• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required as is 20 years of participation in forums (not all true). There are daily reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

Capacitor upgrade in crossover - You CAN'T handle the TRUTH! - Part 3

KaiserSoze

Addicted to Fun and Learning
Joined
Jun 8, 2020
Messages
699
Likes
588
Maybe this will help (WWW search)

Humble Homemade Hifi Capacitor Test
Among other things, this involves different mechanical resonances within capacitors (depending on the structure). You should read the embedded Clarity Cap white paper.

Cyril Bateman's Capacitor Sound articles
With the appropriate test equipment, it was possible to measure the differences between capacitors.


How does this help? I implore you to say what you think the claims or conclusions made in the paper are. I want to know what you think it says before I bother to read the paper and figure out what it says exactly. I've been down this road too many times before, where someone tells me that I need to read this or that paper, and then when I read the paper I discover that the conclusions are not especially clear.
 

R Swerdlow

Member
Joined
Jun 9, 2020
Messages
74
Likes
113
I've read the Humble Homemade HiFi Capacitor Test before. I think it's worse than worthless – it's misleading.
  • There was no explanation of how he performed his listening tests. I can only assume the author did all the listening, unblinded, and wrote all the evaluations himself.
  • The author never demonstrates whether he or anyone else can actually hear differences in a speaker's sound due to different capacitors in the crossover network. He went on to elaborately describe the differences he claims to have heard, but skipped the part where he established that he or anyone can indeed hear these differences. He deserves a few points for his creative evaluations and an extensive number of comparisons, but I docked him for those points, and more, because he never did scientific controls of any kind.
  • He never did negative controls, where two identical capacitors were compared while he was blind to their identity. Perhaps he assumed he could always tell when capacitors were identical. Why not try to measure the amount of 'false positives' that listeners report? This isn't about dishonesty; it's that no one is infallible.
  • He never did any positive controls. These are designed to measure what levels of noise or distortion he actually could reliably detect. Varyied levels of noise would have to be added on top of the sounds or music passage used during a listening test. These results could identify what was the lowest level of noise or distortion that he or other listeners could reliably hear. No such tests were done. The "humble" author apparently assumed he could always hear differences no matter how subtle they might be.
The Clarity Cap white paper makes a weak attempt at comparing capacitors by a listening test. But it fails to do any scientific controls, uses only 30 listeners, which are too few to allow statistically significant conclusions. It's data fails to establish that listeners can reliably hear differences in loud speakers due to different capacitors.

The Cyril Bateman capacitor articles aren't bad. He does show how he can measure very low capacitor distortion levels using electronic lab bench equipment. He never does any listening tests. In fact, he suggests that the levels of distortion he does detect are significantly lower than acoustically measured distortion levels in speakers. In none of his articles, is there any mention of speaker sound quality differences due to capacitor performances, much less any direct measurements of sound qualities. So, it is wrong to cite Cyril Bateman's electronic measurements as conformation that capacitors can differ, while ignoring his suggestion that the low distortion levels he measures are much lower than acoustic distortion levels in loud speakers.
 
Last edited:
OP
C

ctrl

Major Contributor
Joined
Jan 24, 2020
Messages
1,315
Likes
4,436
Location
.de, DE, DEU
I miss a complete list of the devices in the signal chain here used during your measurements!?
You have asked me exactly the same questions in a German forum, so you know the answers.
For everyone else: Audio interface MOTU UltraLite mk4, Marantz PM7001 as measuring amplifier.
More details about the test setup can be found in part one.


From my experience you have to look at the whole signal chain. E.g. an amplifier must be able to drive the audio signal precisely in the time axis so as not to blur it. If this does not happen in your case, this may result in the same capacitor plots for you.
Amplifier timing is a critical factor in such audible sound gains.
Every straw in the high-end-marketing-speech is grasped at :rolleyes:
In the German forum you made the same nebulous hints without presenting concrete facts.
What should enable a "good" amplifier to follow the test signal more precisely, so that a capacitor comparison is significantly influenced in the time axis?

Damping factor and amplifier negative feedback you can't mean, because a tweeter hardly generates any induction voltage, unlike a sub-woofer. In addition, 99% of passive loudspeakers have a voltage divider in the tweeter branch of the crossover, which ruins even the highest damping factor (because of the additional impedance).
Therefore my test setup is even more precise than the usual replacement of a capacitor in the high frequency branch of a crossover, because in my measurements only the capacitor and 4m speaker cable (OFC 2.5mm²) are in the signal path.

American competition law encourages manufacturers to use "comparative advertising" because it benefits the consumer.
However, I am not aware of a single case where a manufacturer has demonstrated how superior its audio capacitors are to the competition.

If high-end capacitors would improve the time response of a tweeter, there would certainly be countless measurements. Because the effort for this would be minimal.

If you lend me a high-end capacitor for measurement, I'll be happy to perform measurements with my Hypex Fusion-Amp.
 

KaiserSoze

Addicted to Fun and Learning
Joined
Jun 8, 2020
Messages
699
Likes
588
I've read the Humble Homemade HiFi Capacitor Test before. I think it's worse than worthless, it's misleading.
  • There was no explanation of how listening tests were performed. I can only assume the author did all the listening, unblinded, and wrote all the evaluations himself.
  • The author never demonstrates whether he or anyone else can actually hear differences in a speaker's sound due to different capacitors in the crossover network. He assumed that differences can be heard, and he went on to elaborately describe the differences he claims to have heard. He deserves a few points for his creative evaluations and numerous comparisons, but gets docked for more points because he never did scientific controls of any kind.
  • Did he ever do negative controls, where two identical capacitors were compared while he was blind to their identity? No. Perhaps he assumed he could always tell when capacitors were identical. Why not try to measure the amount of 'false positives' that listeners report. This isn't about dishonesty, no one is infallible.
  • Did he do any positive controls? These are designed to measure what levels of noise or distortion he actually could reliably detect. Varying levels of noise would have to be added on top of the sounds or music passage used during a listening test. These results could identify what was the lowest level of noise or distortion that he or other listeners could reliably hear. No such tests were done. The "humble" author assumed he could always hear differences no matter how subtle they might be.
The Clarity Cap white paper makes a weak attempt at a listening test comparing capacitors. But it fails to do any scientific controls, uses only 30 listeners, too few to allow statistically significant conclusions. It fails to make the point that listener's can reliably hear differences in loud speakers due to capacitors.

The Cyril Bateman capacitor articles aren't bad. He does show how to better measure capacitor distortion levels on an electronic lab bench. He never does any listening tests. In fact, he suggests that the levels of distortion he does detect are significantly lower than acoustically measured distortion levels in speakers. In none of these articles, is there any mention of speaker sound quality due to capacitor performances, much less any direct measurements of sound qualities.

Very good post!
 

KaiserSoze

Addicted to Fun and Learning
Joined
Jun 8, 2020
Messages
699
Likes
588
Among the various reason to be skeptical of the claims that different types of capacitors sound inherently different, there is this fact:

For the most part, the people who make this claim are the same people who claim that power cables make a difference, who claim that expensive speaker cables sound different from speaker cable made from regular 14-gauge or 16-gauge "lamp" wire, who claim that different tone arms sound different, and who claim that analog vinyl is superior to CD. These people are naturally predisposed to think that any and all differences lead to differences in sound quality, and they reason they think this way is simple: they do not know enough to have any intuitive sense of what differences are or are not likely to affect the sound. They believe what they read when they read that different types of capacitors sound different because they do not know enough to have a reason to doubt that this could be true.

For anyone who believes this notion about different types of capacitors, and who believes it with sincerity sufficient to convince the skeptical scientific community, there are basically three steps you would need to take, in a specific order, before it would make any good sense to expound a notion that you can't back up with evidence.

1. Conduct proper tests to establish that there are at least some people who can conclusively demonstrate the ability to hear the difference between two different types of capacitors, when both are chosen to be as closely matched in parameters as possible.

2. Conduct extensive experiments to isolate which secondary parameter it is that is responsible for the difference in sound.

3. Study the differences in design and manufacturing of the different types of capacitor to assess whether the design and manufacturing of the less expensive type of capacitor could be modified such that the less expensive capacitor can have the same values for the key secondary parameter that the more expensive type has.

Anyone who is truly serious about this would need to complete all three of these steps and then publish the findings in a single paper.
 

pogo

Addicted to Fun and Learning
Joined
Sep 4, 2020
Messages
687
Likes
165
CTRL,
Yes I know the amplifier you used, but I missed this information for further discussions in this forum.
If the NCore was implemented well in the Hypex Fusion-Amp I would prefer this amplifier for such measurements.
Background:
I've already heard the NCore and it has the ability to work out the details well. With the Marantz, as with all similar amplifiers of this type, which I was also able to hear some of them on my speakers, I suspect that it cannot reproduce the detail. I switched from a Pioneer A-717MKII to the NAD C388 and the NAD was able to better define voices / instruments and better present other sound events in addition to them. Like getting to know your recordings new.

I can also understand the repeated statement after a cap upgrade in a crossover that the sound change can result from slightly different capacitor values, but does not apply to my listening experience. Even with good equalizing, you won't get such a result.

Other questions that arise to me:
Which is the right measurement method to show the differences?
Examination of other types of tweeters (another possible bottleneck)?

Do not understand me wrong, but I am still interested in a meaningful measurement / analysis of this phenomenon and would have dealt with it more deeply myself. It's still on my ToDo list. First of all, I will increase my amplifier performance with the NAD M33 and I hope that the DIRAC room correction will perhaps be the cherry on top.
 

Octalman

Member
Joined
Mar 23, 2020
Messages
19
Likes
8
Kaiser, guess you don't remember writing this : "At present we have a fellow who willingly put himself in the spotlight because he wanted to increase his revenue. He is obviously not a scientist, as evidenced by his utter lack of understanding of scientific process. Scientists are rightfully annoyed when they encounter some bozo running around making preposterous claims all the while displaying a complete lack of understanding of what would be expected of him before any of his claims would be taken seriously by the scientific community."

Your won words hang you.
 

SIY

Master Contributor
Technical Expert
Joined
Apr 6, 2018
Messages
8,984
Likes
20,628
Location
Alfred, NY
How about this test? Pick a loudspeaker test subject that has tested well and received high marks during listening evaluation in the $1500 - $2000 retail range. Have Amir conduct a double blind listening test and document the listening results.

You're making claims, get off your butt and run experiments instead of trying to pass them off onto others. Especially one with a fantastically ignorant (or deliberately false and misleading) "procedure."
 

KaiserSoze

Addicted to Fun and Learning
Joined
Jun 8, 2020
Messages
699
Likes
588
Kaiser, guess you don't remember writing this : "At present we have a fellow who willingly put himself in the spotlight because he wanted to increase his revenue. He is obviously not a scientist, as evidenced by his utter lack of understanding of scientific process. Scientists are rightfully annoyed when they encounter some bozo running around making preposterous claims all the while displaying a complete lack of understanding of what would be expected of him before any of his claims would be taken seriously by the scientific community."

Your won words hang you.

This is funny as hell. For what possible reason would you have thought that I was referring to you? I have no idea who you are and do not care who you are. If by some chance you are the person I had in mind, I had no knowledge of this. It should have been apparent when you read this that I was referring to someone who profits from selling people expensive capacitors to replace or augment the perfectly good existing capacitors in their speakers. Let me be clear that I was not referring to you, and was not even aware of you specifically, however if the shoe fits, i.e., if you are someone who profits from selling people expensive capacitors to replace or augment the perfectly good existing capacitors in their speakers, then I approve of you taking my comment personally if this is what floats your boat.
 

pogo

Addicted to Fun and Learning
Joined
Sep 4, 2020
Messages
687
Likes
165
CTRL,
the main application of the EMX-7150 microphone is the measurement in buildings to optimize the room acoustics, studio acoustics or the home theater (see also product data sheet).

According to my research, for more precise laboratory measurements (including results in phase, time, impulse transmission, delay, etc.) of loudspeakers, it should be at least an Earthworks M23.
(German WWW Sources: Jochen Schulz Sound Engineering & Jobst-Audio)

Suitable preamp: e.g. an Earthworks 1021.
 
OP
C

ctrl

Major Contributor
Joined
Jan 24, 2020
Messages
1,315
Likes
4,436
Location
.de, DE, DEU
I've already heard the NCore and it has the ability to work out the details well. With the Marantz, as with all similar amplifiers of this type, which I was also able to hear some of them on my speakers, I suspect that it cannot reproduce the detail.
What is the technical background that leads you to this assumption?
As already mentioned the damping factor is not important for the tweeter.
I am not an expert in electronics, but the slew rate should not play a role due to the low voltage (around 1-2V). Since it is a "classic" amplifier (cut-off frequency -3dB at 100kHz), the same applies to the "gain-bandwidth product" because of the low voltage.
So what is the reason?

I have never read anywhere in the world that it has been said that with amplifier XY a capacitor upgrade will unfortunately be inaudible.
The opposite is true, a capacitor upgrade should supposedly immediately result in a dramatic sound improvement without any restrictions!

Really smart high-end capacitor seller and reviewer should update their websites and add the words "If your amplifier is able to follow the improved signal" everywhere ;)


Which is the right measurement method to show the differences?
For this reason I used the acoustic measurement. This corresponds one to one with reality. There is no "secret" vibration of the air molecules that has not been discovered so far.
The sound pressure resolution of the measuring microphone exceeds that of the human ear - this was shown in the first part, see 2) Capacitors - low capacitance deviations.

If someone lends me a high-end capacitor I'm willing to repeat a few measurements.
I could for example increase the resolution in the time domain extremely to see if there is any change (which I don't expect).



Examination of other types of tweeters (another possible bottleneck)?
Have you ever read somewhere that a capacitor upgrade was made dependent on the tweeter used? I have not.
A capacitor upgrade is said to always work, from expensive high-end speakers to cheap DIY speaker kits.
Expensive high-end capacitors supposedly lead to clearly perceptible sound improvements even in cheap tweeters where the voice coil floats in ferrofluid.

So a definite no, it's not because of the specific tweeter that I measured no difference in the audible frequency range between "normal" film capacitors and an expensive high-end capacitor.


the main application of the EMX-7150 microphone is the measurement in buildings to optimize the room acoustics, studio acoustics or the home theater (see also product data sheet).
OMG you are right, all those years wasted on useless measurements...
What does it say on the website:
1599730083810.png

There is really nothing about near field measurements of tweeters to analyze capacitors :eek:
Pardon the irony, but this is simply BS.


According to my research, for more precise laboratory measurements (including results in phase, time, impulse transmission, delay, etc.) of loudspeakers, it should be at least an Earthworks M23.
(German WWW Sources: Jochen Schulz Sound Engineering & Jobst-Audio)
Just two minutes of research and you get evidence that even the Earthworks M50, for example, has slightly higher distortion than the EMX-7150 - which is not unimportant for measurements with high sound levels, such as my capacitor measurements at 100-110dB. Both microphones deliver impeccable results.
1599731466856.png 1599731485321.png

Regarding noise the M50 with 23.1dB(A) is slightly superior to the EMX-7150 with 26.3dB(A) - but this is not important for near field measurements where the sound pressure level is around 100-110dB.
1599732084174.png 1599732096442.png

What are the technical, fact-based reasons for your statements and conclusions? More expensive is always better, is not one of them.
Why should it not be possible to measure sound changes caused by a capacitor upgrade with an EMX-7150, but it is with an Earthworks M23?
 

SIY

Master Contributor
Technical Expert
Joined
Apr 6, 2018
Messages
8,984
Likes
20,628
Location
Alfred, NY
OMG you are right, all those years wasted on useless measurements...
What does it say on the website:
View attachment 82315
There is really nothing about near field measurements of tweeters to analyze capacitors :eek:
Pardon the irony, but this is simply BS.

Just two minutes of research and you get evidence that even the Earthworks M50, for example, has slightly higher distortion than the EMX-7150 - which is not unimportant for measurements with high sound levels, such as my capacitor measurements at 100-110dB. Both microphones deliver impeccable results.
View attachment 82318 View attachment 82319

Regarding noise the M50 with 23.1dB(A) is slightly superior to the EMX-7150 with 26.3dB(A) - but this is not important for near field measurements where the sound pressure level is around 100-110dB.
View attachment 82324 View attachment 82325

What are the technical, fact-based reasons for your statements and conclusions? More expensive is always better, is not one of them.
Why should it not be possible to measure sound changes caused by a capacitor upgrade with an EMX-7150, but it is with an Earthworks M23?

If you use the Cal file that accompanies the EMX-7150, the response is as flat as a lab-grade measurement mike. Coincidentally, I just finished up some measurements on the Earthworks M23R, and one of the mikes used as a reference (besides the PCB Piezotronics mikes I usually use) was an EMX-7150+cal file sent to me by Earthworks.
 
OP
C

ctrl

Major Contributor
Joined
Jan 24, 2020
Messages
1,315
Likes
4,436
Location
.de, DE, DEU
If you use the Cal file that accompanies the EMX-7150, the response is as flat as a lab-grade measurement mike.
I have measured with cal file. In this case, since only relative sound pressure differences were measured, it does not matter.

If you are very small-minded, you can point out that measurements without a cal file would have been even more accurate in this case, because the cal file does not correct the phase frequency response (therefore the measurements would have been "more accurate" in a relative FR comparison without a cal file active).

The calibration file itself hardly uses any correction. Without the calibration file, the EMX-7150's accuracy is still < +-1dB (20-20000Hz).
Calibration file FR:
1599740103087.png



Coincidentally, I just finished up some measurements on the Earthworks M23R, and one of the mikes used as a reference (besides the PCB Piezotronics mikes I usually use) was an EMX-7150+cal file sent to me by Earthworks.

Can you give it away? Are the two microphones (M23, EMX-7150) on the same level in terms of distortion?
 
Last edited:

Octalman

Member
Joined
Mar 23, 2020
Messages
19
Likes
8
So Kaiser, calling an un-named person a Bozo is what branch of "science"? Glad you think it is so funny. For anyone here at ASR posting graphs, charts, measurements, and/or lecturing about equivalent circuit equations, provide the following information along with the proclamations of "proven", final answer, complete, exclamation point.

1) Brand, Model, and serial number of all equipment used in the experiment.
2) When were the items in 1) calibrated.
3) If calibrated, traceable to the Nation Bureau of Standards?
4) If not calibrated, explain why not.
5) What is the length, resistance, inductance, and capacitance of all cabling used.
6) Detailed diagram of the test configuration
7) Detailed test procedure
8) How is the unit under test (UUT) identified, stored, and handled
9) Complete description and details of the test location and environment
10) Procedure used to eliminate Bias of the tester.

If those 10 points are present, then a good start toward following a "scientific" process. Otherwise, simply an experiment - not scientific certainty.
 

Killingbeans

Major Contributor
Joined
Oct 23, 2018
Messages
2,939
Likes
5,070
Location
Bjerringbro, Denmark.

paddycrow

Senior Member
Forum Donor
Joined
Oct 28, 2019
Messages
340
Likes
535
Location
Grand Haven, MI
My $0.02- quality is not a precise engineering metric. Saying capacitor A is a higher quality than capacitor B doesn't tell me much. By what metric(s)?

Carry on...
 
Top Bottom