Coupling capaitors are generally used to block different DC voltages between stages.
You can use a coupling capacitor as a safety feature where there is no DC voltage present.
Where the AC voltage exceeds the DC voltage, one would need to use a non polar plastic capacitor or a Bi-Polar electrolytic capacitor.
Where the AC voltage can never exceed the DC voltage one would use a polar electrolytic if the AC impedance needs to be moved down below the frequency range of signals that need to pass through. Using a Bi-Polar here should not give any benefit.
You can use a coupling capacitor as a safety feature where there is no DC voltage present.
Where the AC voltage exceeds the DC voltage, one would need to use a non polar plastic capacitor or a Bi-Polar electrolytic capacitor.
Where the AC voltage can never exceed the DC voltage one would use a polar electrolytic if the AC impedance needs to be moved down below the frequency range of signals that need to pass through. Using a Bi-Polar here should not give any benefit.
Oh absolutely. I was just wondering if the specifications can in fact point out some objective selection criteria between some of the affordable and commonly obtainable capacitors for coupling usage. Although I prefer the sound of the Elna Silmic IIs in that application, I have no idea if this is reflected by the specifications in any way. As far as I can tell, it sounds indistinguishable from metalized polypropylene and non-metalized polypropylene plastic capacitors for identical applications. Of course, that's just me and my equipment/music. I'm curious what everyone else has to say about that. Also, I have tried back-to-back polarized Silmic IIs to mimic a bipolar, I am not sure I can hear the difference in applications without DC-bias either.
There is a body of opinion that believes that when there is no DC voltage across the coupling capacitor and no AC voltage acrosss the coupling capacitor then the capacitor can never contribute any distortion to the AC signal passing through it.
I am one of those believers.
This seems to point to most of the parmaters used to assess capacitor performance as having no influence on the capacitors ability to pass the signal.
There is one parameter: Dielectric Absorption (DA), which may have an audibility concern. I don't know, but J.Curl and a couple of others have mentioned it.
I am undecided.
It would seem that if the F-3dB frequency is moved low enough then the capacitor can even be an electrolytic, even though this is agreed to be the worst performer as far as passing audio signal goes, is good enough to have no audible effect.
I am convinced.
I use two polar back to back, or four polar back to back in parallel with front to front, where I know there could be insufficient DC to prevent reverse polarity.
I am one of those believers.
This seems to point to most of the parmaters used to assess capacitor performance as having no influence on the capacitors ability to pass the signal.
There is one parameter: Dielectric Absorption (DA), which may have an audibility concern. I don't know, but J.Curl and a couple of others have mentioned it.
I am undecided.
It would seem that if the F-3dB frequency is moved low enough then the capacitor can even be an electrolytic, even though this is agreed to be the worst performer as far as passing audio signal goes, is good enough to have no audible effect.
I am convinced.
I use two polar back to back, or four polar back to back in parallel with front to front, where I know there could be insufficient DC to prevent reverse polarity.
The coupling cap in the LM3886DR is metalized polypropylene.
As far as the perceived sonic differences between Brand A and Brand B and how they may/may not correlate with measurements of various parameters, I suggest to measure a bunch of caps and setting up a double-blind experiment with a significant participant population (N > 30 preferably) to determine whether there is in fact an effect or if the perceived sonic differences are really just confirmation bias. Alternatively, you could visit a college/university library and search the academic journals for such an experiment. I suspect Sean Olive and Floyd E. Toole have looked at it already. While such an experiment would be interesting to conduct, it's a bit beyond the scope of this thread. Should you wish to undertake it, I suggest starting a new thread in the Parts forum.
I'm not aware of any science that indicates any issues with the PP cap in the LM3886DR. My measurements back this up as well.
Tom
@Tom and Andrew,
Thanks for the comments.
I have no doubt that Tom's LM3886DR is as good as that stated by the measurements. I actually have the Modulus 86 V2.0 that I still have to find an enclosure for. It is nice to hear your summaries because this field (electronic engineering, particularly analog design) is not my profession and some of the opinions (particularly regarding boutique parts) are difficult to substantiate for me.
Thanks for the comments.
I have no doubt that Tom's LM3886DR is as good as that stated by the measurements. I actually have the Modulus 86 V2.0 that I still have to find an enclosure for. It is nice to hear your summaries because this field (electronic engineering, particularly analog design) is not my profession and some of the opinions (particularly regarding boutique parts) are difficult to substantiate for me.
That certainly seems to be a good advice and I thank you for it. I am a little unclear regarding what can be considered "boutique" though - would Nichicon KZ or Elna Silmic II qualify ? They are a bit more expensive than general purpose electrolytics but not substantially so. What about Wima polypropylene capacitors ?
Some years ago I was having trouble getting a SKA to work properly.
It turned out that the SKA was suspectible to the PSU ringing. It had Wima MKP caps across the rectifier as suggested by the Designer.
I changed to MKT and then it was much better. Later it blew up.
There was clearly an instability problem and that made the low ESR MKP caps incompatible. Shame the Designer did not specify a snubber for the PSU rather than lonely caps.
The major manufacturers are reliable. They make dozens of different duty devices to suit their customers. If Audio needs some oddballs to perform well, then there is probably something wrong with the Audio design, not with the component.
But don't let a capacitor (anywhere) become an inadvertant filter !
Filters will alter the wanted signal. Generally our amplifiers are required to PASS the signal, not to filter the signal. We filter the interference instead.
It turned out that the SKA was suspectible to the PSU ringing. It had Wima MKP caps across the rectifier as suggested by the Designer.
I changed to MKT and then it was much better. Later it blew up.
There was clearly an instability problem and that made the low ESR MKP caps incompatible. Shame the Designer did not specify a snubber for the PSU rather than lonely caps.
The major manufacturers are reliable. They make dozens of different duty devices to suit their customers. If Audio needs some oddballs to perform well, then there is probably something wrong with the Audio design, not with the component.
But don't let a capacitor (anywhere) become an inadvertant filter !
Filters will alter the wanted signal. Generally our amplifiers are required to PASS the signal, not to filter the signal. We filter the interference instead.
Last edited:
I think unfortunately the market is flooded with an overwhelming majority of products that appeal to boutique-connoisseurs. I am glad that there are some products around (DIY or otherwise) that appeal to the objectively measurable instead of dabbling in subjectivity. I suppose some psycho-acoustics are involved as well, as sometimes, a non-neutral sounding design can be made more palatable with coloration from some other boutique parts. I would think that starting with a top-notch design is far more crucial to the overall quality of a piece of electronic, rather than slapping on boutique and expensive band-aids to a hobbled design. Then again, it is hard to master good analog design, at least in my opinion.
The lone caps work just fine, actually. That's what you see in many commercial designs. It's also what Blencowe recommends.
If you insist on using a snubber, you'll have to tune it to match the parasitics in your circuit and transformer. There's no "one size fits all".
Now, why the lone cap didn't work for you is unknown. I'm guessing it wasn't rated for the ripple current or the voltage across it.
Tom
There was an article in Linear Audio a few years ago whose conclusion was that 1k Ohms in series with 1000 pF is an almost universal snubber for audio-related power supplies. There was another article (by the guy who did the 'alternative' power supply board for Modulus and chip amps that advocated more complex snubbers tuned to the individual application.
If one is trying to optimize the snubber, empirical measurements are almost a given since you almost never have an accurate circuit model for the transformer involved in addition to the rest of the supply. For recent projects I've just used the 1K + 1000pF figuring that it is better than nothing. Nothing, in fact, is the alternative due to limited time.
If one is trying to optimize the snubber, empirical measurements are almost a given since you almost never have an accurate circuit model for the transformer involved in addition to the rest of the supply. For recent projects I've just used the 1K + 1000pF figuring that it is better than nothing. Nothing, in fact, is the alternative due to limited time.
read Bateman's "capacitor Sound" series
his measurements prove your belief wrong
the thin spontaneous oxide on one foil of a Polar Al electro contributes measurably higher distortion than a same V, F non/bipolar with full thickness oxide grown on both foils
for low added (measurable) distortion when only Al electros fit, use bipolar, use as high a V rating as you can afford, even use 2 bipolars in series
reducing V/m across the dielectric seems to be a powerful, verified principle
his measurements prove your belief wrong
the thin spontaneous oxide on one foil of a Polar Al electro contributes measurably higher distortion than a same V, F non/bipolar with full thickness oxide grown on both foils
for low added (measurable) distortion when only Al electros fit, use bipolar, use as high a V rating as you can afford, even use 2 bipolars in series
reducing V/m across the dielectric seems to be a powerful, verified principle
coupling capacitors are always high pass filters
if you want to defend posting your incorrect belief about Al electrolytic distortion mechanisms by saying in most applications it doesn't matter, fine
you're still wrong on the fundamental facts and repeated that wrong understanding to the community
Bateman's work should have killed off the idea of series polar electrolytics of alternating polarity for AC anytime a real bipolar cap is available
if you want to defend posting your incorrect belief about Al electrolytic distortion mechanisms by saying in most applications it doesn't matter, fine
you're still wrong on the fundamental facts and repeated that wrong understanding to the community
Bateman's work should have killed off the idea of series polar electrolytics of alternating polarity for AC anytime a real bipolar cap is available
Last edited: