Voltage rating on filter caps

[Gtr0]

Hi Folks.

Quick question to fill the nooks and crannies of my knowledge on power filtering and voltage ratings.

Two caps rated at 500 Volts in series can accept 1000 Volts, correct? Then Two caps in parallel will still only be able to accept a 500 Volt max. Also correct?

I find that in the info widely available on the internet, voltage isn't as popular a topic when talking parallel as they are when talking series. So honestly I am a bit uncertain.

And what are people - or how - are they using these types of caps... let's say your average F&T 500 volt cap. What max voltage should be pumped into one? There is so much depending on what the wall voltage is in any particular moment to say that a power tranny that dumps 480 Volts DC from the rectifier would be fine... because one tranny that puts out 480 Volts today can be 490 tomorrow, or even more.

Is there a "rule" in such case?

Thank you!

[Malcolm Irving]

Basically correct, but two 'equal' value caps in series are not very good at sharing the voltage equally between them, so it is common to put two balancing resistors alongside. The junction point of the two resistors is connected to the junction point of the two caps. Typical values are around 100k to 200k. The resistors dissipate power continuously, so a good safety factor on their power rating is desirable. Also watch out for the voltage rating of the resistors.

The voltage limit on a cap is 'instantaneous'. If a critical voltage across the cap is exceed, even for an instant, the insulation layer will 'punch through' and the cap will be destroyed (may even blow up). For a particular cap, it is hard to tell what that critical voltage is, so manufacturers try to rate their caps at a safe level, where that should never happen.

[mozz]

You should be able to get the datasheet from the manufacturer, that would tell you the surge voltage and a lot of other specs.

[Chuck H]

Further...

The series caps are (were) typically used where a suitable cap of a particular value/rating is either unavailable or cumbersome to a design for cost or space reasons. Another thing to consider is inrush current. When voltage is first applied to the caps there is a moment where the unloaded voltage is on the caps as they charge. In this circumstance the caps may see a hundred more volts than when the amp has stabilized. This would be a good reason to use series caps with balancing resistors. When choosing filter caps it's a good idea consider the unloaded voltage of a design.

[Gtr0]

Thank for the info. Oddly I did check out the datasheet and couldn't pinpoint the relevant info for inrush... however I do have a 100k 2W resistor across the standby switch which is supposed to help with inrush current, letting some bleed by when amp is first turned on and in standby.

As far as having them in series, yes I do have balancing resistors on mine, and yet the input voltage is in fact exceeding the single un-series'ed cap rating of 500, even though it is only by 5 volts. At any rate I have ordered a new tranny that should bring that voltage down to around 480.

So the question is would you use a single 500V cap for 480V supply???

Thanks!

[Chuck H]

So the question is would you use a single 500V cap for 480V supply???

Only if that's 480V Unloaded or the 500V caps have a peak voltage spec above the unloaded voltage. If that's 480V loaded I would use a pair of 350V caps in series (with balancing resistors). Even with the bleed resistor on the standby switch.

[J M Fahey]

Two caps rated at 500 Volts in series can accept 1000 Volts, correct? Then Two caps in parallel will still only be able to accept a 500 Volt max. Also correct?

Yes.

I find that in the info widely available on the internet, voltage isn't as popular a topic when talking parallel as they are when talking series.

Not sure to follow you.
Capacitors have a voltage rating which must be respected, not sure why would anybody argue that which is the "normal" way to use them.

Now when not enough voltage rating is available, then it becomes a hot topic and series connection starts being talked about.
As in: unusual facts are News .... everyday standard things not that much

And what are people - or how - are they using these types of caps... let's say your average F&T 500 volt cap. What max voltage should be pumped into one? There is so much depending on what the wall voltage is in any particular moment to say that a power tranny that dumps 480 Volts DC from the rectifier would be fine... because one tranny that puts out 480 Volts today can be 490 tomorrow, or even more.
Is there a "rule" in such case?

Unwritten but wise rule is to have some safety factor, 10% *at least* and if possible 20% comes to mind.
So I would not trust a 500V cap with more than 450V ; if anything because 10% is a *mains* voltage variation you can find any day of the week.
So do NOT use a 500V cap with 480/490V by any means, you are playing with fire ... literally.

Want to save money and play it safe?
Manufacturers who do this for a living use 2 series capacitors each one rated 350V or 400V DC , easily available anywhere, any amount you want, tons of brands to choose from, you name it.
And get a HUGE safety rating, go figure.

Thank for the info. Oddly I did check out the datasheet and couldn't pinpoint the relevant info for inrush... however I do have a 100k 2W resistor across the standby switch which is supposed to help with inrush current, letting some bleed by when amp is first turned on and in standby.

Answers above in general addressed surge *voltage* , when caps get chargedn to full unloaded pieak voltage which is easily 20% above working voltage, until heaters warm up and tubes start pulling it down.
Surge current was not exactly addressed.

As far as having them in series, yes I do have balancing resistors on mine, and yet the input voltage is in fact exceeding the single un-series'ed cap rating of 500, even though it is only by 5 volts.

Who cares about "single un-series'ed cap rating" if they are NOT being used that way?

At any rate I have ordered a new tranny that should bring that voltage down to around 480.

If you use them in series, no need to spend on a new transformer just for a little voltage lowering; if you plan to use a single "unserie´d" cap, your new safety margin is not enough.

So the question is would you use a single 500V cap for 480V supply???

No way!!!!

[Gtr0]

Yes.

Not sure to follow you.
Capacitors have a voltage rating which must be respected, not sure why would anybody argue that which is the "normal" way to use them.

Now when not enough voltage rating is available, then it becomes a hot topic and series connection starts being talked about.
As in: unusual facts are News .... everyday standard things not that much

Unwritten but wise rule is to have some safety factor, 10% *at least* and if possible 20% comes to mind.
So I would not trust a 500V cap with more than 450V ; if anything because 10% is a *mains* voltage variation you can find any day of the week.
So do NOT use a 500V cap with 480/490V by any means, you are playing with fire ... literally.

Want to save money and play it safe?
Manufacturers who do this for a living use 2 series capacitors each one rated 350V or 400V DC , easily available anywhere, any amount you want, tons of brands to choose from, you name it.
And get a HUGE safety rating, go figure.

Answers above in general addressed surge *voltage* , when caps get chargedn to full unloaded pieak voltage which is easily 20% above working voltage, until heaters warm up and tubes start pulling it down.
Surge current was not exactly addressed.


Who cares about "single un-series'ed cap rating" if they are NOT being used that way?

If you use them in series, no need to spend on a new transformer just for a little voltage lowering; if you plan to use a single "unserie´d" cap, your new safety margin is not enough.

No way!!!!

Thank you.

I have not come even close to doing so, been using them in series and parallel where needed making sure I typically have a few hundred volts of breathing room in caps. This really is just to more formally put myself better in the know so to speak. When in doubt I always go the safer route and I just wanted to know what the norm typically is.

Of course occasionally space is a consideration and while I could easily use cans, I just wanted to make sure I was properly informed so I can make the proper... and safe... decision.

Thank you! :-)

[mozz]

Can someone decode this F&T datasheet? I think this surge means 110%
Surge voltage US UR ≤ 315 V: US = 1,15 UR ; UR > 315 V: US = 1,10 UR

https://www.ftcap.de/en/downloads/

[Chuck H]

Well, in the technical notes it say's this:

[eschertron]

I just wanted to know what the norm typically is.

As was hinted at, the norm is what the engineers and bean counters agree on

If you can get your hands on 575V caps, then you might be safe with a 480V supply (about 20% margin, what Juan considers minimum-to-acceptable). But if the 575V cap costs way more than two 350V caps, then make room in the design and pat yourself on the back for being safer and more frugal. Besides, with a series pair of caps in the design portfolio, you can easily use power supplies up to 580-ish volts unloaded on the same 350 + 350 totem pole.

[Gtr0]

I suppose caps of many voltage ratings exist, but my amp supply places typically carry a max of 500v.. So I tend to just use 450 or 500 volt caps in series, then paralleled if I need to. So far it has proven to be of small issue when it came to size and chassis installation.

I find by bigger issue tends to be adding them up to target a specific µF. So for example an amp I have built has 2x 5U4G rectifier tubes, according to the datasheet I am able to use as high as 40µF capacitence for the reservoir. Tranny puts out 505 VDC when using UF4007s and 486 when using 5U4Gs. So if I wanted to target 80µF total capacitance I either find 4 X 80 µF caps (JJ or Atoms, but Atoms are literally 4 to 5 times the price at my supplier) at 350 - 500 Volts, OR ...

100 + 100 in series = 50µF
50 + 50 in series = 25µF

50 and 25 in parallel = 75µF

of course all 350 - 500 volt caps.

Is it acceptable to use different values in power filtering stages?

Right now I am using 4 x 47µF 450volt caps, totalling 47µF 900v - sounds a bit low for a high gain amp, but the amp does sound AWESOME. So I am probably too worried about nothing - as usual.

[J M Fahey]

Well, we can´t argue *awesome*, can we?
I bet 50uF will be fine with 5U4 , since in the old days a nominal 40uF cap could have had *up to* 60uF (+50/-30% tolerance).

Being in Argentina, meaning very high Courier cost, Customs Tariff, Tax, lost time at Customs "liberating" shipments, etc. , I rather use "what is readily available over the counter" and for Tube amps I am blessed by the abbundance and low cost of 100 to 220uF caps rated 350/400V, which are used in TV supplies ... as well as in many other SMPS, so I can find them *anywhere*, as many as I want and for peanuts.

Don´t even *try*to get 450/500V caps.

[Gtr0]

Well, we can´t argue *awesome*, can we?
I bet 50uF will be fine with 5U4 , since in the old days a nominal 40uF cap could have had *up to* 60uF (+50/-30% tolerance).

Being in Argentina, meaning very high Courier cost, Customs Tariff, Tax, lost time at Customs "liberating" shipments, etc. , I rather use "what is readily available over the counter" and for Tube amps I am blessed by the abbundance and low cost of 100 to 220uF caps rated 350/400V, which are used in TV supplies ... as well as in many other SMPS, so I can find them *anywhere*, as many as I want and for peanuts.

Don´t even *try*to get 450/500V caps.

Initially in this particular amp, I was using F&T 2x 220µF in series, 110µF total (with balancers of course)... the pair of 5U4Gs were JJ 5U4GB. And everything was great for a month or so, then one day I happened to see a small flash in one of the 5U4GBs... so I immediately swapped out the 220µFs and all I had lying around were 47µF 450v, so I used 4 of those in series/parallel in place. Been fine ever since. Popped in some EH 5U4G also, for no reason other than one of the JJ's had experienced what I thought was internal arcing.

So I assume pushing the datasheet boundaries depends on brand, though I was over 50µF to more like 55 (-10% / +30% of course )

[Mick Bailey]

Sometimes in older equipment you find caps marked 'WV' after the voltage figure. This is Working Voltage. The cap manufacturers rated these caps to withstand the voltage with an additional margin built in. So you'd find a maybe a 350WV cap operating on or close to its rating with no problems. Nowadays I look at a cap's voltage and add my own margin. This isn't shared by many manufacturers who will use a cap with no margin at all - sometimes even using a cap in circuit where it's over-voltage. I've had a few TV SMPS repairs where the secondary caps are running slightly over their rated voltage by design, and after a few years these have failed. All that was needed to effect a repair and give reliability was to replace the caps with adequately rated parts.