Filter caps over voltage on startup

[Wes]

What can be done about filter caps going over their voltage ratings when an amp first turns on?

Is it necessarily bad for a 16uf 475V cap to drift upwards to ~500V for 10 seconds or so at start up?

The only thing I can think to do is install a totem of two caps to handle this as it is done in many amps.

[Chuck H]

That's one way to handle it. And that's how I did it. But many amps just ignore the problem. I think I've read here that the Fender HR series amps all have this "problem". There was a short run of early models with filters of questionable merit and they did indeed fail a little disproportional in numbers. Later amps use better filters and don't seem to have the problem. Most caps are rated for a certain over voltage condition much greater than 25V and much longer than 10 sec. So it's PROBABLY nothing to worry about. If this is a build you're designing you could set up the board to use either a single cap or a totem arrangement. Then try the single cap and if it gives you any trouble you can change to the totem. Assuming your goal is to NOT use a totem if you don't have to. If this is just an amp you already have and you noticed this condition... Has it been a problem so far? If not, just ignore it. If this is a new build that is already layed out and built but you've noticed this problem, I say just use it awhile. If it's not a problem, it's not a problem. If it becomes a problem you can fix it when it presents.

[daz]

I don't understand why all the available ones are always 450 or 500v. Why don't they make them 600 instead and charge a bit more or offer those as a higher priced alternative? Seems stupid when they are so close to most amp's B+.

[soundguruman]

It's way better to use 2 in series, with bypass resistors. The over-voltage will shorten the life span.

600V caps are getting too hard to find and too expensive.

All the longest lasting amps have this series / bypassed setup.

It's not too hard to change HRD to the series deal. It's been done before.

[Chuck H]

It's way better to use 2 in series, with bypass resistors. The over-voltage will shorten the life span.

600V caps are getting too hard to find and too expensive.

All the longest lasting amps have this series / bypassed setup.

It's not too hard to change HRD to the series deal. It's been done before.

Agreed. It's not much work or expense to be sure about your filters reliability. And I use totems whenever there is the likelihood of even a momentary over voltage. But if it's not a problem and anyone wants to rely on the over voltage condition spec of the caps, for their own reasons which aren't yours or mine, I think that's fine too. I'm not even sure a small, short duration over voltage condition WILL shorten the caps life span significantly. Unless the over voltage is punching holes in the caps construction.

[Chuck H]

I don't understand why all the available ones are always 450 or 500v. Why don't they make them 600 instead and charge a bit more or offer those as a higher priced alternative? Seems stupid when they are so close to most amp's B+.

Components are made overseas in enormous batches. There's not enough demand for 600V caps to entice any of the actual makers of these components to offer a supply. The cost would be prohibitive. Modern caps are so low in impedance that series caps aren't a problem and most amp builders have adapted well enough. In fact, I would recommend staying away from 500V caps for the most part because they're too expensive and usually of lesser quality than the available 350V caps. A pair of Nichicon PW's at 350V will cost less and outperform a single F&T or Sprague 500V cap in every way.

[Enzo]

I don't understand why all the available ones are always 450 or 500v. Why don't they make them 600 instead and charge a bit more or offer those as a higher priced alternative? Seems stupid when they are so close to most amp's B+.

You ask the wrong question. Put it in context. Why do guitar amp makers insist on running B+ higher than standard cap voltage ratings?


Much as we might like to think we are important, guitar amps are a TINY corner of electronics. Designing a line of 600v electrolytics just for that would be an unwarranted expense. Common 350v caps will stack simply and reliably, so as far as industry is concerned it is not an issue. We might like our tubes, but the majority of guitar amps run on something closer to 40v than to 400v.

[daz]

Theres also the matter of capacitance. If you want 50uf like marshalls typically use, the typical value that would be 1/2 of that so that in parallel you have the proper value would be 22uf which gives you 44uf. I know it's not much different, but if you do that to 3 preamp nodes you have a total difference of 18uf less. May seem like it's not much but it can change the feel. If you design the amp around a typical value and you then decide to do this you won't likely be able to recreate the exact filtering by doubling up.
I wish i hadn't read this thread.

[Tom Phillips]

Electrolytic caps actually have a "working" voltage and a "surge" voltage rating. In the past caps often had both numbers printed on the case. See the attached photo for an example showing an old 40 µF Sprague TVA-1908 with a voltage rating of "500VDC / 550VDC surge." That surge rating was specifically provided to allow the designer to plan for “over voltage” at start up or other times during circuit operation. We would need to see the detailed data sheet for the currently available capacitor to know the surge rating of a particular cap. A capacitor from a reputable manufacturer would probably have a built in voltage surge rating. A no name or house re-branded part may take liberties and push the limits of the stated working voltage rating. This marketing approach is just the same as that used by suppliers of all kinds of items. With capacitors it’s all part of the tradeoff between useful operating life and operating conditions. For electrolytic capacitors used in a guitar amp the most significant operating conditions are the temperature and the applied voltage. Keep those parameters low and the caps will last longer.

As already pointed out, your existing capacitors may be fine for practical purposes. If your amp has some weaker area they may last the life of your amp. However, if you want to extend the life of the caps then you do have the option of converting to a totem pole design with an overall higher voltage rating. Another option is just to change them out after 5 to 10 years of use if you still own the amp.

Cheers,
Tom

[Rhodesplyr]

I don't understand why all the available ones are always 450 or 500v. Why don't they make them 600 instead and charge a bit more or offer those as a higher priced alternative? Seems stupid when they are so close to most amp's B+.

As I understand it, it's a matter of the basic chemistry of a single electrolytic capacitor. What I've read is that 450V is pretty much the voltage limit of the standard aluminum electrolytic capacitor recipe and that going beyond that takes some tweaking of the design and construction.

Supposed 600V electrolytic capacitors have always been few and far between, and I have no idea how that rating was achieved.

That said, the 475V rated Mallory FP cans used in Leslie amps typically hit 510V DC for 5-10 seconds or so until the 6550s heat up enough to drag the B+ down. Many have been doing this ever since the 1960s, and a surprising percentage of them still test good. The new CE Distribution cans, unfortunately, fail at turn-on far more often than the originals.

On the other hand, Dynaco MkIII amps can hit 580V DC at turn-on, and these are somewhat notorious for blowing up the 525V rated originals. In fact, one of these turned into a roman candle right in front of me.

Most electrolytics can survive brief surges above their working voltage rating. It's only with hindsight that we can tell how much is too much.

[Enzo]

Theres also the matter of capacitance. If you want 50uf like marshalls typically use, the typical value that would be 1/2 of that so that in parallel you have the proper value would be 22uf which gives you 44uf. I know it's not much different, but if you do that to 3 preamp nodes you have a total difference of 18uf less. May seem like it's not much but it can change the feel. If you design the amp around a typical value and you then decide to do this you won't likely be able to recreate the exact filtering by doubling up.
I wish i hadn't read this thread.

Yes, it is implicit in stacking caps that you need to calculate the total. A 50uf stack needs to be two 100uf series caps. These days cap tolerances are much closer than they were. Back in the days of Leo Fender, caps had tolerances like -20/+80% or even -40/+100%. Some makers used round numbers like 50uf, while others used industry sequence numbers like 47uf. For the same cap. The modern value of 22uf used to be labelled 25uf. All are approximations at best. Your 50uf cap could be perfectly fine and measure anywhere from 30uf to 100uf.

We can find examples like 50uf trying to result from 22uf caps, but who says an amp will cooperate and want exactly 50uf? what if your design calculations come up with 54uf or 38uf? Then no off the shelf part will match your needs. Given the tolerances, two caps marked the same can be farther apart than the stacked value is from your original.

[Diablo]

I've sourced 600 volt e-caps from Weber. I don't know if they'll stand up to higher voltage than a 500v rated cap, but they give me more peace of mind.

https://taweber.powweb.com/store/capord.htm

[Leo_Gnardo]

You can sidestep electrolytics entirely and use film caps rated typically 630V. See what Solen has on offer in the Antique catalog. (tubesandmore.com). Yes they're big, maybe a little 'spensive but I've used 'em when I don't feel I can trust electrolytics. You could use the film caps for the first & maybe second HV node then electrolytic for the remainder.

Carr has been using film filters in his amps for over a decade now. Whether you like 'em or not, the power supply is built to last.

Electrolytics currently supplied with 600 or 800V ratings are typically series-wired pairs without balancing resistors. I'd rather use a Solen 22/630 as the replacement for the primary filter in a B15 for instance than one of the modern "dynamite sticks."

[R.G.]

As I understand it, it's a matter of the basic chemistry of a single electrolytic capacitor. What I've read is that 450V is pretty much the voltage limit of the standard aluminum electrolytic capacitor recipe and that going beyond that takes some tweaking of the design and construction.

That is correct. Getting over 450-500V requires some creative electrochemistry, and I doubt that there are any "real" 600V caps, only ones adapted to live with that increased leakage and tolerate it.

Once again, science to the rescue. It's very easy to put a small timer into an amp that adds a startup load sufficient to keep it below "disaster overload" or "warp core implosion" for 10-15 seconds of a typical tube warmup. This fixes the startup peak issue.

What's REALLY bad is when the caps won't quite take it and you pull the output tubes out. Now the voltage stays over for the entire duration.

Probably the simplest thing is stacking caps. When stacking caps, BLEEDER RESISTORS ACROSS EACH CAP ARE MANDATORY, NOT OPTIONAL. This is because the DC voltage per cap is initially the inverse of the capacitances as they balance dynamically, but becomes the ratio of the leakage currents when they get stable. Leakage currents are not guaranteed in cap specs. You need to force DC balance across the series caps with bleeders.

Which is a good anti-death step anyway, as the bleeders run the caps down and may keep you from getting killed one day.

In the Workhorse amps, I put an LED in series with the bleeders so I could SEE which caps had voltage on them and which didn't. But that's just me.

[Tom Phillips]

...In the Workhorse amps, I put an LED in series with the bleeders so I could SEE which caps had voltage on them and which didn't. But that's just me.

Cool idea. Especially when one is constantly doing open chassis tweaking.