Burning in hi-pass capacitors

[Herman Trivilino]

Re: CAPS - FR graphs - pink noise and sweeps

From a purely objective perspective, there appears to be no significant difference between the cooked and non-cooked caps for any of the manufacturers. I really doubt that the minor fluctuations in the plots can be detected by the human ear in any sort of music program.

The only way to determine if you're right about this is to conduct listening tests.

I will still get around to the subjective tests. Unfortunately, that is a much tougher proposition.

Oh, yeah. It sure is.

[bkeane1259]

CAPS - FR graphs - pink noise and sweeps

I will eventually get around to corralling some guinea pigs for the cap listening tests, but I figured I could run some FR graphs in the meantime. I measured each cap with pink noise and a quick sweep. One cap is cooked for just about 200 hours and the other is not.

Mic placement: Basically I stuck the mic about 1 inch from the tweeter.



I ran 1 minute of pink noise with 500 averages and plotted the non-smoothed results for each pair of caps, and then separated the pairs by 10db.

In this first graph the bottom pair are the Clarity ESA caps, the middle pair are the Sonicap G1's and the top pair are the Dayton 5%.



For the second test I did a quick spectrum sweep and then separated the pairs by 10db.

Same parings as before-- The bottom pair are the Clarity ESA caps, the middle pair are the Sonicap G1's and the top pair are the Dayton 5%.



From a purely objective perspective, there appears to be no significant difference between the cooked and non-cooked caps for any of the manufacturers. I really doubt that the minor fluctuations in the plots can be detected by the human ear in any sort of music program.

Now if anyone is a TrueRTA guru and would like to point out a better methodology for this test, please do so, and I will remeasure.

I will still get around to the subjective tests. Unfortunately, that is a much tougher proposition.

[bkeane1259]

Re: Burning in hi-pass capacitors

Just wondering if you got around to running the tests and if you have any results to show us.

Dave.

My initial attempt was a flop (I posted about it). I revised the testing methodology per the advice from a few helpful posters and decided to try straight up pink noise -- not well received by my sons. They were the only second-round testers. I have not tried since.

I'm still going to. I just recently bought measurement equipment and software, so I'm attempting a self-administered crash course in FR measurement -- just not enough time in the day.

I plan to try again with both listeners and measuring equipment. I've certainly NOT ditched the idea. I still have the testing caps (cooked and new). I still don't know which are which. They are presently collecting a bit of dust while I get my rear into gear.

[oldnwrinkly]

Re: Burning in hi-pass capacitors

Just wondering if you got around to running the tests and if you have any results to show us.

Dave.

[bkeane1259]

Re: Burning in hi-pass capacitors

Thanks for all the tips, guys!! :D

I really want it to be done correctly.


We'll revisit the testing next weekend.

[dantheman]

Re: Burning in hi-pass capacitors

Herman, I wasn't saying the data shouldn't be interpreted, just that the test shouldn't be designed to try and prove a position. IOW, we shouldn't care what the results are, just how we got to them. That's like intro to engineering.

Level matching is utterly essential for sure Carlspeak. Thanks for bringing it up.

Dan

[carlspeak]

Re: Burning in hi-pass capacitors

Well, it can in the mind of an individual constitute proof or disproof. For example, if I were bkeane and were unable to the difference between the NPE caps and the film caps with my own self-designed test, it would constitute proof that I was unable to hear a difference that I was previously convinced I was able to hear.

Conversely, if I were able to verify my ability to hear the difference, that might cause me to conclude I'd proven my point. However, speaking for myself, I would need more evidence before I was convinced that films sounded better than NPE's when used in series with a tweeter. All I'd have really proved is that I can hear a difference in one specific case. Catch me on another day, listening to a different kind of music with different speakers in a different room, I might instead conclude that the NPE's sound "better". Or, I might not be able to verify that I can hear a difference.

I know what you mean, in the end it's just data. Nevertheless, interpretation of data can and has led to significant advancements in civilizations. The philosophers can go on and on about how the data proves nothing, meanwhile engineers and scientists are using it to develop new and improved technologies that make my existence a lot better than it otherwise would have been.

Bkeane can't loose site of the fact that different ESR's (NPE vs film is an example) will most likely result in different SPL's out of the speaker with only just the simple switching going on, resulting in film types being favored or at least easily differentiated over the NPE's because the NPE's play a bit softer. Volume level matching is CRITICAL in any BT or DBT cap comparison test.
This SPL difference may be what Bkeane heard with his quick test before the more rigorous film vs film tests were run.
The Clarity Cap's study is an example with its initial, large un-trained panel of listeners able to differentiate between NPE's and film types. However, when it came to comparing film to film, CC had to resort to a trained golden ear'd panel and a special test protocol.

Match those volume levels and listen for other audible attributes besides louder vs softer.

[Herman Trivilino]

Re: Burning in hi-pass capacitors

That's excellent Herman. I just want to add: remember, it's data collection, nothing more or less. It's not an attempt to prove a point or anything at all. So you need a test that will accurately assess capacitor audibility. Expectation and sight bias need to go out the window.

Well, it can in the mind of an individual constitute proof or disproof. For example, if I were bkeane and were unable to the difference between the NPE caps and the film caps with my own self-designed test, it would constitute proof that I was unable to hear a difference that I was previously convinced I was able to hear.

Conversely, if I were able to verify my ability to hear the difference, that might cause me to conclude I'd proven my point. However, speaking for myself, I would need more evidence before I was convinced that films sounded better than NPE's when used in series with a tweeter. All I'd have really proved is that I can hear a difference in one specific case. Catch me on another day, listening to a different kind of music with different speakers in a different room, I might instead conclude that the NPE's sound "better". Or, I might not be able to verify that I can hear a difference.

I know what you mean, in the end it's just data. Nevertheless, interpretation of data can and has led to significant advancements in civilizations. The philosophers can go on and on about how the data proves nothing, meanwhile engineers and scientists are using it to develop new and improved technologies that make my existence a lot better than it otherwise would have been.

[dantheman]

Re: Burning in hi-pass capacitors

Your test subjects need to listen with pen and paper in hand. Let them hear what A and B sound like, as many times as they want. If they claim they can tell a difference, the game is on! Tell them that you're now going to run a test to see if they really can tell a difference.

Announce that this is Run 1. Flip a coin and play A if it's heads, B if it's tails, but don't let them see the result of the coin toss. Ask them to write down either A or B on their paper next to Run 1. (Again, they should have the opportunity to listen to both A and B as many times as it takes them to decide, but they can't ask for them by name. They have to say something like "Let me hear that first one again".)

Then flip the coin again, announce that this is Run 2, and repeat. And of course, neither can the test subjects.

It'll be tedious, so you may have to bribe your test subjects to get them to continue.

And remember, the person flipping the switch can't know which cap (A or B) is the burned-in cap, or the NPE cap, or whatever.

That's excellent Herman. I just want to add: remember, it's data collection, nothing more or less. It's not an attempt to prove a point or anything at all. So you need a test that will accurately assess capacitor audibility. Expectation and sight bias need to go out the window.

Dan

[carlspeak]

Re: Burning in hi-pass capacitors

.............
What in the world is there to "break-in" on a cable? How can you possibly distinguish performance "before" and "after" a break-in, over a period of weeks?

You can't, because psychoacoustics is a marvelous thing.....one aspect of it is the brain's ability to adjust to sounds over the long term, thus giving the allusion of 'break-in' when, in fact, the brain is 'breaking -in'.

[jimofoakcreek]

Re: Burning in hi-pass capacitors

+1. No moving parts=no burn in.

IMO any cap manufacturer that recommends burn in is a purveyor of snake oil and should not be rewarded with your business.

My limited knowledge and my strong intuition along with my ears more or less agree with you.

However, I will say this. I have found tubes to grow more quiet after a few hours of "break-in". I also get an impression of subtle sonic changes as well, perhaps just my imagination.

What I find incredulous is the "Cable Asyum" forum at another Web site. There, the audiophiles have no qualms about spending $500 - $1K+ for cables. Then they post rave reviews after a 100+ hours of cable "break-in".

What in the world is there to "break-in" on a cable? How can you possibly distinguish performance "before" and "after" a break-in, over a period of weeks?

[Herman Trivilino]

Re: Burning in hi-pass capacitors

I need to rethink how to conduct this test. I'm open to suggestions.

Your test subjects need to listen with pen and paper in hand. Let them hear what A and B sound like, as many times as they want. If they claim they can tell a difference, the game is on! Tell them that you're now going to run a test to see if they really can tell a difference.

Announce that this is Run 1. Flip a coin and play A if it's heads, B if it's tails, but don't let them see the result of the coin toss. Ask them to write down either A or B on their paper next to Run 1. (Again, they should have the opportunity to listen to both A and B as many times as it takes them to decide, but they can't ask for them by name. They have to say something like "Let me hear that first one again".)

Then flip the coin again, announce that this is Run 2, and repeat. And of course, neither can the test subjects.

It'll be tedious, so you may have to bribe your test subjects to get them to continue.

And remember, the person flipping the switch can't know which cap (A or B) is the burned-in cap, or the NPE cap, or whatever.

[Herman Trivilino]

Re: Burning in hi-pass capacitors

I paralleled my Jantzens with two 0.22uF and one 0.1uF Sonicaps, getting me to 8.94uF on the Jantzen side. The other speaker was running the cheapo NPE + mylar bypass -- the measurement there was 9.0uF.

There was absolutely nothing scientific about this test....just for fun basically, yet revealing.

This is a perfect comparison, then, for your test. If the difference really is obvious, your tests will confirm it.

[carlspeak]

Re: Burning in hi-pass capacitors

I suggest you seriously consider using a simple pink noise track as one of the listener tracks. Some years back in the 'old' days when Stereophile was still located in Santa Fe, NM, they ran a series of speaker evaluation panel tests using their own writers. I still have the original articles. There were numerous music tracks and a pink noise track. The one track that had the most consistent scores good or bad among all of the panelists was the pink noise track. All the others, had much more variability and thus confusion ensued about which speaker rated best.
Listening to pink noise requires some patience. The sound should be well balanced and a gentle hiss. Any distortion or peaks in the response are more easily detectable.

[bkeane1259]

Re: Burning in hi-pass capacitors

Tried the "can you tell the difference between a cooked cap and a new cap" test tonight with a few friends and relatives. It was a complete mess.

Most participants noticed a difference, but had a hard time articulating exactly what the difference was. I had 8 listeners ranging from young teens to 40-somethings. I was prepared to go several rounds, but I ended up changing my original source music (1 a-Capella, 1 normal song, 1 drum track), to songs that were more widely known by ALL participants.

I need to rethink how to conduct this test. I'm open to suggestions.

Here's what tonight's setup looked like. I was the engineer.....I'm not participating in the listening.