Is 50v cap too high for an amp with 32v power supply?

[Mp3]

Hi everyone, upgrading the caps in my FX Audio D802, but one problem is the power supply is 32v and the main power caps are rated for 35v.... so i want to put in something with a higher tolerance, but all i can find is 40v or 50v, and there are only a couple choices in 40v, my question is 50v too high voltage for use with a 32v power supply? Or should i get some better quality 35v caps?

[Rerouter]

It is fine as long as
1. They fit
2. The ESR is not rubbish for the voltage.

[exe]

Actually, I suggest using 50V parts to put some safety margin:

1) the cap voltage rating is *absolute maximum* rating that is not to be exceeded. Working close to maximum voltage is pushing it to the limit
2) parts with higher voltage last longer (under same conditions)
3) parts with higher current last longer (under same conditions)
3) there are different temperature ratings. Cheap ones have 85C rating. Higher temperature parts (105C and 125C) are more durable, but cost more and have bigger size.

Just be sure new parts meet ESR and max ripple current requirements.

PS here is seems to be a nice article: http://www.learningaboutelectronics.com/Articles/What-does-the-voltage-rating-on-a-capacitor-mean

[radiolistener]

capacitor should have 30-40% voltage margin. For 32 V, you're need at least 32 * 1.4 = 45 V rated capacitor.
In your case it's better to use 50 V (or higher) rated capacitor.

Voltage rating on the capacitor is the maximum voltage limit for the capacitor. If you exceed it, it may explode or make short circuit. Also note that capacitor voltage rating is decreased with time, so if you plan to use your device for a long time, you're need enough voltage margin.

[Zero999]

capacitor should have 30-40% voltage margin.

That's a bit of a sweeping statement. In reality, it depends on the capacitor and how well regulated the supply voltage is. For example, on a well-regulated 5V supply, a 6.3V electrolytic capacitor would be fine. Using a much higher voltage rating than necessary has disadvantages: increased cost, size and ESR.

For 32 V, you're need at least 32 * 1.4 = 45 V rated capacitor.
In your case it's better to use 50 V (or higher) rated capacitor.

Voltage rating on the capacitor is the maximum voltage limit for the capacitor. If you exceed it, it may explode or make short circuit. Also note that capacitor voltage rating is decreased with time, so if you plan to use your device for a long time, you're need enough voltage margin.

Yes, this is true. In this case if the capacitor is being replaced with a higher voltage unit, it'll be fine.

Hi everyone, upgrading the caps in my FX Audio D802, but one problem is the power supply is 32v and the main power caps are rated for 35v.... so i want to put in something with a higher tolerance, but all i can find is 40v or 50v, and there are only a couple choices in 40v, my question is 50v too high voltage for use with a 32v power supply? Or should i get some better quality 35v caps?

Why do you think the capacitors need upgrading? Is it an old amplifier and you think there's a high change they'll need replacing? If not and the capacitors are good, there's no point, just leave it as it is!

[radiolistener]

on a well-regulated 5V supply, a 6.3V electrolytic capacitor would be fine

that's a way how Chinese guys doing it. In order to make electronics more cheap and less reliable. They don't care about reliability and durability. Because if it will fail, they will get more money. For 5V working voltage, 6.3V capacitor is 25% voltage margin - this is the minimum voltage margin for capacitors. And this is why Chinese electronics fails very often and don't works for a long time.

In opposite, when you're need reliability and durability, such as military electronics or something like that, there is need at least 30-40% voltage margin for capacitors.

[Zero999]

on a well-regulated 5V supply, a 6.3V electrolytic capacitor would be fine

that's a way how Chinese guys doing it. In order to make electronics more cheap and less reliable. They don't care about reliability and durability. Because if it will fail, they will get more money. For 5V working voltage, 6.3V capacitor is 25% voltage margin - this is the minimum voltage margin for capacitors. And this is why Chinese electronics fails very often and don't works for a long time.

In opposite, when you're need reliability and durability, such as military electronics or something like that, there is need at least 30-40% voltage margin for capacitors.

The cheap Chinese crap, has crappy capacitors, which don't meed their specifications and have a high ESR, so will fail quickly. Decent capacitor brands are far more reliable.

De-rating also depends on the capacitor. Tantalum capacitors definitely need de-rating, aluminium not so much so. Running at too lower voltage i.e. too little DC bias isn't good either, as some voltage is good to keep the capacitor's oxide layer in good condition. I wouldn't recommend a 100V capacitor, in an application rated to only 10V.

It's simply not true that using a 6.3V aluminium capacitor on a 5V supply, compromises reliability. Capacitor size roughly quadruples, every time the voltage is doubled (E = ¹/₂CV²), along with progressively having a poorer ESR, if the package size isn't increased enough. A 10% to 25% safety margin is fine, giving a roughly 20% go 56% oversize, but going far beyond that, it starts to get silly (40& higher voltage rating is nearly double the size!), especially if the power supply voltage is well-regulated and it's not being operated at high temperatures.  It makes perfect sense to use a smaller, lower ESR capacitor, than a much larger one, which will quite likely have a higher ESR.

As far as we're aware, this is a home audio amplifier, not a military grade piece of equipment! Even in designs for aerospace and defence, there's no blanket rule on de-rating capacitors. It depends on a number of factors such as: how long it will be in service for, before it will need to be replaced, the temperature, size, weight and space restraints. Using much larger capacitors than necessary, might be a problem in an aerospace design, which needs to be as light as possible. I used to work for a defence contractor and we had a reliability engineer who would look into the mean time before failure of different components in a system. Electrolytic capacitors are nowhere near as unreliable, as many believe them to be.

In this case a 35V rated capacitor on a 32V rail, might be OK, if that's the nominal voltage and it doesn't get higher than that, or it could be marginal, if the power supply is poorly regulated. It also depends on whether the capacitor is subject to high ripple current or is hot i.e. next to a heat-sink. Of course, if the capacitor needs replacing and you can only get higher voltage ones, in the same value and footprint, then go for it, otherwise don't bother.

[radiolistener]

In this case a 35V rated capacitor on a 32V rail, might be OK

I don't agree, 35V is just 9% margin for 32V working voltage.
It can be easily exceeded by usual power supply voltage tolerance which is about ±5..10%.

[Zero999]

In this case a 35V rated capacitor on a 32V rail, might be OK

I don't agree, 35V is just 9% margin for 32V working voltage.
It can be easily exceeded by usual power supply voltage tolerance which is about ±5..10%.

Note, I said might, because we don't know what the tolerance is. The original poster just said 32V. It could be the nameplate, or measured voltage. The power supply might be regulated or not. The tolerance could be 1% or 50%. It's unknown. It might be fine, or it could be on the verge of blowing up.

In any case, I agree, a 50V capacitor is not a bad idea.

[David Hess]

A higher voltage capacitor will work fine and will even have advantages.  The larger size yields higher ripple current rating and longer operating life.  Since more modern capacitors tend to be smaller for the same capacitance and voltage, a higher voltage part may be necessary to meet other specifications.

It is fine as long as
1. They fit
2. The ESR is not rubbish for the voltage.

ESR usually drops with higher voltages up to about 150 volts where a different electrolyte has to be used.

Below is an example where I probably should have used even higher voltage replacements just to get a physically longer part for better mounting.

[Zero999]

A higher voltage capacitor will work fine and will even have advantages.  The larger size yields higher ripple current rating and longer operating life.  Since more modern capacitors tend to be smaller for the same capacitance and voltage, a higher voltage part may be necessary to meet other specifications.

ESR usually drops with higher voltages up to about 150 volts where a different electrolyte has to be used.

You're right, the ESR does reduce if the case size is also increased. My point was that if the case size stays constant the ESR tends to increase, with voltage rating. If you can fit a bigger capacitor, in the same place, by all means up the voltage rating, but if it means having to bodge it, then don't bother, as it can cause other failure points.

The datasheet linked below suggests above 35V, the ESR increases, if the case size remains the same. This is of course assuming the capacitance normally falls too. If the voltage is upped and the case size not increased, I suspect that squeezing a higher capacitor value and voltage rating, into the same case increases the ESR. Yes, if this is an old design, you'll probably be able to get modern, higher voltage, lower ESR, parts with the same capacitance into the same or smaller case style, but I doubt the same applies to modern parts.
https://docs-emea.rs-online.com/webdocs/1500/0900766b815005ec.pdf

Below is an example where I probably should have used even higher voltage replacements just to get a physically longer part for better mounting.

Yuck! In that case lager capacitors should have been used purely to avoid those bodge wires, which should be insulated, to reduce the risk of short circuits.

[Rerouter]

should make clear on the voltage margins, you don't want to exceed them, and you can sometimes extend there life by running them at a lower voltage, so if your power supply is very well regulated and low ripple current it generally doesnt matter,

On a device like an amplifier where I presume its more or less straight out of a rectifier, then you need to leave some margin, however this margin should be based on the AC voltage that your multimeter measures before the rectifier, and not the DC voltage measured after it,

Measure AC voltage before rectifiers, times that by 1.414 (Changing RMS voltage into peak to peak voltage)
Measure your mains AC voltage, see where your in your countries power specifications, and add margin to account for it being at that maximum, (its relative, so if you measure 240V and your mains allows up to 250V, then times it by 250/240)

And finally you add your compfort margin, generally atleast 5% to account for surges on the mains,

If you meet this your set, if you exceed it, then so long as you still meet my first 2 points.

[innkeeper]

Hi everyone, upgrading the caps in my FX Audio D802, but one problem is the power supply is 32v and the main power caps are rated for 35v.... so i want to put in something with a higher tolerance, but all i can find is 40v or 50v, and there are only a couple choices in 40v, my question is 50v too high voltage for use with a 32v power supply? Or should i get some better quality 35v caps?

What are you trying to accomplish?
What is the problem your trying to solve, or, what are you trying to improve?

[vk6zgo]

should make clear on the voltage margins, you don't want to exceed them, and you can sometimes extend there life by running them at a lower voltage, so if your power supply is very well regulated and low ripple current it generally doesnt matter,

On a device like an amplifier where I presume its more or less straight out of a rectifier, then you need to leave some margin, however this margin should be based on the AC voltage that your multimeter measures before the rectifier, and not the DC voltage measured after it,

Measure AC voltage before rectifiers, times that by 1.414 (Changing RMS voltage into peak to peak voltage)

You stuttered a bit, -------I know you really meant peak voltage.

Measure your mains AC voltage, see where your in your countries power specifications, and add margin to account for it being at that maximum, (its relative, so if you measure 240V and your mains allows up to 250V, then times it by 250/240)

And finally you add your compfort margin, generally atleast 5% to account for surges on the mains,

If you meet this your set, if you exceed it, then so long as you still meet my first 2 points.

[vk6zgo]

This 6.3v thing happens even in the "best of families".

Sony loved using them, & must have known something, as they did last, (ten years or so), but their capacitance as read on a digital  display LCR meter was usually well down & "losses" (it didn't talk "ESR") were up, so I would replace  them "on principle" with 10v rated caps, which were usually in the same package.
I didn't have the same sources as Sony, 6.3v caps were very rare at my usual suppliers, & the alternative sources didn't inspire confidence.

You can still get caught out with some other manufacturers.
I replaced the faulty "biggish" NP smoothing cap for filtering the "boost HT" on an Electrohome monitor with its "modern equivalent" marvelling at how they had reduced the size, till it started to overheat & blow "steam" out of the end.

Low ESR equivalent NP caps in that value & rating were unobtainium, so it ended up with a "Christmas tree" of polyester "greencaps" instead.
Ugly? Messy? Not best practice?-----Yep! But it worked, & that is the primary requirement.

[radiolistener]

This 6.3v thing happens even in the "best of families".

They doing it intentionally in order to implement planned obsolescence policy for home electronics. 
Manufacturers are not interested in reliable and long life home electronics, because in such case they will lose profit.

[langwadt]

This 6.3v thing happens even in the "best of families".

They doing it intentionally in order to implement planned obsolescence policy for home electronics. 
Manufacturers are not interested in reliable and long life home electronics, because in such case they will lose profit.

if you build something to last 10 years that is going to be replaced in 3 years anyway, it'll be too expensive and no one will buy it

[schmitt trigger]

Below is an example where I probably should have used even higher voltage replacements just to get a physically longer part for better mounting.

Love that 70s-era electronics assembly. Just like Disco music, it brings back old memories.

[Doctorandus_P]

Is there any indication the 35V caps are failing?
(I am assuming we're discussing some Electrolytic Power Caps here, it has not been mentioned explicitly)
32 V is just fine for a 35V (or higher, the sky is the limit) rated cap.

[vk6zgo]

This 6.3v thing happens even in the "best of families".

They doing it intentionally in order to implement planned obsolescence policy for home electronics. 
Manufacturers are not interested in reliable and long life home electronics, because in such case they will lose profit.

if you build something to last 10 years that is going to be replaced in 3 years anyway, it'll be too expensive and no one will buy it

These were Broadcast level equipment---10 years is incredibly youthful in that environment!
Sony designed stuff that would handle component value changes.
Too expensive to do that for consumer stuff!

[David Hess]

Below is an example where I probably should have used even higher voltage replacements just to get a physically longer part for better mounting.

Yuck! In that case lager capacitors should have been used purely to avoid those bodge wires, which should be insulated, to reduce the risk of short circuits.

That is actually how it was originally designed!  That large capacitor was what was originally installed and it used the same wiring arrangement.

If you look closely, those wires are *not* in parallel.  Instead, that is a Kelvin connection to each capacitor lead to minimize capacitor ESR.

I should have used even higher voltage replacements to get a long enough case length to reduce the wiring length but to make up for that, I used thicker wires than the original which is shown below.

[oskimac]

make sure to keep an eye on the polarity when replacing those caps... i have seen professional you tubers failing at that...    

[exe]

If you look closely, those wires are *not* in parallel.  Instead, that is a Kelvin connection to each capacitor lead to minimize capacitor ESR.

I understand why we need Kelvin connection on a power shunt, but I've ever seen Kelvin connection on capacitors. Could you please tell the idea behind?

[David Hess]

make sure to keep an eye on the polarity when replacing those caps... i have seen professional you tubers failing at that...    

I heard about some goofy Australian guy doing that once. 

If you look closely, those wires are *not* in parallel.  Instead, that is a Kelvin connection to each capacitor lead to minimize capacitor ESR.

I understand why we need Kelvin connection on a power shunt, but I've ever seen Kelvin connection on capacitors. Could you please tell the idea behind?

If the inputs are floating, which they are in this case, then running the inputs directly to the capacitor leads and then making the output connections to those points results in a lower ESR than if the capacitor leads were just extended or run in parallel.  The same idea applies to a single point ground.

Where this matters is that the high charging current on line voltage peaks flows through a minimum of resistance so adds a minimum of ripple to the output.  If for instance you added a 1 ohm resistor in series with one of the capacitor leads, then for every amp of input ripple current, the output ripple voltage would increase by 1 volt.

It can be though of as placing the capacitor as directly between the rectifier and load as possible.  It is actually a pretty common configuration but usually it is not so apparent because the capacitor is directly mounted to the printed circuit board.