Cap overvoltage risk?

[Khron]

Greetings everyone...

The question i'm posing is this: are 35V electrolytic caps too close to the 30-32V voltages i'm thinking of subjecting them to?

Operating conditions: I intend on using some as reservoir caps for some amps i'm building (not on the amp pcb's). Supply rails will be from an SMPS (half-bridge with feedback), so that should take care of pretty much any chance of over-voltage in case the mains go up.

[kevinkr]

Depends a lot on the manufacturer's derating curves, and your highest anticipated operating temperature. (Include self heating due to ripple if present.) In general for good electrolytic life (AL types) you should derate to about 80 % of nominal. IMO I would not use a 35V part at more than 28V.. Go for a 50V part - you'll never have cause to regret it. (Exploding caps are messy.)

[Khron]

The (main) reason i'm askin is that i have a number of 35V caps at hand (Panasonic mainly, so quite trustworthy), and i'd like to put them to (good) use. 50V caps i'd have to buy, and at the moment, that's not a too desirable option

[Mooly]

Providing the DC voltage ( and other parameters in the data sheet ) are not exceeded they will be fine. Maximum ratings are however exactly that !! a 35 volt part at 32 volts is running at around 8% below max. --- that's just about acceptable for long term reliability.
Kevins advice is very sound however, and you must take into account any ripple current and max temperatures. Heat and passing AC current is the big killer of electroylitics. It's the leading cause of switch mode PSU failure.
For ex a 100 mfd 35v cap across a 32 volt rail as decoupling for an OpAmp say is fine. A 2200mfd across an AC bridge with "around 32 volts" and perhaps with the rail supplying a couple of amps --- not so good --- and a 100 mfd 35v on a 32 volt rail in a SMPSU supplying an amp or so with a ripple component of perhaps 100khz or more TOTAL DISASTER. Get the idea

[Nordic]

Here is another option, apply some regulation... infront of the regulator the caps needs to provide a margin for line variations, but after the reg I think it is safe to use caps closer to the output voltage, wich should not be seeing spikes...

[S.G. Mahbub]

Dear khron,

For use of appropriate capacitors, you have to calculate lot of parameters, specially when you use it in smps. However, when you use capacitor filtering dc when you convert sine ac to dc, your requirement stands 30v ac(rms value) x1.41 =44/45v. So, there is a common practice to use 2x rms value and hence you should use 60v capacitor and it is for safer side. When you use filtering dc when converting square wave ac to dc, your requirement is 30 v capacitor as rms and peak value for square wave ac and dc value is almost same but you can have some margin for security purpose and should use 45 to 50v capacitor and for modified sine ac rms and peak value is slightly differs and hence peak value has to be catered for and hence using 50v capacitor is safe I think.

Moreover, when use use capacitors in smps in high frequency circuit, you should look into the ESR(equivalent series resistance)value of the capacitor as with the passage of time, electrolytes inside the electrolytic capacitor reduces and ESR value increases which is very dangerous for long lasting of the circuit.

Instead of using single value capacitor it is useful to use more higher valued capacitors as that will reduce the ESR value substantially in an smps circuit.

With thanks.

[Khron]

I'd like to apologise in case i haven't been too clear...

At the output of the aforementiones SMPS there will be 2 pairs of Panasonic FC 1000uF / 50V caps, with a choke between them, so ripple should be well reduced.

Adding to that, the SMPS will be housed separately from the amps, and these 35V caps i mentioned, i intend to use as reservoir caps, in the amp housing.

Thus, i guess / hope / think that these 35V caps will be reasonably safe...

[Mooly]

They probably will

[S.G. Mahbub]

Dear khron,
You are right. They probably will.
Thanks.

[S.G. Mahbub]

Dear Khron,

Probably you and Mooly are right. They probably will. But should you depend on probability in smps? If you are confident, you can use and test. But if you want to use capacitor in smps, you should follow the principal. In smps, there is no separate rule for main and reservoir. In circuit, everything matters and more so if you use high frequency. Your main 1000 uF capacitor of 50 volts also may not suffice if those are not of low ESR and good quality.

You probably know better than me as I could guess, but what I said, that is from experience. Capacitors are one of the main culprit for causing smps circuit failure. Though I am not very much conversant with amplifiers, but as my knowledge goes, the load on the reservoir capacitor is more than that on main capacitor in amplifier circuit. So, safety margin should be maintained.

With thanks.