Voltage rating on filter caps

[jmaf]

Here in my region they began fining energy consumers who had a lot of reactive power on their bills. It's government-run eletrical utility and as such they'll find a way to fine you for everything. They installed electronic measuring systems and the electronics can detect reactive power where the old magnetic wheel meters did not detect anything. So this upgrade generated a immense wave of fines - a lots of work for electricians, some of whom are good friends of mine and made a nice buck installing capacitor banks for bars, restaurants and such which were the first targets of massive reactive consumption fines.

Well this gave me the opportunity to learn about power factor correction capacitors, which are heavy duty non polarized caps.

What I learned is that capacitors are actually rated for reactive power. These PF correction caps have a KVAR, or KVA-reactive, rating which is how much inrush reactive current x voltage reactive it can handle. In the world of amplifiers we're used to thinking that caps can simply handle whatever we throw at them as long as voltage is fine, but that's not the case at all, the reactive power is what's important. A tiny 50 uF capacitor will not be able to handle the same inrush current as a big bulky 50uF capacitor rated for the same voltage.

The caps they use for PF correction have escape valves which let out excessive pressure built into the cap during high load operation, because caps aren't perfect they do heat and let out vapor. We don't have this on amplifier caps, thus we see the bulge on top of caps that've been through abuse.

All in all messing around with higher power electrical caps taught me a lot about amplifier caps. The 3 phase voltage here is 380 V star configuration, with each phase having 220 VAC. PFC caps are rated 500V, giving them over 30% margin. So for whatever standard they're following, +30% seems to be their safety margin for 24x7 utility capacitors. You should apply at least the same to amplifier caps IMO, if your peak voltage, without power tubes on, is 500V then your totem should handle at least 650 to 700 VDC.

Reactive power is fascinating, it heats the wires, creates a mess if you don't do it right and will knock circuit breakers right out, all while producing nothing. Reactive power produces nothing, it's just there to fill the capacitors and inductors so they can function. Fluorescent lamps, refrigerators, SMPS and every single wall wart from computers to phones you plug in these days produces reactive power. Eventually everyone will need to learn more about these tin can beasts called power factor correction capacitors.

[Chuck H]

I built an amp a few years ago that had an unloaded rail voltage of 505VDC. This voltage appears AT EVERY NODE IN FRONT OF THE CHARGING CURRENT. That amounts to THAT voltage appearing across the dielectric. (<period) Some builders, manufacturers included, do not recognize this and only provide an adequate safety margin for the first filter node. It's worth noting that it's not unusual to see a preamp decoupling cap fail before the main ripple filter. Caps don't mind "working" (current) but they do hate over voltage circumstances.

The caps I've switched to are only available up to 450V rating. I could have gone with 500V caps available through another source/brand, but that still leaves no margin for spikes in wall AC or rating. So I did this: (fourteen filter caps!)

[Leo_Gnardo]

Here in my region they began fining energy consumers who had a lot of reactive power on their bills. It's government-run eletrical utility and as such they'll find a way to fine you for everything. They installed electronic measuring systems and the electronics can detect reactive power where the old magnetic wheel meters did not detect anything.

There have been outfits in the USA trying to flog their "money saving" power gadgets for years. Some have capacitor banks. Others have some sort of "magic" transformer or choke system. So far no investigations, fines or court cases that I've heard of. And the companies always tell you "big manufacturing plants use these, so it's a good idea to have them put into your home too." Yeh I can imagine how much an aluminum smelter could save for instance, and have cap/trans banks the size of cargo trailers to do it. The home rigs look like they could fit into a lunchbox.

Folks I hope you'll indulge me in another moment. jmaf's comment brought this on. Maybe someone can provide enlightenment. And if not, I'm sure it would be best to "let sleeping dogs lie" - or is it lay. Hmmm, what ever.

A strange story about my home set up. Couple of years ago a windstorm knocked a big branch off the neighbor's tree, and it hauled the AC supply wires off my house. Neutral and one phase were still connected, the other phase was clearly disconnected at the pig-on-the-pole transformer. I had the old wheel style meter. Although I was using current on the working phase, the meter no longer moved. Of course this situation couldn't last long, it was bit of a hassle hopping over the live lead-in wires laying on the sidewalk. Almost a week later after buying a $100 permit to allow the damage to be rectified, my electrician furnished me with new down-lead wires and even an new breaker box - the old one was way obsolete, rated at only 60 amps, and loaded with weird thermal/overcurrent breakers besides. Power company took this opportunity to install a new digital meter, presumably unaffected by any power factor correction or any other phase funnies. Strangely my electric usage since then measures about 30 to 40% lower than what I had before. I dunno what to ascribe that to. But I'll count my blessings. I got away cheap, about $800 for the electrician and parts but I'll save that over the course of a couple years reduced electric bills. Is there any reason the old meter indicated a higher electrical usage than the new one? I'm avoiding talking to the power company about this. "Twisting the dragon's tail" could bring unwanted attention and they might "fix" my new meter, additionally charge me for a couple years of underbilling.

[Leo_Gnardo]

Chuck you get a like for the label on your reverb tank. And because you deserve it on general principles.

[Chuck H]

Chuck you get a like for the label on your reverb tank. And because you deserve it on general principles.

Thank you

That's the "Double Agent". Two speakers, two power tubes, two channels, two on board effects and it shares loyalties between British and American allegiance Only five preamp tubes and it has a big, wet reverb, bias wiggle trem and topographically changes from a BF Fender to a VOX AC50 with a flip of the switch. One of a kind and this sucker worked the first time I powered it up (he said with no modesty whatever).

[g1]

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

Yes. And I think you meant in parallel arrangement, but it can be done in series (totem pole) arrangement also. See image of 65 Deluxe Rev. filter caps., 220uF 100V in series with 47uF 500V. Note that respective bleeder/balancing resistors are also different from each other (100K and 470K).

[g1]

Is there any reason the old meter indicated a higher electrical usage than the new one? I'm avoiding talking to the power company about this. "Twisting the dragon's tail" could bring unwanted attention and they might "fix" my new meter, additionally charge me for a couple years of underbilling.

You could be right. Or it could be the opposite and your old meter may have been out of cal. and overbilling.
Did you have to turn in your old meter?
Here we do not have privatized electrical utility. You can request cal. for your meter. If it's out and you were over-billed you are entitled to some compensation, but I'm not sure how the details. If it was under-billing before, you do not have to pay up retroactively.
If the meter meets spec., you get dinged a small fee, I think it cost us $25 when we had it checked. In our case it was in spec. so we had to pay the cal. fee.

[Leo_Gnardo]

Did you have to turn in your old meter?

Still wandering with yer indulgence... Yes, the power company took away the old meter. I'm sure it's long been recycled iow scrapped & pulverized. To boot, they were taken over a couple years ago by a Canadian outfit, Fortis Energy, and when it comes to dealing with crustomers they're mean as snakes. Time was, as a customer AND a stockholder too I could get the attention of some minor Vice President of Paperclips and shake some action out of them. Now it's "pay your bill & shut up, eh?" They just gave themselves a 12% raise on gas & electric rates. That oughta buy some back bacon for the brass hats.

[J M Fahey]

Yes. And I think you meant in parallel arrangement, but it can be done in series (totem pole) arrangement also. See image of 65 Deluxe Rev. filter caps., 220uF 100V in series with 47uF 500V. Note that respective bleeder/balancing resistors are also different from each other (100K and 470K).



[ATTACH=CONFIG]48908[/ATTACH]

NOTICE that C32 500V rating "should" have been ample for rated <400V supply ... yet they had to add extra 100V margin with a kludgy dissimilar cap; that speaks volumes about 500V cap reliability.

Personal conspiracy theory: electrolytic Chemistry is excellent up to 400V; reasonably good up to 450V and Marketer's BS at 500V.

Notice 500V rated caps had disappeared for ages from dealerīs shelves, only lately they were reintroduced ... and only for the Guitar Amplifier market ....
Complete with blue or varnished cardboard wrappings, and old brands suddenly reppeared , clearly a Mojo based selling point

[Tom Phillips]

If one were interested in checking the calibration of your utility meter you could try comparing it to the power use registered by a Kill-A-Watt power meter. You would not be able to measure the whole house usage but it would be relatively easy to do the following:
1) Pick a time when you don't need whole house electricity
2) Hook your refrigerator to the Kill-A-Watt monitor. (Doesn't need to be the Refer. Just a convenient example)
3) Turn off all the circuit breakers except the refrigerator circuit
4) Let it run a while and compare the elapsed energy usage registered on the utility meter vs. that registered on the Kill-A-Watt monitor.
Cheers,
Tom

[Leo_Gnardo]

If one were interested in checking the calibration of your utility meter you could try comparing it to the power use registered by a Kill-A-Watt power meter.

Funny thing, I have had a Kill-A-Watt for a long time. And I have used it on my fridge. The current icebox is a low tech wonder that uses only a single KWh per day. Sure would be an inconvenience to continue with the rest of the house off line long enough to make a statistically significant measurement. FWIW I have installed a couple of LED lamps instead of tungstens, but I don't think that was enough to explain the change. So .... I'll enjoy the savings and smile. Meanwhile, any thoughts of installing solar have drifted away. At this point in my life I really doubt solar would enable enough savings to offset the cost of installation. Fortis' net metering would pay me 1.25 cents per KWh while they charge me 18 cents for the juice plus the privilege of having it delivered. Subtract the cost of installation and maintenance, also the $30 monthly upcharge for the privilege of being on a net metering program, I still wind up in the loss column. Let's also factor in I'd need to have a new roof installed so they can expect 30 years of uninterrupted solar service. Don't get me wrong, I've been a BIG fan of solar for a long time. At this point the numbers just don't work out. If I could reliably predict I'd be staying in my house well past age 95 of course it might be a different story.

[the fatch]

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!

Another point to bear in mind is that caps are rated to operate for a number of hours at their rated voltage, and operating a cap at a lower voltage will extend its working life.

[Gtr0]

Yes. And I think you meant in parallel arrangement, but it can be done in series (totem pole) arrangement also. See image of 65 Deluxe Rev. filter caps., 220uF 100V in series with 47uF 500V. Note that respective bleeder/balancing resistors are also different from each other (100K and 470K).



[ATTACH=CONFIG]48908[/ATTACH]

NOTICE that C32 500V rating "should" have been ample for rated <400V supply ... yet they had to add extra 100V margin with a kludgy dissimilar cap; that speaks volumes about 500V cap reliability.

Personal conspiracy theory: electrolytic Chemistry is excellent up to 400V; reasonably good up to 450V and Marketer's BS at 500V.

Notice 500V rated caps had disappeared for ages from dealerīs shelves, only lately they were reintroduced ... and only for the Guitar Amplifier market ....
Complete with blue or varnished cardboard wrappings, and old brands suddenly reppeared , clearly a Mojo based selling point

Thank you the tip, J.M.F - I will keep that in the very most forefront of my mind.

According to Merlin:

Also, resistors must be added in parallel in order to encourage equal voltage sharing between the capacitors. The resistors should be equal to 50/C or less, so two 100uF capacitors would each need a 500000 ohm resistor (470k would be the obvous choice). These also act as bleeders when the amp is switched off.

Found here.

50/C or less - means what exactly? Is there a limit as to how less it could be... according to the schem attached in the quote, it looks to be (50/C)/2. (?)

[the fatch]

Another point to bear in mind is that caps are rated to operate for a number of hours at their rated voltage, and operating a cap at a lower voltage will extend its working life.

Did some research and found that the working life of an electrolytic cap is proportional to max. working voltage/operational voltage. E.g. if a cap rated for 2000 hours at 350V is operated at 250V its working life will be extended to 2000 x 350/250 = 2800 hours.

[Malcolm Irving]

... Is there a limit as to how less it could be... according to the schem attached in the quote, it looks to be (50/C)/2. (?)

Not really, but lower value resistors would create more heat and pull down the B+ voltage a bit - and higher watt rated resistors would be needed.

Low value resistors would balance up the voltages faster.

High values like 1M would still achieve a balance, but it would take more time to get there (in the meantime the cap getting more than its fair share of voltage might blow).