Capacitance must remain the same, while voltage can be higher (not lower as it will blow).
Example : 16V 470uF > Can be replaced with > 25V 470uF

It's also vital to maintain the same polarity with the replacement capacitor. Usually cylindrical ones have a big stripe on one side.

You can actually get away with a higher capacitance depending on what you're replacing, but in general, yes, stick to like spec components.

As a rule, same uF, same or higher voltage, similar ESR (resistance), same or higher temp rating (e.g. 105C) and same or higher endurance (e.g. 1000 to 5000 hours).

NEVER EVER replace with lower voltage (BOOM!). Higher voltage rating is safer (but more expensive).

Higher temp rating is safer (but more expensive).

Higher endurance is safer (but more expensive).

For some PC related applications higher uF and/or lower ESR can be of interest.

The japs make the best caps (Chemicon, Hitachi, Nichicon, Panasonic, Rubycon, Sanyo).

fyy wrote:

Lets say you have a 10v 1000uf capacitor. So the max voltage this capacitor can receive is 10v, and its capacitance is 1000uf. If this capacitor needed to be replaced, is there anything wrong with replacing it with a higher spec capacitor? What if I replaced it with a 25v 1500uf capacitor?

No problem with higher voltage, in fact it's better (safer). About the higher capacitance, it depends on the application but broadly speaking you should be OK.

There are a lot more parameters to capacitors than just rated maximum voltage rating and rated capacitance.

So the basic law is to replace with exactly the same kind, unless you know what you are doing.

Usually you want the
-same type of capacitor (aluminum, tantalum, ceramic, polyester, whatever, as different types are meant to be used in different places because of their differences)
-same capacitance rating (because sometimes there is a minimum limit and sometimes there is a maximum limit, for example to keep some control loop stable)
-same voltage rating (as with some capacitor types, the actual capacitance depends on the voltage the capacitor is connected to)
-same temperature rating (if the original one was rated for 105 degrees, a replacement rated only to 85 degrees has a much shorter lifetime)

There might even exist subtypes that are specific to the used circuit, for example there are low-ESR electrolytics for use in switching power supplies and in linear power supplies you can just use standard electrolytics.

Having said all this, the earlier posts do apply when replacing aluminum electrolytics in general case - use same or higher voltage rating, use same capacitance rating or slightly higher if you don't have the same, but not too much higher.

I mostly agree with the other posters.

You should try to keep the capacitance the same. In filter applications, you can generally go a little higher. 1000 -> 1500 is probably about the most I would bump it up. Keep the voltage the same or higher -- keep in mind higher voltage caps will be physically larger so make sure it will fit.

Awesome replies guys, thanks alot. Answered everything I needed

TELVM wrote:

NEVER EVER replace with lower voltage (BOOM!). Higher voltage rating is safer (but more expensive).

Actually, in computer PSUs and motherboards, you usually always CAN replace with lower voltages, since you're really only dealing with these voltages, specifically:

0.5V - 1.5V (CPU)
3.3V
5V
12V

So for the CPU, you can always use 3.3V caps instead of 6.3V that the manufacturers used (Polymers are a good example. Most people who mod the VRM from electrolytic to polymer go from 6.3V caps to 2.5V caps). For the 3.3V rail, if there's a 10V cap, you can use a 6.3V cap instead, and the same goes for the 5V rail. And for the 12V rail, you always need at least a 16V cap. And this doesn't apply to capacitors filtering non-switched circuits like a wall wart for example, because the voltage on those fluctuate depending on the load.

So in this case, if you have a 10V cap, you can safely go to 6.3V. The quality and spec of the cap is more important than the voltage of the replacement. Replacing a lower voltage cap with a higher voltage cap isn't an improvement, to the contrary, it can be diminutive if the replacement capacitor has a higher ESR and can handle less ripple than the original.

^ I know and I do, but unseasoned padawans to the dark smoking side of the electrical force by lower voltage caps swayed might be.

Don't forget jitter and varying voltage in cheaper PSUs.

TELVM wrote:

^ I know and I do, but unseasoned padawans to the dark smoking side of the electrical force by lower voltage caps swayed might be.

🤣, I'm not ashamed to say I was just that a couple of years ago. I used to buy higher voltage caps, especially the small value 5mm ones, thinking they were better. Recently I actually noticed that in many datasheets, a higher voltage 5mm cap of the same value often has inferior specs.

I seem to recall a review of solid cap recapping, and it was rather mixed, though the problems were no doubt partly due to making do with significantly lower values - if you can get the same value and required voltage in solid, it would probably be ok - in the main, I think solid's worked ok in the motherboards, but attempting to recap PSUs with solids of lower values was a complete failure.

Another key point, bigger capacitance and bigger voltage is likely to mean too big to fit the space, as well as bigger price than needed

fyy wrote:

Lets say you have a 10v 1000uf capacitor. So the max voltage this capacitor can receive is 10v, and its capacitance is 1000uf. If this capacitor needed to be replaced, is there anything wrong with replacing it with a higher spec capacitor? What if I replaced it with a 25v 1500uf capacitor? So now the capacitor has a higher max voltage and a higher capacitance. Is there any no-no's or "gotchas" that you all know about?

Basically, just give me the basic "capacitor replacement" laws.

If you're actually replacing capacitors in something, you'd be much better off just telling us what it is so we can properly direct you. I wasn't extremely thorough in reading everyone's replies, but I don't think anyone even touched on temperature range. It's possible that a given capacitor would be better in it's application if it was the 105-degree variety rather than the typical 85-degree. Often this is what the capcitor actually needs to survive better in it's application, rather than higher voltage or capacitance. Higher voltage and capacitance aren't going to help the cap live any longer if the problem is the capacitor is next to a voltage regulator and it's getting cooked because it's only an 85-degree cap when it should have been a 105 or even 125-degree cap.

How about tolerance? If we cannot obtain capacitor with the same tolerance, do we better use wider or tighter tolerance?

Well, I'm not an expert when it comes to this stuff, but it is my understanding that all capacitors are supposed to operate within 20% of their specified capacitance in order to be functional. So that means you can expect a 1000uF capacitor to give between 800 and 1200 uF of capacitance, if I understand this correctly.

Evert wrote:

Well, I'm not an expert when it comes to this stuff, but it is my understanding that all capacitors are supposed to operate within 20% of their specified capacitance in order to be functional. So that means you can expect a 1000uF capacitor to give between 800 and 1200 uF of capacitance, if I understand this correctly.

I remember reading somewhere in an audiophile forum that replacing the capacitors in your speaker's crossover with tighter tolerance yields better, more transparent sound, tighter and more articulate bass, better stereo imaging, etcetera, but I don't exactly remember the details. Nonetheless, this is an interesting topic, since I'm planning to re-cap my speaker's crossover and my amplifier.

Well, the Audiophiles (or Audiophools) are a strange crowd. To my understanding, it makes a significant difference when it comes to vintage audio equipment, but I think the so-called gains are marginal when it comes to new stuff. I personally would just replace capacitors to improve the stability, longevity or reliability of something.

Evert wrote:

Well, the Audiophiles (or Audiophools) are a strange crowd. To my understanding, it makes a significant difference when it comes to vintage audio equipment, but I think the so-called gains are marginal when it comes to new stuff. I personally would just replace capacitors to improve the stability, longevity or reliability of something.

Well, my speakers and amplifiers are vintage stuff (JBL 120Ti and Sansui AU-7900, respectively). Last year I was going to buy replacement capacitors for the speakers' crossover, but the JBL 120Ti maintenance manual doesn't say anything about voltage rating and tolerance -only capacitance. Then I was told that using tighter tolerance makes better sound -less frequency overlap between drivers, so the sound becomes cleaner. Not sure if it's true or not.

Well, I'll applaud you for trying it out and telling us about it. The problem is that audio is very subjective. What sounds great to one person, sounds completely shit to another.