Capacitor Types Explained

[chartz]

Honestly I don’t believe in the sound of different good origin fresh new caps. I tried, and failed to hear anything.
But electrolytics will degrade over time, thus compromising sound quality. Treble is usually the first to go.
Wima et al will never degrade. I find Philips multicoloured caps are also completely reliable.
Once I found a terrible sounding new one, it was a fake Chinese one.

 

[Arkless Electronics]

Honestly I don’t believe in the sound of different good origin fresh new caps. I tried, and failed to hear anything.
But electrolytics will degrade over time, thus compromising sound quality. Treble is usually the first to go.
Wima et al will never degrade. I find Philips multicoloured caps are also completely reliable.
Once I found a terrible sounding new one, it was a fake Chinese one.

There are a few places where a cap can be pretty critical to sound quality but they are few and far between IME.
Only on a couple of occasions have a I heard a big effect and they were in said critical applications. A 100uF tant sounded bloody awful in one such application and normal service was resumed by replacing it with a polyprop.
All capacitor applications and types are exaggerated into having night and day differences on an almost hourly basis on forums etc of course.
As I've mentioned before I have set up a jig to allow me to add all sorts of different caps all in the range about 1 - 10uF to the signal path between pre and power amp. I think I could hear the electrolytic but it was such a slight difference I couldn't convince myself. Polypropylene, polyester and polycarbonate all sounded identical as far as I could tell. Again I couldn't convince myself that I could hear any difference between bypassed and any of these cap types in signal path...
Even with the electrolytic the effect of moving my head a couple of inches in any direction was infinitely bigger than the difference between electrolytic and no cap! Hence when one reads tales of night and day improvements, sometimes by changing just for an identical spec capacitor of a different brand, I would take it with a sack of salt!
I do use good quality caps of normal provenance in my work, "just to be sure", for reliability and yes because customers expect certain types to be used. I don't use "boutique" caps as once you are using a quality polyprop at say £4 each there is nothing further to be gained by using £30+ each ones which can sometimes be a rebranded version of the £4 one but with a black outer sleeve and gold writing etc

 

[sq225917]

Styrene for small signal duty, lytics for heavy psu duty, smt ceramics where nothing else does and film caps pretty much wherever they fit.

The only location where I think models makes a difference is in the feedback position, a la 47uf in naim circuits.

 

[Arkless Electronics]

A problem with SMD is that most of the best capacitor types are not available in SMD versions and in cheaper applications high K ceramic tend to get used... PEN and PPS can be used in SMD types and offer some solace but no polyprops or polystyrene etc

 

[martin clark]

There are a few places where a cap can be pretty critical to sound quality but they are few and far between IME.
...
I do use good quality caps of normal provenance in my work, "just to be sure", for reliability and yes because customers expect certain types to be used. I don't use "boutique" caps as once you are using a quality polyprop at say £4 each there is nothing further to be gained by using £30+ each ones which can sometimes be a rebranded version of the £4 one but with a black outer sleeve and gold writing etc

That I think is a fine summary on its own.

i. Use good-quality industrial branded parts in any use.
ii. Don't worry too much; and (while ceramics remain a thing apart as already discussed) - the key issue for coupling and feedback-loop uses is just using one cap large enough in capacitance there is a minimum of ac voltage 'dropped' across it in use, in the bandwidth of interest. A 'bigger' value cap will do better simply for this reason; and even a 'crappy' huge electrolytic can do well-enough for the same reason.

...Which is where the OP came in - just picking a bigger cap to suit his condition of use.

Is there a level of finesse beyond such a simple 'cut'? I think so - but its a very long way into the long grass. In DIY we are free to choose to chase that, but do it on an informed basis. And don't buy giant w -anky handmade caps on the basis of nice photos and reveries from web-shill-ijits you've never heard of - what you need is the consistency, and QC, of a mass-produced item; and in locations that matter - also, usually, the physically-smallest appropriate thing.

ETA: reasons for these statements = whole other rant.
ETA2: next time you see some marketing suggest 'we've used amazing polyprops in the feedback circuit' but they are the size of beercans = point lost... layout considerations matter, even more...

 

[mansr]

Precisely why I said not to use ceramic over 1000pF.... C0G (AKA NP0) are available up to around this value.

C0G is readily available up to 100 nF and for silly money as high as 470 nF.

For power supply decoupling close to ICs, X7R is fine. Physically larger capacitors necessitate longer, and thus higher inductance, traces. Not much of an issue for analogue audio, but it matters with high-speed digital devices.

 

[JimmyB]

Was/Is there any basis for those DNM caps that I seem to remember they used. T-caps or something.

 

[Gervais Cote]

All in all the only cap types needed are electrolytic, polypropylene and polystyrene. For ALL jobs in audio one of those types will be the best you can get.

Agree 100% according to the multiple tests I performed myself on amplifier, preamp tuner and so on but does the same apply for speaker crossovers ?

 

[Arkless Electronics]

C0G is readily available up to 100 nF and for silly money as high as 470 nF.

For power supply decoupling close to ICs, X7R is fine. Physically larger capacitors necessitate longer, and thus higher inductance, traces. Not much of an issue for analogue audio, but it matters with high-speed digital devices.

Much above 1000pF and they are likely to be non C0G unless you make a point of specifying, double checking, that they are C0G was my point...

You can get SMD MLCC ceramic caps of 10 and even 100uF for decoupling which have about the lowest ESR and inductance available but they should never see an audio signal of course.

 

[jagdesign]

Thanks for all the input guys, think I've sort of got my head around it

 

[orangeart]

Don't forget caps don't just have capacitance, they have some series resistance too so a cap used in a filter will give a different transfer function as it interacts with it's load often.

 

[Tony L]

Don't forget caps don't just have capacitance, they have some series resistance too so a cap used in a filter will give a different transfer function as it interacts with it's load often.

Indeed. This is why I’ve been trying to make a point about replacing like for like unless you consciously want to alter the voicing of the component. As I understand it PIO caps have very slight leakage and series resistance from new, the valve circuits they were placed into clearly factored this in as it was well known and designed/voiced the product accordingly. Replace them with some “perfect measuring” modern film caps or whatever and you change the voicing, and almost always not for the better. A real not seeing the wood for the trees thing.

 

[Arkless Electronics]

Don't forget caps don't just have capacitance, they have some series resistance too so a cap used in a filter will give a different transfer function as it interacts with it's load often.

Only really likely with speaker crossovers which I guess is what you mean here. Even then it's way more likely with some filter topologies than others of course. If a crossover is made with film caps it is no longer relevant as their ESR is so low. It would effect most those crossovers which have been made with the cheapest and nastiest electrolytics with the highest ESR. If the crossover has been designed to allow for this then using better caps can indeed put things out of kilter.

 

[Arkless Electronics]

Indeed. This is why I’ve been trying to make a point about replacing like for like unless you consciously want to alter the voicing of the component. As I understand it PIO caps have very slight leakage and series resistance from new, the valve circuits they were placed into clearly factored this in as it was well known and designed/voiced the product accordingly. Replace them with some “perfect measuring” modern film caps or whatever and you change the voicing, and almost always not for the better. A real not seeing the wood for the trees thing.

I completely disagree. PIO do have worse leakage from new than other types... about the worst of the many ways in which they are poor compared to modern ones, but if it's high enough to have any effect on anything then the cap is a dud. Whilst higher than a polyprop it should still be completely negligible. It would not have been factored in and no way would they have used PIO if they had any other options.... they were got shot of PDQ when better tech became available. Series resistance has no effect whatsoever in a coupling cap in a valve amp.
Such amps would not have been "voiced" in any way involving capacitors and nor is this even particularly possible.
With modern amps replace like with like but I would do the exact opposite in a vintage amp as there will almost always be a modern part available which will be vastly better than the original and which they would have used if they had been available. Modern resistors don't drift in value, hiss or crackle and modern caps don't go leaky. In some areas performance can be improved, such as where bulk capacitance is required but the tech of the time restricted them to say 16uF and now we can fit a 1000uF part instead.

 

[mansr]

Much above 1000pF and they are likely to be non C0G unless you make a point of specifying, double checking, that they are C0G was my point...

You can get SMD MLCC ceramic caps of 10 and even 100uF for decoupling which have about the lowest ESR and inductance available but they should never see an audio signal of course.

I'm assuming Farnell's search function is truthful. If you were meaning to say that C0G is the most common ceramic type up to around 1000 pF, then I agree.

 

[orangeart]

Indeed and that is where most of my experience is. Of course loads of crossovers especialy older ones have electrolytic caps. Changing them for foil of course makes a huge difference to the sound, which is often perceived as improvement because all of a sudden there are things appearing in the mix that just were not heard before. Of course in reality, there is now a whacking great hump in the FR where that flute is playing that just wasn't there previous to the 'upgrade'

 

[Arkless Electronics]

I'm assuming Farnell's search function is truthful. If you were meaning to say that C0G is the most common ceramic type up to around 1000 pF, then I agree.

Most common was a part of it but there was more to it than that... I was also trying to put noobs off from even looking for ceramic caps above about 1000pF to avoid the wrong dielectric being used!

 

[Tony L]

With modern amps replace like with like but I would do the exact opposite in a vintage amp as there will almost always be a modern part available which will be vastly better than the original and which they would have used if they had been available.

We will never see eye to eye on this. The overwhelming consensus, and it really is overwhelming amongst those who really know and love vintage valve amps, be they hi-fi or guitar, is to replace like for like (type wise) wherever it is possible to do so. The indisputable fact is say a Leak rebuilt with poly caps and metal film resistors etc is actually worth far less than a non-working untouched example, yet rebuild them properly/tastefully and you can add £hundreds to the value. Basically it is far better to start from scratch than have to pull someone else’s work out and start again. I’d never advocate anyone spending good money to actually devalue a valuable classic product, which is why I will always challenge your opinion (and that is all it is) on this subject on this website.

 

[Arkless Electronics]

We will never see eye to eye on this. The overwhelming consensus, and it really is overwhelming amongst those who really know and love vintage valve amps, be they hi-fi or guitar, is to replace like for like (type wise) wherever it is possible to do so. The indisputable fact is say a Leak rebuilt with poly caps and metal film resistors etc is actually worth far less than a non-working untouched example, yet rebuild them properly/tastefully and you can add £hundreds to the value. Basically it is far better to start from scratch than have to pull someone else’s work out and start again. I’d never advocate anyone spending good money to actually devalue a valuable classic product, which is why I will always challenge your opinion (and that is all it is) on this subject on this website.

No we are at polar opposites on the subject and likewise, I will always challenge your opinion. The difference is yours is indeed opinion, and that of a technical lay person, whereas I bring electrical fact to the subject. The desirability or otherwise of authenticity in vintage equipment is one where opinion and personal taste are the main factors. The actual technical merits of electronic components are not. As in so many areas of hi fi "the overwhelming opinion" one will read is that of non engineers and based on hearsay, Chinese whispers, what journalists say etc combined with commercial vested interest and is almost always simply wrong.

I cannot spell out more clearly that personally I have no interest in the value of vintage equipment or in maximising it. My aim is ALWAYS to IMPROVE vintage equipment and to bring out the very best performance from it.
I am often contacted by people asking me to modify their modern amp for better performance. This is not always easy as the designers have access to the same parts and technology that I do. With vintage equipment however many original aspects were so poor that transformations can be accomplished and the amp made literally twice as good. Generalising, the older something is the more "night and day" an improvement is possible!

Those with say a Leak Stereo 20 for rebuild should be aware that if they bring it to me I will hugely improve it. If this is not what they want then they should take it elsewhere.

 

[mansr]

Most common was a part of it but there was more to it than that... I was also trying to put noobs off from even looking for ceramic caps above about 1000pF to avoid the wrong dielectric being used!

The mistake is thinking of "ceramic" as a single type. Avoiding them entirely just because some types are unsuitable for some applications is throwing the baby out with the bathwater. By the same logic, electrolytics should be shunned too since they are polarised and blow up if used incorrectly.