power supply caps in audio output

Recently I purchased replacement caps for the analogue output section of my Pioneer cd player which called for 100uf/ 25v caps. Instead of using the standard two pin caps electrolytic caps (Elna or Cerafine) and because I have had excellent experience with Mundorf's 4-pin power supply caps in power supplies, I decided to swap out the cd player caps with the Mundorf's (the Mundorf's were rated for 100uf @100v) . I asked Mundorf about this idea and they did not recommend it because they said the caps would "break". I asked them to explain, but nothing came back.
I decided to do it any way.
With the correct wiring for these special (and expensive) caps, I installed them and as soon as I turned on the player I heard the wonder of these caps. They really are something special and I loved them for about a month. Then the left channel failed: it started with intermittent popping, and eventually that channel failed completely.
So my question is this to those might know (since Mundorf isn't saying): what is it about a/c which makes these caps incompatible with any thing but d/c? Why can they not be used for a/c or in any output section? Would over-rating the voltage even more help them last longer?
Thanks in advance.
 
I don't have any knowledge specifically about Mundorf capacitors except they get to play with the same laws of physics as everyone else. The dialectric layer in an electrolytic capacitors is formed by polarising voltage across the capacitor. If an electrolytic capacitor is not biased the dialectric layer can erode. Your application provides no bias voltage to maintain the dialectric layer.

Many manufacturers only guarantee their power supply capacitors from to meet spec when 'fresh'. If they have been in storage for a few years, before use it is necessary to apply a small bias voltage across the capacitor to reform the dialectric layer to prevent damage. From the research on capacitor distortion by Bateman over many years I would suggest you use a bipolar electrolytic such as Nichicon Muse, or an appropriate bipolar from your manufacturer of choice.
 
Actually standard polarised electrolytics are usually perfectly fine at passing low voltage, line level, signals. I mean like a 50V DC part wouldn't have any trouble with a 2V line signal, especially a 500V part. These mundorfs must have some internal construction that differs from your typical lytic.
 
Actually standard polarised electrolytics are usually perfectly fine at passing low voltage, line level, signals. I mean like a 50V DC part wouldn't have any trouble with a 2V line signal, especially a 500V part. These mundorfs must have some internal construction that differs from your typical lytic.
I shared with you the same reason that I thought a cap rated for 500v could tolerate a 25v a/c voltage. Could it be that we were both mistaken given the other two suggestions? Obviously the cap did fail so perhaps reverse voltage, ripple, absence of bias, etc. was the culprit.
 
I don't have any knowledge specifically about Mundorf capacitors except they get to play with the same laws of physics as everyone else. The dialectric layer in an electrolytic capacitors is formed by polarising voltage across the capacitor. If an electrolytic capacitor is not biased the dialectric layer can erode. Your application provides no bias voltage to maintain the dialectric layer.

Many manufacturers only guarantee their power supply capacitors from to meet spec when 'fresh'. If they have been in storage for a few years, before use it is necessary to apply a small bias voltage across the capacitor to reform the dialectric layer to prevent damage. From the research on capacitor distortion by Bateman over many years I would suggest you use a bipolar electrolytic such as Nichicon Muse, or an appropriate bipolar from your manufacturer of choice.
Is there any way of providing bias to a cap before it gets installed? Would a small amount of d/c voltage across the cap for a short period of time reform it making it suitable for an a/c application?
 
Is there any way of providing bias to a cap before it gets installed? Would a small amount of d/c voltage across the cap for a short period of time reform it making it suitable for an a/c application?
Anecdotally Peter Walker (Quad) ensured that during the power on phase there was an offset voltage across coupling capacitors to reform or preserve the dialectric layer. This was achieved with asymmetric power supply topology. Even so, many of the coupling capacitors in Quad units are redundant, and the best capacitor in an audio signal path is no capacitor.

My suggestion is to use components appropriately designed for the task in hand as I posted above. If you really want to experiment with the performance of coupling capacitors you should read Bateman's work, rather than taking random shots in the dark. After all it's just applied physics.
 
Well what I would first do is see what operating conditions the coupling caps are going to be subjected to.

In other words turn on and turn off voltages. Like the John said above. These could present reasonably high reverse bias that the capacitor wouldn't be happy to see.

25V AC for the output of a CD player seems extremely high and almost an order of magnitude larger than I was thinking the CD player would output. 5.6V AC is typical.

It's possible that the CD player always has a DC potential on its outputs. Some output configurations do and require the coupling cap to remove the DC. If this is the case then you may have connected the mundorf in so that it was subjected to a reverse bias all the time the CD player was on. Popping, and actual popping rather than simply an intermittent connection, implies a momentary blip of DC getting through, hence the pop. If the CD players output was always biased to -7 volts and you connected the mundorf in the wrong way round that could have caused it to fail.

How much reverse bias a polar lytic can take is dependent on its construction? Obviously the mundorfs can't take much at all.
 
Well what I would first do is see what operating conditions the coupling caps are going to be subjected to.

In other words turn on and turn off voltages. Like the John said above. These could present reasonably high reverse bias that the capacitor wouldn't be happy to see.

25V AC for the output of a CD player seems extremely high and almost an order of magnitude larger than I was thinking the CD player would output. 5.6V AC is typical.

It's possible that the CD player always has a DC potential on its outputs. Some output configurations do and require the coupling cap to remove the DC. If this is the case then you may have connected the mundorf in so that it was subjected to a reverse bias all the time the CD player was on. Popping, and actual popping rather than simply an intermittent connection, implies a momentary blip of DC getting through, hence the pop. If the CD players output was always biased to -7 volts and you connected the mundorf in the wrong way round that could have caused it to fail.

How much reverse bias a polar lytic can take is dependent on its construction? Obviously the mundorfs can't take much at all.
I used 25v in my response not because I had measured it but because the highest rated cap in the circuit is rated for 25v. It probable and likely that the caps see much less voltage than 25v and so my thinking was that a cap rated for 500v could certainly handle whatever voltage was in the system-- even reverse and/or bias voltage (if there was any).
 
my thinking was that a cap rated for 500v could certainly handle whatever voltage was in the system-- even reverse and/or bias voltage (if there was any).
The forward voltage rating has little to do with the level of reverse bias voltage an electrolytic capacitor can handle. Any electrolytic capacitor is likely to quickly break down if a reverse voltage of >1.5V is applied. This is because the insulator is an anodised aluminium dielectric layer with a breakdown voltage in the reverse direction of approximately 1.5V. Above the breakdown voltage the layer is eroded by electrochemical reduction.
 
High voltage electrolytics aren't designed for low voltage use. Having said that using them as a coupling cap for line-level audio isn't really using them at all (only tiny currents involved), so I would not expect any failure that quick. Was the failure you report actually provably down to the change of caps? Channels can fail for many reasons.
 
I cannot be sure. It was my assumption at the time and further discussions here seemed convincing.
However, I would like to think that the problem occurred elsewhere but until I get another cap of some sort to replace it, I won't know for sure. (I do like the sound of the 4-pin cap there). In addition I have ordered another tube for the amp which also started to fail at the same time.
The possibility remains that the cap was ok and the failure was elsewhere since the other one shows no sign of failure; even when I compared the two after measuring their respective resistance and capacitance, it was inconclusive.
 
Thank you gentlemen for your insights and suggestions.
Problem solved.
It was not the Mundorfs at all (though they still might fail later). It was my Jensen bypass polypropylene caps which failed. They were very old and though they measured properly at the beginning, the little voltage and current took out the last bit of life in them.
Now I have music once again. Happy happy joy joy.
 
It was my Jensen bypass polypropylene caps which failed. They were very old and though they measured properly at the beginning, the little voltage and current took out the last bit of life in them.
Oh wow, that's interesting to know! I've not heard of this kind of failure before. The Jensens are oil filled under vacuum; perhaps the vacuum got out of yours. If you had a vacuum pump to keep them full of vacuum it could help, but I don't know how often would it need to run. Did you follow the recommended oil change interval and use an authorised service center? Did you have a lead alignment done at the same time?

I was considering the Jensen Metallised Polypropylene Capacitors for bypass in my passive preamp, but there were some niggling puzzles I couldn't find answers for.* I'm still really interested to hear how they might be solved. What should I do with the preamp's lid which no longer fits? Is it best to place it on the left or the right side? What is the possible effect on balance? Or underneath? And I have heard the may be residual electrostatic fluidizing in the powder-coated lid. Should I strip the powder-coating on the inside, outside, or both?

*Where to put them was one, but I think that issue has been circumvented.

 
Last edited:
I used 25v in my response not because I had measured it but because the highest rated cap in the circuit is rated for 25v. It probable and likely that the caps see much less voltage than 25v and so my thinking was that a cap rated for 500v could certainly handle whatever voltage was in the system-- even reverse and/or bias voltage (if there was any).
Why don't you spend some time on sites like Wikipedia to discover what the properties of electrolytic caps are? And other components.
Lots of good info there, then come here & ask informed questions. Reverse voltage on an electrolytic is not a good condition.
And can result in an explosion.:eek:
 
  • Like
Reactions: 1 user
Why don't you spend some time on sites like Wikipedia to discover what the properties of electrolytic caps are? And other components.
Lots of good info there, then come here & ask informed questions. Reverse voltage on an electrolytic is not a good condition.
Time-consuming research is not "in the books" for DIY'ers.
Just "dive right in" and ignore any professional design because you want to make it better than they did in the research labs.
:yikes:
Oh, and adhere to the "recapping club" that started on the internet blogs, become a member of the sheep herd.