80hinhiding
Well-known member
- Joined
- Sep 27, 2016
- Messages
- 97
.
-
- Thread starter 80hinhiding
- Start date
Help Support GroupDIY Audio Forum:
The amount of capacitance (farads) probably matters more than the dielectric... But if you are serious make specific observations with more details about the specific capacitors and what you hear.
JR
JR
scott2000
Well-known member
Following this to get some more insight on how this works....
So, did you try it with just the 22uf cap and listened....
and then added the bypass cap and heard the change ?
So, did you try it with just the 22uf cap and listened....
and then added the bypass cap and heard the change ?
This is a can of worms. There are several old threads on how capacitors can affect audio signals on the forum, so i’d start there if you haven’t already.
Agree with John that capacitance value would be the first guess, so be sure you’re measuring the actual capacitance rather than relying on nominal value. Especially since you are hearing frequency response differences.
Assuming you’re already doing that, just a couple of thoughts. There’s an old blog post from an extinct blog that documented an interesting experiment on the linearity of different dielectrics by sampling the current across different caps and integrating the voltage to generate a lissajous curve (as with a B-H curve for looking at core permeability characteristics in transformers). The resulting curves purport to show the affects on linearity of an audio signal. The results, as the guy who did the experiments interpreted them, weren't out of line with common perceptions of different dielectrics, with PIO, polystyrene and polypropylene possessing the best linearity, and ceramics and tantalums possessing the worst. Properly DC-biased electrolytics weren't terrible, but weren't as linear as most common film types.
So FWIW, if you buy into this experiment then you'd expect that your electrolytics would've generated more distortion (of some sort) of the signal than the film types you tested.
PDF I saved from Wayback Machine attached.
Take it for what you will. I'm no audiofool and I can't hear a huge difference between PIO, styrene, polypropylene and polycarbonate types in my builds, if any. Polyester I'm on the fence about. But in my BA-283 clones I have tested electrolytics and tantalums (which was used in the originals) on the input, and there's a definite difference between them.
Cyril Bateman also did a whole series of experiments to determine the effects of different dielectrics on audio, and even documented a "DIY" (if you're severely obsessed or have infinite free time and patience) capacitor distortion tester. The series may be worth digging into if it's something you're really interested in:
https://linearaudio.nl/cyril-batemans-capacitor-sound-articles
BT
Agree with John that capacitance value would be the first guess, so be sure you’re measuring the actual capacitance rather than relying on nominal value. Especially since you are hearing frequency response differences.
Assuming you’re already doing that, just a couple of thoughts. There’s an old blog post from an extinct blog that documented an interesting experiment on the linearity of different dielectrics by sampling the current across different caps and integrating the voltage to generate a lissajous curve (as with a B-H curve for looking at core permeability characteristics in transformers). The resulting curves purport to show the affects on linearity of an audio signal. The results, as the guy who did the experiments interpreted them, weren't out of line with common perceptions of different dielectrics, with PIO, polystyrene and polypropylene possessing the best linearity, and ceramics and tantalums possessing the worst. Properly DC-biased electrolytics weren't terrible, but weren't as linear as most common film types.
So FWIW, if you buy into this experiment then you'd expect that your electrolytics would've generated more distortion (of some sort) of the signal than the film types you tested.
PDF I saved from Wayback Machine attached.
Take it for what you will. I'm no audiofool and I can't hear a huge difference between PIO, styrene, polypropylene and polycarbonate types in my builds, if any. Polyester I'm on the fence about. But in my BA-283 clones I have tested electrolytics and tantalums (which was used in the originals) on the input, and there's a definite difference between them.
Cyril Bateman also did a whole series of experiments to determine the effects of different dielectrics on audio, and even documented a "DIY" (if you're severely obsessed or have infinite free time and patience) capacitor distortion tester. The series may be worth digging into if it's something you're really interested in:
https://linearaudio.nl/cyril-batemans-capacitor-sound-articles
BT
ruffrecords
Well-known member
rackmonkey said:
There is more up to date info in Doug Self's book Small Signal Audio Design. It should be on every serious audio designer's bookshelf.
Cheers
Ian
caps can make a difference in old guitar amps,
as they get leaky, the amplifier stages get biased differently, which may be why Neil Young says all Fender tweed deluxe amps sound different, plus other factors like speaker damage and filter caps drying up ,
I hope Neil Young will remember... ;D
as they get leaky, the amplifier stages get biased differently, which may be why Neil Young says all Fender tweed deluxe amps sound different, plus other factors like speaker damage and filter caps drying up ,
I hope Neil Young will remember... ;D
scott2000
Well-known member
That does it. I'm recapping everything again.....
ruffrecords said:
Yes, I have that book. Self references the Bateman tests from 2002/03 as the definitive work on the subject, and his tests agree with Bateman’s. Looking over that chapter again, the interesting insight Self offers is that lowering the voltage across a cap lowers the distortion. So paralleling/bypassing caps in a circuit like Adam did will reduce the voltage across each and thus the distortion.
It’s worth noting that we’re splitting hairs here though. The measured distortion levels are in the 0.001-0.002% range.
Bateman also did an updated article in 2013 for Linear Audio (the Self book is from 2010), which I bought this morning for 0.99 euros. Main insight there is that cap construction has an effect. E-cased caps are better than stacked film in this regard, evidently. But again, this stuff is small, especially if you’re talking about tube circuits where 5% THD is not considered bad.
But both Self and Bateman agree that Polyester (PET) is the worst of the film types.
BT
CJ said:
As a Southern Man i’m not sure what i’m supposed to do with that.
But CJ raises the leakage issue, which is a big factor with electrolytics. My cheap little MK-328 and more expensive DER EE DE-5000 both measure leakage. You might take a look at that if you have the tools just to see how the electros are affected by that. Not much of a factor for film types though.
BT
I did my own research into this back in the 70s and yes there are differences between caps, but they also have sundry specifications that quantify these differences, no voodoo left to be discovered this century.
JR
JR
merlin said:
Merlin, welcome! Indeed, forgot to mention your excellent book’s coverage of that. Would be interested in any key points overlooked here. Seems like as good a thread as any to catch everything of interest on this topic once and for all.
Individual article here: https://linearaudio.net/article-detail/2222
Cheers,
Bryan
merlin
Well-known member
- Joined
- Aug 17, 2011
- Messages
- 260
Don't overlook that for polyester caps, distortion goes down as voltage rating goes up. A high voltage polyester is often almost as good as a polystyrene for the same size signals.rackmonkey said:
Also don't forget that a small polarizing voltage will reduce distortion in electrolytic caps.
There is lots of good technical information about differences between capacitors. In general the most audible flaws express themselves when there is changing terminal voltage, so capacitors in filter EQs where there is changing terminal voltage at audio frequencies, are more critical than simple DC blocking applications. Filters in general should be film and/or higher linearity dielectric.
JR
PS: Our own Sam Groner did some nice research into quantifying distortion in electrolytic capacitors. (on his website along with lots of other good information).
JR
PS: Our own Sam Groner did some nice research into quantifying distortion in electrolytic capacitors. (on his website along with lots of other good information).
true dat^
if you want a good test bed to test caps for sound, build a pultec eqp1a with clip leads for low and hi cap values, plug and play.
if you want a good test bed to test caps for sound, build a pultec eqp1a with clip leads for low and hi cap values, plug and play.
clintrubber
Well-known member
CJ said:
As other said, many articles around.
FWIW, let's add some more few:
Last year there was an interesting (but sure subjective) short series on caps for guitar amps, by
Udo Pipper in the German magazine Gitarre & Bass. Of interest when you want to recap an old guitar amp (and can read German).
Bye
buildafriend
Well-known member
Can you post a picture of the schematic you're referring to?
One thing that helped me to click with what a cap is doing to an audio circuit was reading the section on capacitors in the book "the art of electronics." The author states that when they are used for frequency filtering they are a short circuit to high frequencies. Now picture one going to ground from the signal and one in series with the signal. What would each one do? This is somewhat of an oversimplification though.
I can run some sweeps and get the actual response/noise differences on some common caps if you want to see the differences.
One thing that helped me to click with what a cap is doing to an audio circuit was reading the section on capacitors in the book "the art of electronics." The author states that when they are used for frequency filtering they are a short circuit to high frequencies. Now picture one going to ground from the signal and one in series with the signal. What would each one do? This is somewhat of an oversimplification though.
I can run some sweeps and get the actual response/noise differences on some common caps if you want to see the differences.
The circuit of the Quad II power amplifier uses foil based caps between driver and output stage ,the originals which have an aluminium sleve which is securely clamped to chassis apply a given capacity to ground from each grid which is essential to the correct opperation of the feedback network in that unit .I remember way back in the day when I put together my first, what I hoped would be hifi power amp ,I had these lovely Nitrogol paper in oil .1 uf caps , I securely bolted them to chassis without fully understanding the implications ,capacitance from output stage grids to ground was huge and things started to nose dive around 10khz , its worth remembering there are more variables than just dielectric material when it come to how caps sounds in a given situation .
