If capacitor longevity is your aim, you would want to run them at a lower voltage than their quoted working voltage. I normally aim for 90% of this voltage for this type of application.
Transformer regulation is important in the calculations. For example, a 38-0-38Vac transformer with a quoted regulation figure of 6% will produce 38 + 6% Vac off-load, or 40.3-0-40.3 Vac. Add to that the worst-case mains tolerance (+/- 10%, I believe in the USA) and potentially, you could see 44.3-0-44.3V. This would result in a DC supply of around 61V DC. This is perhaps a little too close to call for 63V capacitors but is a reasonable worst-case assessment.
If your transformer is recycled, measure it's off-load voltage and the mains input voltage. Then calculate the output voltage when the mains is at the top of its tolerance (120V + 10%). You might be fine with 36-0-36V types. Generally, higher power transformers have a better regulation figure than their lower power counterparts. Hope that helps.
Under what conditions did you obtain that figure?
Transformer under no-load / full-load?
Mains voltage at min / max / nominal level?
Even at no load with 5% regulation typical of a transformer we would use for this application and mains voltage 5% higher than rated, you still should be fine, at ~58V. I use a 600 VA 40-0-40 V transformer for my HB and biased at 50 mA per device I end up with 57 V rails. If that's too tight for you, just use 80V caps. No need to over analyze it. 
You feel the need to criticize me when the nominal voltage is 10v lower than the cap's nominal rating? What's the point in that? Do you really think that a 38-0-38vac transformer will stress 63v caps?
I am not criticising you in the slightest, Terry. All I am showing to anyone interested is that adding up the worst-case parameters with a 38-0-38V transformer, off-load, taking into account mains tolerances, might bring the rail voltage near to the limit of 63V as Daniel had asked. Good design practice should examine such scenarios.
secondary voltage is Mains supply * rated secondary / rated primary * transformer regulation.
Maximum peak is sqrt(2) * sec volts
Maximum smoothing cap voltage is max pk - 1V
ex:
a 115:38+38Vac 7% transformer fed from 127Vac results in sec Vac = 127*76/115*1.07=89.8Vac
Max peak=1.414*89.8=127Vpk
max smoothing cap voltage = 127-1 = 126Vdc
If that is equally split between two caps then each cap sees 63Vdc.
Any interference will use up the spike tolerance of the capacitor.
Maximum peak is sqrt(2) * sec volts
Maximum smoothing cap voltage is max pk - 1V
ex:
a 115:38+38Vac 7% transformer fed from 127Vac results in sec Vac = 127*76/115*1.07=89.8Vac
Max peak=1.414*89.8=127Vpk
max smoothing cap voltage = 127-1 = 126Vdc
If that is equally split between two caps then each cap sees 63Vdc.
Any interference will use up the spike tolerance of the capacitor.
Last edited:
This advice is unsafe because it does not consider worst case conditions and yet gives the mistaken impression that there is nothing to be worth checking !!!!
Take an Antek 6438, specified at 38.6V no load, that's 10% higher than specified, feed it 126V (5% high) and you get 62.4V. Most US mains voltages I've seen don't exceed 120V and I've used a fair number of Antek transformers and never seen one more than a couple percent high. So, the scenario is unlikely, and still doesn't exceed rated voltage.
As DIYers we should have a good understanding of good design practices, but we aren't designing our projects for maximum life at maximum operating conditions. AFAIK, temperature has more impact on a capacitor's life than being close to max rated voltage. Keep them cool and you'll have decent life. Even if you cut the life to 2,000 hours, that doesn't mean that they blow up at 2,000 hours. My HB is running much of the time I am home, but that still only amounts to an average of 3-4 hours a day. I'm not as fast a builder as Terry, but a year or two and I move on to a new amp, with new caps in its power supply. "Warranty" repairs if required would be part of the fun of DIY.
As DIYers we should have a good understanding of good design practices, but we aren't designing our projects for maximum life at maximum operating conditions. AFAIK, temperature has more impact on a capacitor's life than being close to max rated voltage. Keep them cool and you'll have decent life. Even if you cut the life to 2,000 hours, that doesn't mean that they blow up at 2,000 hours. My HB is running much of the time I am home, but that still only amounts to an average of 3-4 hours a day. I'm not as fast a builder as Terry, but a year or two and I move on to a new amp, with new caps in its power supply. "Warranty" repairs if required would be part of the fun of DIY.
In the example I gave in post #1301, I calculated 61 VDC with the following assumptions:
a) The transformer regulation was 6% (not unrealistic, but could be higher or lower).
b) The mains in the USA is 120V +/- 10%.
"... a 38-0-38Vac transformer with a quoted regulation figure of 6% will produce 38 + 6% Vac off-load, or 40.3-0-40.3 Vac. Add to that the worst-case mains tolerance (+/- 10%, I believe in the USA) and potentially, you could see 44.3-0-44.3V. This would result in a DC supply of around 61V DC".
Voltage regulation in transformers is the difference between the no load voltage and the full load voltage. This is usually expressed in terms of percentage.
What are the implications of this?
A 63Vdc cap will last a long time when operated at 63Vdc.
If the mains is @ 127Vac permanently and the there is no interference the 63V cap will still last a long time.
But Mains always has interference.
All the spikes above 127Vac will enter the cap as spikes above 63Vdc. Those spikes wear out the capacitor, using up life.
If the cap were operated at 60Vdc + the same spike incidents then it will last longer. and similarly if operated at 50Vdc+spikes it will last longer still.
63V caps do not blow up at 63.01Vdc.
Their life is reduced each time a spike uses up a little bit of their spike tolerance.
Last edited:
Take an Antek 6438, specified at 38.6V no load, that's 10% higher than specified, feed it 126V (5% high) and you get 62.4V. Most US mains voltages I've seen don't exceed 120V and I've used a fair number of Antek transformers and never seen one more than a couple percent high. So, the scenario is unlikely, and still doesn't exceed rated voltage.
As DIYers we should have a good understanding of good design practices, but we aren't designing our projects for maximum life at maximum operating conditions. AFAIK, temperature has more impact on a capacitor's life than being close to max rated voltage. Keep them cool and you'll have decent life. Even if you cut the life to 2,000 hours, that doesn't mean that they blow up at 2,000 hours. My HB is running much of the time I am home, but that still only amounts to an average of 3-4 hours a day. I'm not as fast a builder as Terry, but a year or two and I move on to a new amp, with new caps in its power supply. "Warranty" repairs if required would be part of the fun of DIY.
Both of these replies failed to mention the rated primary voltage and what effect that has on the worst case operating conditions.
In the UK we can have 220Vac or 230Vac or 240Vac rated transformers.
The equivalent in the USA and Canada would be 110, 115, and 120Vac.
I agree. The transformer primary rated voltage should be chosen to match the local mains voltage and not be significantly lower.
As an aside, my company imports motorised equipment from Germany and they have to fit 240V mains transformers for our UK market. So much for European harmonisation.
I have lived 62 years in the USA and have NEVER seen 127V on the mains. I have a Watts Up meter plugged into an outlet above my work bench at all times. The highest I have ever seen is 122V. Lets suggest everyone build amps so they can weld with them as long as we are spending other peoples money.
Once upon a time I wanted to use truly good caps in PS so I opted for 63V RIFAs, expensive but good. What worried me was that with 2x40V 625VA toroid with 5% regulation and potential of mains reaching ocassionally 255V (nominal 240V) I'd be puting caps into danger zone. 100V rifas were too expensive. I asked manufacturer and reply as usual was a bit evasive.
All depends on caps capabilities withstaning surge currents and short term transients as well as mains overvoltage periods and manufacturers margins. The latter one can learn from someone inside as actual Ur of caps varies from the nominal. Depending on adopted QC standards one can get 65V cap rated 63V with certain probability.
When it comes to surge and transients different manufacturere adopt somewhat different standards here (multiplier (m) for 1.15 to 1.21 (Us=Ur*m) and various periods of time and different frequency of repetitions and temperature) so if one demands longevity from caps then ons should "derate" Ur by a sensible margin or check if under worst V conditions Ur is not exceeded. Caps working well below their Ur and well below their temp ratings have many times longer lifespan than their nominal.
In reality usually the problem is with mains being lower than nominal and higher they are usually at night when most are asleep so as still4given monitors US mains voltage then there is no reason to disbelieve him and thre is notable chance that true Ur of caps is somewhat higher than their nominal Ur.
cheers,
All depends on caps capabilities withstaning surge currents and short term transients as well as mains overvoltage periods and manufacturers margins. The latter one can learn from someone inside as actual Ur of caps varies from the nominal. Depending on adopted QC standards one can get 65V cap rated 63V with certain probability.
When it comes to surge and transients different manufacturere adopt somewhat different standards here (multiplier (m) for 1.15 to 1.21 (Us=Ur*m) and various periods of time and different frequency of repetitions and temperature) so if one demands longevity from caps then ons should "derate" Ur by a sensible margin or check if under worst V conditions Ur is not exceeded. Caps working well below their Ur and well below their temp ratings have many times longer lifespan than their nominal.
In reality usually the problem is with mains being lower than nominal and higher they are usually at night when most are asleep so as still4given monitors US mains voltage then there is no reason to disbelieve him and thre is notable chance that true Ur of caps is somewhat higher than their nominal Ur.
cheers,
I didn't dawn on me that most of you are in Europe. I was wondering where you came up with the idea that mains hit 127V here in the states. My guess is that you needed that number to make the math work out to support your stand that 63V caps are not sufficient for a transformer that will likely spend it's days producing +-50V when connected to an amp. There are no 127V mains. Not going to happen. The question was asked by someone here in the states so I feel pretty good about my answer. So go ahead an plan for the "worst case" but if that requires 127V mains you're barking up the wrong tree.
The calculation method I offered has nothing to do with geographical location, given that the correct transformer primary voltages are adopted for that location. So, for a nominal mains voltage of 120Vac, the transformer would normally be chosen to have a 120Vac primary rating.
Please re-read my posts #1301 and #1311:
#1301:
Transformer regulation is important in the calculations. For example, a 38-0-38Vac transformer with a quoted regulation figure of 6% will produce 38 + 6% Vac off-load, or 40.3-0-40.3 Vac. Add to that the worst-case mains tolerance (+/- 10%, I believe in the USA)...
#1311:
In the example I gave in post #1301, I calculated 61 VDC with the following assumptions:
a) The transformer regulation was 6% (not unrealistic, but could be higher or lower).
b) The mains in the USA is 120V +/- 10%.
Why you might believe I needed to deceive the OP by manipulating the figures is beyond my comprehension.
The whole point of my post was to show the calculation method involved. Initial search results I discovered for mains tolerances in your country were +/- 10%. If you know this figure is more like +/- 5% (or whatever the actual figure is), why not simply state the same and apply it in your own calculations?
Confucius say: "Give a welder a voltmeter on his birthday and he will appear happy, but not be entirely fulfilled. Give him a bag of six-inch nails and his eyes will light up".
For the sake of completeness, I'd like to mention the effect of the rated transformer primary voltage. Here in the UK, we can obtain transformers with primary ratings of 220, 230 and 240Vac. In the USA and Canada, this might be 110, 115, and 120Vac.
Our UK mains tolerance is 230V +10% - 6% and a newer EU harmonisation figure to include the UK in Western Europe states 230V +/- 10%. In either case, the maximum voltage the UK should expect is 254.4Vac.
In the examples below, I will use three transformers all having a 0-38Vac secondary rating and one of 220Vac, 230Vac and 240Vac primary ratings. These will be transformers a), b) and c) respectively. These transformers will be assumed to all have a regulation figure of 5%.
To determine the maximum off-load ac secondary voltage, divide the maximum mains input voltage by the transformer's rated primary voltage, add the regulation percentage and use this as a multiplier for the rated secondary voltage:
Transformer a): ( ( 254.4 / 220 ) + 5% ) x 38V = 46.1Vac
Transformer b): ( ( 254.4 / 230 ) + 5% ) x 38V = 44.1Vac
Transformer c): ( ( 254.4 / 240 ) + 5% ) x 38V = 42.3Vac
Connecting a standard bridge rectifier (two diode drops of around 0.6V each at low current) and smoothing capacitors gives DC levels of approximately:
Transformer a): (46.1Vac - 1.2V) x 1.414 = 63.5VDC
Transformer b): (44.1Vac - 1.2V) x 1.414 = 60.7VDC
Transformer c): (42.3Vac - 1.2V) x 1.414 = 58.1VDC
This is a significant deviation from a nominal mains voltage situation where the input voltage matches the rated transformer primary voltage and the transformer is operating under full load. Under such circumstances, the output would be around 51VDC.
Our UK mains tolerance is 230V +10% - 6% and a newer EU harmonisation figure to include the UK in Western Europe states 230V +/- 10%. In either case, the maximum voltage the UK should expect is 254.4Vac.
In the examples below, I will use three transformers all having a 0-38Vac secondary rating and one of 220Vac, 230Vac and 240Vac primary ratings. These will be transformers a), b) and c) respectively. These transformers will be assumed to all have a regulation figure of 5%.
To determine the maximum off-load ac secondary voltage, divide the maximum mains input voltage by the transformer's rated primary voltage, add the regulation percentage and use this as a multiplier for the rated secondary voltage:
Transformer a): ( ( 254.4 / 220 ) + 5% ) x 38V = 46.1Vac
Transformer b): ( ( 254.4 / 230 ) + 5% ) x 38V = 44.1Vac
Transformer c): ( ( 254.4 / 240 ) + 5% ) x 38V = 42.3Vac
Connecting a standard bridge rectifier (two diode drops of around 0.6V each at low current) and smoothing capacitors gives DC levels of approximately:
Transformer a): (46.1Vac - 1.2V) x 1.414 = 63.5VDC
Transformer b): (44.1Vac - 1.2V) x 1.414 = 60.7VDC
Transformer c): (42.3Vac - 1.2V) x 1.414 = 58.1VDC
This is a significant deviation from a nominal mains voltage situation where the input voltage matches the rated transformer primary voltage and the transformer is operating under full load. Under such circumstances, the output would be around 51VDC.