I am thinking about scavenging some capacitors for use in a circuit on household (120VAC) power. How can I determine whether a capacitor is rated to handle that voltage (without damaging it)?

 

The maximum voltage is given by the manufacturer.

 

Right, but if I'm scavenging from old equipment I can't readily determine the capacitor's manufacturer. Is there a way to determine this if you only have access to the capacitor itself?

 

If not stamped in their bodies in either plain figures or coded (which would be very unusual), try to locate a similar size, material and capacity capacitor that shows the voltage stamped on it to get a reasonably near guess.
If you are unsure; do not bother reusing those that cannot be identified, there will be plenty of properly marked others you can canibalize from other equipment.

 

If you can't read the voltage rating on the cap nor know the manufacturer and part number, unfortunately the correct answer as far as I know is there is no way to know for sure.

That said, it might be possible to use a DC power supply and an ammeter to slowly increase the voltage across a capacitor and monitor the current. A perfect capacitor should show an increase in current when the voltage is increased slightly until the capacitor is charged to the new value. Then monitor the current needed to hold it at that voltage. If you studied a variety of capacitors, you might be able to see and recognize behavior that told you the capacitor was approaching breakdown. (I'm saying you might see this behavior, not that it exists -- I haven't done the experiments.) Unfortunately, when you reach a point of breakdown, you may irreversibly damage the dielectric; thus, what you're looking for are symptoms of reversible breakdown. Hopefully, there's an expert on the board who has measured capacitors and knows whether this type of behavior occurs.

Maybe there are ways of stimulating the capacitor with more "exotic" signals (e.g., narrow high voltage pulses) that can reversibly measure breakdown characteristics.

 

for safety's sake, id recommend either just buying new capacitors or looking for devices where the breakdown voltage is marked. Because you are making a circuit dealing with high voltage, it would be dangerous making it with old and potentially underrated caps.

here is a peice of *duh* advice:
one way you can know if a cap you are looking at will handle 120v is if its already hooked up to a 120v circuit. Generally manufacturers will use caps only a margin of safety above the voltage in their respective part of the circuit, so don't even bother looking in low voltage circuits

 

the process to identify the breakdown voltage is fairly simple: pick up a high voltage high value resistor (>1M), and put it in serial with the capacitor and apply high voltage to it.

if the capacitor is good, you will measure a high voltage across it, near the voltage you had applied to the rc network.

if the capacitor is no good, you will measure a low voltage across it. the resistor is there to limit the current going through the capacitor if the capacitor does break down.

wear a pair of safety glasses when you do this because capacitors may explode when their voltage rating is exceeded.

 

I see nothing wrong with reusing components, including capacitors,

provided you know what you are doing.

You should know, or someone here should have reminded you that there are two sorts of capacitors - polarised and non polarised - and these types are not in general interchangeable.

If you are salvaging caps out of for instance old washing machines, or other machinery and intend similar use this can be OK. These caps will be non polarised.

I would test these by running up a few samples on a variac. You could substitute increasing transformer tappings with some ingenuity.

If you are salvaging caps from TVs and other electronic equipment, these caps will probably be polarised.

Polarised caps need approximately the correct DC voltage (of the correct polarity) across them or they will have a very short life.

These can be tested for shorts at low DC voltage and again tested at successively higher levels, by the resistance method mentioned by Millwood, all the while looking for excessive leakage.