Hi there, I was just wodering if one could use large electrolytic capacitors with a low voltage rating (ie 50mfd , 50 000 micro farad 15volt) in some configuration on higher voltages, maybe 50 volt dc. The question relates to a large stock of computer grade capacitors, but all with a low working voltage.
Thanks,
Roelof

 

The 50V will most probably exceed the voltage limit of the dielectric and an arc will result on the 2 plates inside the capacitors => capacitor "explodes".

 

I've never actually tried any overvoltage tests to see how far capacitors could be pushed to failure, but all of my experience says don't expect reliability if you do this.

Some Capacitors explode when they fail, I guess something to do with them storing quite a lot of energy, so if you do want to try driving capacitors over their rated voltage, I suggest you shield yourself from your project!

What you could do is connect several of them in series, maybe with voltage equalizing resistors both to force equal voltage division and to safely discharge them when power is shut off. That's a completely acceptable way to get a higher voltage rating from low voltage capacitors.

 

Thanks for the reply, that is what I am looking for, to connect them in series/parallel with resistors to handle the high voltage . How do I do that ? Maybe some example or easy formula will help.
Thank you.
Roelof

 

If you connect two perfect equal-valued capacitors in series, the result has half the capacitance and twice the voltage rating as each individual capacitor. When the capacitors are connected to a DC source, they each will charge up to half the total voltage.

The problem is that capacitors aren't perfect. If one has more leakage than the other then it will have less voltage across it, forcing the other to carry more voltage, and if the voltage imbalance is great enough, the capacitor carrying the too-high volage can explode.

A way to force them to be better balanced is to connect equal value, high-value resistors in parallel with the capacitors. This way, the resistors divide the voltage much more equally than the capacitors would and help keep the voltage division balanced across the capacitors. Plus, when the power is shut off, they will bleed off the charge from the capacitors over a few seconds or minutes of time, making the circuit much safer to work on.

What you need is to use resistors that are much less than the capacitor internal leakage but are high enough in value that they don't waste much power from your circuit.

A megohm might be a good value to start with.

Plenty of people get away with combining capacitors in series without using resistors like this; this is just a suggestion to help make the circuit more reliable.

 

Thank you for your reply. I assume that the normal will apply, ie: 3 equal caps in series, a third of capacitance and 3 times the voltage etc.
Regards
Roelof

 

That's correct, but I would go for much lower value equalising resistors - 10K or even better 1K to ensure the voltage is properly shared across the series capacitors.

With 1M you would get 1V difference in voltage for each 1uA difference in leakage, which could easily overvoltage one or more caps and cause problems or confetti..

 

Thank you all. I will try a few ideas and let you know. The capacitors are all rather large, so I assume the possibility for small differences is larger than normal. So I am going for the 1K resistors. (These are all computer grade 50K to 100K micro farad)
Roelof

 

If the resistors are dropping 15VDC, and are 1KΩ, then they will be disappating .225W, which means if they are ¼W resistors they are too close to rating. Use ½W resistors if you do this.

 

Thanks, will be using at least 2watt resistors. (have some in stock)