytic action will etch away pcb traces and component leads and leave corrosion deposits that contunue to rot over time. Electrolytic action can cause both acid and alkali corrosion once there's moisture and voltage present, nothing will withstand this over time (especially acid corrosion), the only way is to prevent moisure ingress and condensation.
'VENT' is not a brand of capacitors. It's stating that the can has indentations in the top that will safely let out the smoke when the cap goes short and overheats instead of it blowing up with a large bang.
Did you say SMD parts? I can't tell from the above.
Maybe it's time to transition to a toaster over for SMD reflow instead of hand soldering for more reliable results.
That's how we do boards.
Toaster oven, multimeter with a thermcouple probe, and simple timer.
What seems to work well is oven full on, ramp to 150C, oven on/off to hold around 150C for 90 seconds,
then full on again to 190C, oven on/off for 90 seconds to 120 seconds, holding 190-195C,
then oven off and door slowly open to allow cool down.
Ours has a glass door, I look in with a flashlight, if solder looks melty & shiny after 90 seconds I stop there. Larger boards with bigger components (TO252 parts for example, D-PAKs, need closer to 2 minutues).
There is one other area that you need to address. How well did you clean the circuit boards after all the components were soldered on? Did you use "no clean" flux? Did you use lead-free solder? Lead-free flux will continue to "flux" until completely washed off the board and components.
