Electrolytic voltage rating

[KatieandDad]

This thread seems to be going round and round in circles.

 

[wintermute]

Quite a few years ago our mains voltage was "harmonised" to bring it in line with other countries that used 230 vac. In practice nothing here changed distribution wise. All that happened was that the tolerance changed from the old 240 -/+ 6% to 230 +10/-6 % which gave pretty much the same limits. Appliances (and I guess new transformers too) are "designed" now for 230 volts. So an old 240 volt 30 v secondary and a "new" 230 v 30 volt secondary are actually different and will give different voltages when fed from the same supply.

Thanks for the explanation Mooly That definitely explains why my caclulation were lower.

If I adjust:

multiply 25V by 1.414. Multiply by 1.07 (for 7% regulation) and then by 1.1 (for the 23V extra that the mains may actually be)... that comes out at 41.6V which agrees with Andrews worst case calculations.

7% is actually a very good regulation figure so in a lot of cases the figures would be even higher.

Again in this case though being class A, if the transformer is running close to it's VA rating it should be fine. I would agree in general with Andrew though that 40V caps are too low for a general 25V secondary.

I'm with Rundmaus and would generally prefer to have a 20% margin on the caps WV rating over he nominal rectified voltage.

katieandad I don't think it is going round in circles. This is an important consideration and I think should be made completely clear as someone less experienced may come along and see the thread and figure that 40V caps will be fine for their situation which may not be the case.

I use exactly the same method as Andrew when determining cap size putting in the actual regulation of the transformer being used when determining absolute minimum values for the filter caps ratings. If this is done then there should be no circumstances (short of the mains being outside of specs) where the caps will be over voltage. I think that this is just good engineering practice. Obviously as stated if the load is known to be constant then the regulation of the transformer can probably be taken out of the equation but the above gives a very good indication of worst case scenarios

Tony.

 

[Mooly]

I get the same figures although maths being my worst subject it took me around 5 minutes to work it out. My eyes glaze over at numbers lol

25 * 7 = 175
175/100 =1.75
1.75 + 25 = 26.75 volts AC. The transformer volts off load for the 230 volt specified transformer.

230 volts in and 26.75 out.
230/26.75 gives the ratio of the tranny. Thats 8.598 : 1

So 230 volts mains at its 10% limit above is 253 volts

253 / 8.598 is 29.42 volts ac

29.42 * 1.414 gives 41.6 volts YAY same answer

It would be interesting to know what alibears mains voltage really is.

 

[TonyTecson]

i aim for 80% of capacitor working volts rating, say a cap is rated at 100 volts, i try to use it at 80 to 85 volts.....

it is the high line voltage that we have to consider, say that in your locality 260 volts is the highest voltage, then we consider the secondary voltage at that high primary line and size the capacitor voltage accordingly...

 

[tinitus]

quote.....The transformer has a secondary of 25 -0- 25 and when rectified under no load is about 37V dropping to 35V when loaded with 1.6 amp bias of the amplifier.

can I please ask, are you talking about measured trafo voltages, or theoretical ?


depends on whether you talk about measured actual voltages, or theoretical voltages
classA can be tough on the caps if they are wrong
but I think the type of cap is more important than exact voltage rating
meaning, it really needs to have high current rating, or it runs too hot

a cap deals better with a small over voltage than too high current

 

[tinitus]

Hi all, I have 8 new Epcos 47000microfarad 40V capacitors .......

if they are the big ones(screw mount/terminals), I wouldn't worry about running them at 41V
they are very tough industrial caps
do you think they would of any use if the brake down at such small over voltage

but maybe they will last only 20 years instead of 25 years

btw, someone once told me many caps with different voltage specs are actually more or less the same cap .... just specced differently
more voltage ... less capacity (same physical size)

 

[KatieandDad]

As 40V is very close to the 42V (ish) that everyone is so worried about, you could always try a CRC circuit with the first C being 63V and safe, R being fairly small say 0R47 and then the voltage across the second C should be safely below 40V.

 

[alibear]

Thanks all for your valuable and informative replies, much appreciated. I will go with the informed replies and up the voltage rating.
Andrew T sorry to be a frustration to you. It's not that I "refuse" to read your posts, on the contrary,over the last two years or so that I have been a member of DIY Audio I have realised that your technical knowledge appears to be immense and your posts are of great help to those with inferior knowledge or experience such as myself. ( no I am not creeping, just telling the truth ) But the Forum is huge and it is not possible for me to read every post.
I will now use some 100,000microfarad 50 V that I have instead.
I am sure that in one of Andrew T posts in the past he has stated that you should bypass filter caps with caps of 100th the size eg.
100,000 bypassed with 1,000, then bypassed with 10, then 0.1microfarad.
Is this correct? or am I dreaming

 

[KatieandDad]

50V is an outdated voltage, replaced by 63V in most cases.

How old are the caps? They might need reforming.

 

[alibear]

The capacitors are at my place of work at the moment, will have to wait until Monday to look at them closely. If need be I will post a photo of the details on them to try to assertain their age. From what I remember they appear never to have been used.
Thanks for help
Alan

 

[KatieandDad]

Old caps dry out so be careful.

I would reform them anyway and use that check to see if they are still serviceable.

 

[tinitus]

100,000 bypassed with 1,000, then bypassed with 10, then 0.1microfarad.
Is this correct? or am I dreaming

I can imagine you are not dreaming, and that it have been said .... just forget about it

when things are so 'tightly matched' you need to measure your actual voltages ..... with your trafo
or else its all just theoretical, and more or less waste of time

but if you want to design for '100% safe everywhere' .... then it would be different, and Andrew would be right

 

[palstanturhin]

It is good that this case has come to a resolution and better suited capacitors have been found.

Therefore the rest of my input can totally be ignored:

We know he had 40V rated Epcos capacitors.
We know he had measured 37V no load output voltage.
We know he had measured 35V at 56W.
We assume he has 0.3V diode voltage drop at no load.
We therefore know he has 26,4VACrms at his transformer output at no load, which is really close to a what ever toroid transformer available, for example a TalemaNuvotem.
If the transformer was rated at 230Vin, we know that he must have measured the output at quite precisely at that exact voltage.
If the transformer was rated at 240, we know that the input voltage during his measurements must have been 240v...

So we assume he measured the output at (the worst case) 230Vin.
Now: We take the British over voltage situation of 254V input (at that worst case 230v transformer of ours).

254/230x35=38,65Vout!
And that will be the worst case situation!

But as I said, ignore this...

 

[vacuphile]

Derating voltage is advised to increase lifetime.

http://electrochem.cwru.edu/encycl/misc/c04-appguide.pdf

At this location, elco's are bound to be operating at >40 degrees and at high ripple currents. So, I would say it is really borderline and even slightly across it.

 

[tinitus]

Derating voltage is advised to increase lifetime.

says at operating temperature up to 45degr, no voltage derating needed

and in these terms, I suppose a hifi classA amp is considered as low current

 

[vacuphile]

The major ripple is from the bridge diodes transducting only briefly each cycle. In that brief period, all the Coulombs needed to maintain the current driving the amplifyer have to enter the capacitor. So, the more current you draw, the higher the ripple current in the caps. ClassA for that reason creates per output Watt larger ripple currents in the PS caps than classAB.

45 degrees will easily be seen inside a power amp, it's just over 20 degrees over normal living room ambient temperature. The caps will also heat up by themselves. Anyways, it will work without derating, but it is not good for longevity.

 

[tinitus]

high current would be something like welding
and those caps may be rated with 20A ripple current...each
if they hear we call our amps high current, they will probably laugh at it

 

[vacuphile]

I am not pulling it out of my hat. Please read Elliott Sound Products - Linear Power Supply Design under 'capacitor ripple current'.

 

[wintermute]

So we assume he measured the output at (the worst case) 230Vin.
Now: We take the British over voltage situation of 254V input (at that worst case 230v transformer of ours).

254/230x35=38,65Vout!
And that will be the worst case situation!

But as I said, ignore this...

Unfortunately this is where it goes wrong. The worst case is in fact 213V not 230V as the standard is 230V -6% + 10%

Without knowing the actual mains voltage at the time of the test then all anyone is doing is making assumptions. Yes it is likely that the 40V caps will be ok under the circumstances when load + diode drops are taken into account (as I said earlier).

The maths is not hard, for anyone starting from scratch it doesn't hurt to do it and be safe knowing that no matter what the load is the caps should be fine.

Personally I'm not one to cut corners (in anything). I've too many times seen situations where if a little more had been spent in the first place, the overall cost would have been a lot lower. Everyone though can make their own choices.

The power in my street often goes to 252V, the suburb I used to live in was normally around 238V. If I'd used caps in a power supply that were marginal at 240V, then when I moved house I probably would have had problems.

Tony.

 

[palstanturhin]

I just calculated his input voltage, It was 230!
Error margins included 39.5Vout - still under 40V!
And long life expectations to his capacitors...
230-6% is totally irrelevant here.
He has measurements!
What kind of engineers are you?
He has MEASUREMENTS!
Use them!

To make more accurate calculations we do not need to know the input voltage.
We need to know his transformers make and model.
And we need to know his diodes make and model.

But we can ignore all this now.
There has been an other solution to the case...