cap paper this is a good read

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.
Interesting.

So at +20dBu through 10 stages, distortion was 0.002% vs 0.0004% at 10kHz. Aside from the fact that I'm perfectly fine with 0.002% at 10kHz, what kind of crap circuit has 10 stages of electrolytic coupling? There are cheap op amps with low enough offsets that you could probably get away with 2-3 stages and presumably only the output is going to be vaguely high level and in practice no where near +20dBu.

And I'm not really not at all surprised that there's basically no difference between different brand / series of electrolytics when used in a conventional circuit. So instead of buying fancy "audio grade" electrolytics, go for duability / hours (assuming there's any real difference in durability / hours as well). I guess there are only so many ways to roll a piece of metalized plastic film and stuff it into an aluminum can.

I would be interested in seeing if there is any difference when using an electrolytic as a filter. Traditionally this has been avoided but I wonder if it's still justified. Using larger caps would allow for filters with lower characteristic impedance and thus lower noise. I care way more about noise than I do about triple-zero distortion.
 
I have used polypro 630VDC rated caps in a guitar amp power supply. You might need to add some resistance before the first cap to limit the surge current due to the low ESR.

This can upset a toroidal transformer with solid state diodes. I have not tried it with a tube rectifier.

 
Gus said:
I have not tried it with a tube rectifier.

Interesting, thanks for the observation about toroidal transformers Gus. I actually intend to try it with a valve rectifier with my next guitar amp build, employing a pi-filter after a 6X4 rectifier. I have these 10uF 400V poly caps. It is going to be a tiny wee push pull amp with CV4063 output valves. I have two pairs (of output tubes) just in case the first pair melts. I'm also going to use a 20uF PIO cathode bypass cap (for V1) in the same amp.

Nick
 
I have a brief look at the paper, there is only THD. I have tested lots of different schematics and components, some of them were remarkably matched (THD/IMD/FR) according to my measurements but had a remarkably different sound, so it is telling me, basically, that THD is the same, that's it :)
 
dbelousov said:
I have a brief look at the paper, there is only THD. I have tested lots of different schematics and components, some of them were remarkably matched (THD/IMD/FR) according to my measurements but had a remarkably different sound, so it is telling me, basically, that THD is the same, that's it :)
So did you make any complementary measurements that would show differences?
 
I haven't said this lately but back decades ago when I was spending too much time on the design bench, I found that I could measure things that I couldn't hear, and I could measure pretty much everything that I could reliably hear (back in the 70s I had to roll some of my own test equipment to do that). Nowadays bench test equipment is much better.

If you can't measure something that you can hear, you may not be measuring the right thing (a null test will confirm if there is anything there).

JR
 
dbelousov said:
I have a brief look at the paper, there is only THD. I have tested lots of different schematics and components, some of them were remarkably matched (THD/IMD/FR) according to my measurements but had a remarkably different sound, so it is telling me, basically, that THD is the same, that's it :)
If you can hear a difference but not measure one in THD, IMD or FR, then is it:
(a) Unconscious bias
(b) A transient effect not easily picked up with steady state measurements
(c) Something else?
 
Matt Nolan said:
If you can hear a difference but not measure one in THD, IMD or FR, then is it:
(a) Unconscious bias
(b) A transient effect not easily picked up with steady state measurements
(c) Something else? Poor design
Cheers

Ian
 
dbelousov said:
I have a brief look at the paper, there is only THD. I have tested lots of different schematics and components, some of them were remarkably matched (THD/IMD/FR) according to my measurements but had a remarkably different sound, so it is telling me, basically, that THD is the same, that's it :)
OK, I'll bite. Post the schematics and the changes you made.

Cheers

Ian
 
Matt Nolan said:
If you can hear a difference but not measure one in THD, IMD or FR, then is it:
(a) Unconscious bias
(b) A transient effect not easily picked up with steady state measurements
(c) Something else?

Something else can happen... but like I said a null test can confirm if there is a difference between two paths, or not.

JR
 
I'm done with proving that somebody is wrong or I'm right (it's not exactly the same thing) on the Internet  :) It takes considerable effort to do and doesn't gain anything for me, because the end is always the same. It's silence. So I just present what I know and if one doesn't believe me it is his right. But as a consequence, I have a right too. A right to not care about it :)

Regarding the subject, let the sh* hit the fan. Shunting lytics with PP works and the result is easily "nullable". I use direct-coupled ADC and direct-coupled DAC for tests. Maybe DC is not that important, but mine are DC, so I just present my condition. Audibly, the effect is more prominent with fast-transient stereo tracks. I use oversized Pana FR/FC lytics and Wima 0.033 uF MKP-2 caps.

 
dbelousov said:
Shunting lytics with PP works and the result is easily "nullable".
Running a null test to detect a difference at near triple-zero THD is not at all easy. You would have to trim the minima perfectly to see distortion that far down. I'm not sure if I could even do it. Certainly not with my usual test rig. Maybe with really good ADCs and some custom code to find the minima.
 
I think Ethan winer made a special test device for deep null testing.

He even made a video about that on his youtube channel.

Worth to be watched in my opinion.

Cheers,

Thomas
 
dbelousov said:
I'm done with proving that somebody is wrong or I'm right (it's not exactly the same thing) on the Internet  :) It takes considerable effort to do and doesn't gain anything for me, because the end is always the same. It's silence. So I just present what I know and if one doesn't believe me it is his right. But as a consequence, I have a right too. A right to not care about it :)

Regarding the subject, let the sh* hit the fan. Shunting lytics with PP works and the result is easily "nullable". I use direct-coupled ADC and direct-coupled DAC for tests. Maybe DC is not that important, but mine are DC, so I just present my condition. Audibly, the effect is more prominent with fast-transient stereo tracks. I use oversized Pana FR/FC lytics and Wima 0.033 uF MKP-2 caps.
I have written about this before too...

Back in the 70s while refining a RIAA preamp design I experienced issues with non-ideal electrolytic capacitors in the high gain leg of the RIAA stage. As was the fashion (popular wisdom) back then I experimented with paralleling small film caps across the electrolytic.  BUT I measured the results empirically... I focussed on phase shift related to the electrolytic capacitor's ESL (series inductance), because it was obvious and easily measurable (while not very audible).

In my bench measurements I had to use a film cap at least 10% of the total combined capacitance or larger to get film like behavior.  To fix my 22uF electrolytic would take a 2.2 uF film cap, much bigger and more expensive than I was willing to use. Over the years I just designed the electrolytic caps out of such circuits.

According to my bench tests your 0.033 uF cap could be used to parallel a 0.33 uF cap effectively.

JR

@ Thomas: Yes, Ethan is an old friend, I actually rented him a room in my house back in the 70s. My dog would chase his cat(s) so we had to keep them apart. 
 
dbelousov said:
I'm done with proving that somebody is wrong or I'm right (it's not exactly the same thing) on the Internet  :) It takes considerable effort to do and doesn't gain anything for me, because the end is always the same. It's silence. So I just present what I know and if one doesn't believe me it is his right. But as a consequence, I have a right too. A right to not care about it :)

Regarding the subject, let the sh* hit the fan. Shunting lytics with PP works and the result is easily "nullable". I use direct-coupled ADC and direct-coupled DAC for tests. Maybe DC is not that important, but mine are DC, so I just present my condition. Audibly, the effect is more prominent with fast-transient stereo tracks. I use oversized Pana FR/FC lytics and Wima 0.033 uF MKP-2 caps.
I am not interested in proving anyone right or wrong. I am interested in what people did to what circuits to get results. The trouble is, simply shunting existing caps and getting improved performance does not  necessarily demonstrate the failings of electrolytic capacitors, but more often the failings of the original design. That is why I want to see schematics. But if you are not up for sharing that is your choice.

Cheers

Ian
 
Back
Top