Voltage rating on caps?

[Logistics]

There is absolutely nothing higher than 12V running through a PC, ANYWHERE, correct?

I noticed on a soundcard of mine that some of the little 4mm guys in the sub-100uF values are 50V. Now, I have noticed that this can be advantageous if possible because it USUALLY lowers the ESR and sometimes raises the ripple. And often times, for a given line, you have to use a higher voltage rating to achieve a lower capacitance, anyway.

But technically, these caps can't possible NEED to be rated higher than 16V, right?

[tazwegion]

Re: Voltage rating on caps?

I wouldn't imagine so... but perhaps because it's on a soundcard the capacitor was chosen more for it's effect on sound quality/reproduction/tone than voltage rating

[starfury1]

Re: Voltage rating on caps?

humm yeah well you tend to find smaller values with larger voltages...like 1 uf 50V
Also purchasing power and what available might enter into to it too.
Actual design of caps may determine to some extent what they produce

The other side is they will probably produce as small a range as they can to reduce
manufacturing costs.

It possible both of the above but I am not going to get into the audio VS electrolytic cap debate

I suppose when it comes to building things too having same type size compoments may work to there advantage with loading machines

12 Volts would be tops for the MB well the GP stuff

Oh as far as voltage no....if its 12V 16V will do.....it really depends on what is the purpose of the cap to what specifications you have to look at or what aspect is being utilized with it.

[PeteS in CA]

Re: Voltage rating on caps?

It probably has to do with $$ - smaller lytics are probably cheaper that film types - and what is available at the lower values of C (lower as lytics go). MLCCs have issues with temperature and voltage coefficients (I keep playing that harp, don't I, , I may be ceramophobic).

[Spacedye69]

Re: Voltage rating on caps?

All the small lytics I see are at least 25v. I buy in bulk, and that is usually what I find is available. Most of the ones I have in low esr types are 50v or 63v.

[999999999]

Re: Voltage rating on caps?

A manufacturer buying in bulk can often get a part rated for higher voltage for the same price as the lower rated part, when they're tiny little few-uF value parts. I've said it before and I'll say it again, there is no need to try to match a cap voltage rating to just barely over what the circuit voltage is, but when that is done it is usually because going up any higher in voltage increases the cap cost or size, or reduces uF value (in typical switching circuits we usually see the capacitor failure in).

Rarely there were some old sound cards that ran -12V-0-+12, or 24V rail to rail and while they usually only had rail-ground capacitors, in theory one could have been designe to have a rail-rail 24V potential. These were sound cards with onboard chipamps meant to directly drive speakers, you are hard pressed to find anything like this today as new(er) product.

As for the lower ESR caps on a soundcard, generally if lower ESR benefits the subcircuit they're in, the card has SMD ceramic caps local (on the PCB) to the active stages benefitting from this lower ESR, making the 'lytic cap ESR fairly unimportant unless the system has atypically high ripple current and/or bad PSU in it.

Ignoring the old cards with rare situtions mentioned above, yes in general a 16V cap is sufficient. When it doubt about a cap, the safest way to know is to measure the voltage across it while the (part) soundcard is powered.

[stavr0]

Re: Voltage rating on caps?

I suggest reading http://www.repairfaq.org/sam/captest.htm#ctesr this explains what ESR is and why a circuit might require a capacitor with a specific ESR.

Interesting notes:

Low Temp rating = low ESR -- at first. But as time passes the low temp cap 's ESR will increase at a much faster rate than the higher rated temp one, so the higher temp will be more reliable in the long run.

And most important:

Higher voltage rating means higher ESR value. The ripple current will be the same for both caps and the higher ESR means more power dissipated into the cap, drying out the electrolyte and shortening the capacitor's life a lot.

[starfury1]

Re: Voltage rating on caps?

Higher voltage rating means higher ESR value

I think its the opposite

"higher voltage lower ESR"

This is from what I understand, why crap caps are more likely to be 10 Volts and not say 6V3
So they can get a lower ESR with a cheaper Cap


At the same amount of ripple...true

higher ESR means more power dissipated into the cap, drying out the electrolyte and shortening the capacitor's life a lot.

Well from the heat generated

[999999999]

Re: Voltage rating on caps?

I think its the opposite

"higher voltage lower ESR"

This is from what I understand, why crap caps are more likely to be 10 Volts and not say 6V3
So they can get a lower ESR with a cheaper Cap


At the same amount of ripple...true



Well from the heat generated

Yes, higher voltage results in lower ESR in the types of caps we use, keeping the same family of cap as a constant. ESR is related to the physical size and dimensions and this is why higher voltage = larger (all else same) = lower ESR.

We can see this easily enough on cap spec sheets.

[starfury1]

Re: Voltage rating on caps?

as you said 9's that was my understanding (larger can size)
Just thought I might have been missing something with the comments
(since the link was there which covers a lot of ground)

Cheers

[PCBONEZ]

Re: Voltage rating on caps?

I've looked (closely) at around 50 different P3 and P4 boards in the last 2 months exploring this and some other cap related things.

Those higher voltage small uf caps are found in places where things are likely to be connected or disconnected while the PC is running. - PS/2, USB, and Sound.
- Next to the sound jacks/chip.
- Next to the USB jacks/chip.
- Next to the PS/2 jacks/chip.

My guess is the higher voltage rating is a precaution to handle voltage spikes and sudden load changes when connecting/disconnecting something to a port.
(Particularly if the user makes a mistake.)
- Connecting/Disconnecting the wrong thing.
(Connecting the Speakers to the MIC jack. [MIC jacks output 5v to power the MIC.])
- Disconnecting things with current flowing. (USB powered drive that's running.)
- Connecting a device with a fault. (A busted iPod. A bad keyboard. A shorted MIC.)
- Shorting something by not being deliberate about plugging it in.
- Hot-Swapping the PS2. (Not designed for it but people do it. My wife for example....)

I suspect those caps are there to not to reduce ripple from the PSU going out, but, to keep RF from coming IN through the jack. Wires can act like antennas and they get routed past nasty EMI things like desk fans, monitors, paper shredders, printers, speakers.
-
This is why those are a smaller uF value than you see elsewhere on a board.
They have a different purpose.

~~~~

The voltage rating has nothing to do with the ESR rating.

You probably 'think' it does because the for a given uF you see a lower ESR at a higher voltage. What you are missing is that the can size changed too.
[By can size I mean both diameter and length.]

*Within a given series* of capacitor the ESR changes with THE CAN SIZE.

Go look at the data sheets for MBZ and MCZ. (Just picked an easy one.)
You will see what I mean.
-
Regardless of the voltage (and staying within the series).
All the 8x16 have the same ESR/Ripple.
All the 8x20 have the same ESR/Ripple.
All the 10x16 have the same ESR/Ripple.
All the 10x20 have the same ESR/Ripple.
~~ And on, and on....

You can look at any 'name brand' series and see the same pattern. It is consistent inside any series of the "good brand" caps. (Below about ~680uF things get squirrelly presumably because the 'guts' are mounted in a bigger can than they need to be for PCB mounting/application purposes.)

Those few 'good brands' that don't -seem- to follow this pattern actually do.
- Why? In some series' there are more than one model (design) of cap used in the series. Sanyo WG for instance is actually 4 different models/designs (WG, WGL, WGL2, and WGV) rolled into one series called WG. A few (not many) of the other 'name brand' series do something similar but they don't make it so obvious as Sanyo with sub-series designations in the data sheet and part numbers. [One example of how this can happen is would be two caps in a series with the same can size but available with two different lead spacings. Lead spacing affects ESR.]

None of this applies to some crap cap brands. Even with the same can size the ratings on some crap brand caps jump all over the place as if every single cap is a different design. They possibly are all different. Or perhaps this is further evidence of poor testing methods and/or poor quality controls. Or maybe even plain ole false advertising???

~~~

With the exception of those places where things get plugged in I haven't seen 16v (or greater) caps used on a 3.3v or 5v circuit on a motherboard at all.
- Not saying it's never happened. I just haven't seen it yet.

~~~

If using larger voltage caps to lower ESR were effective you'd see manufactures doing it on their high end boards where cost matters less and reliability more. But they don't. They use better grade caps with the same voltage (or even better type caps which cost even more).

With the exceptions of the VRM and other non-cpu voltage regulators on the board the ESR/Ripple rating requirements for the random caps scattered around a board (I call these "Field Caps") simply aren't that high. They don't do that much compared to VR caps. Even on the P4/800MHz FSB and Dual CPU Xeon/533MHz FSB server boards that I've looked at those caps in the 8mm sizes were rated 1050 or less for ripple and .069 or more for ESR. (For 10mm it's 1250 or worse and .044 or worse.) - The exceptions I've seen were obviously more for marketing purposes on 'premium' boards than because they were actually needed. The same manufacturer's 'entry level' and mid-range boards of the same class (chipset) follow what I said above so those were high grade caps for 'show'. (The server board with all MBZ's was REALLY PURDY! I like royal blue! It goes so well with gold!)

Some of the older P3/100MHz FSB boards I looked at had 8mm caps rated at only 670 for ripple. Ripple around 850-950 was the norm for P3/133MHz boards. A Dual Tualatin server board had 1050 but I found an early i845 P4(PC133-only)400FSB board [MSI] with only 840 for the field caps.

What I'm getting at is that you don't have to raise the voltage to upgrade those 'field' caps for better ESR. Just get better caps. There are PLENTY of options available that can beat those 'stock' ESR/Ripple ratings without raising the voltage at all.
-
If (outside of a voltage regulator) you have anything better than ZL, HD, GK, KZE, WX (which all exceed those 'stock' ratings I mentioned above) then you are already into the over-kill range and you don't need an upgrade.

.

[PCBONEZ]

Re: Voltage rating on caps?

Are there more 9's in your name than there used to be?

[999999999]

Re: Voltage rating on caps?

Are there more 9's in your name than there used to be?

LOL

There are nine nines, there always were and there always will be. You might just be seeing double, I get that way if I drink too much.

[Logistics]

Re: Voltage rating on caps?

What about this 450V cap I'm considering replacing in this LCD monitor? I can't find a 450V 150uF that is the appropriate size, and it can't be any bigger than it is now, really. There's just no more real-estate available. But there can't be anything higher than a DC voltage equal to the incoming AC, correct? So can I just use a 160V, 200V or even 250V as a replacement? ...unless there is voltage doubling going on in these things, but I'm not familiar with that sort of thing.

[arneson]

Re: Voltage rating on caps?

Not sure about the 450v cap rating but all the ones I've replaced have been 400v,
and no you cannot use less.

[999999999]

Re: Voltage rating on caps?

Logistics, what size is it? Would it be possible to drop down from 105C to 85C temp rating to find the right size?

Are you trying to use a very low ESR cap? They tend to be physically larger for same uF and voltage and for a typical bridge rectified (only)rather than boost APFC circuit the ESR isn't so important.


160V is definitely too low even for a 110V AC input with normal line fluctuations. 250V "might" be acceptible but it depends on the circuit. What you could do is temporarily solder a 450V cap on leads to the PCB and measure the resultant voltage very carefully then add a margin to account for AC line variations, or describe the circuit a bit more, what comes between the AC input socket and this cap. If you have a schematic all the better.