Burning in hi-pass capacitors

[Wolf]

Re: Burning in hi-pass capacitors

However, there is a significant and evidently very wealthy audio subculture that purports to be able to hear the differences between a Sonicap and a Solen and a Mundorf no matter how tight the tolerance. They too make the even bolder claim that they can tell with complete certainly that the Solen is harsh and metallic while the Sonicap is smooth and accurate.

I know your stance, j, but I disagree with that statement anyway. Solens are not metallic, but kinda artificially gritty, unless in the lowpass shunt where I like them. Personally thus far, Sonicaps have been hisssy or a little sssibilant. I would not use them with a bright/articulate tweeter. I've not invested capital in any Mundorf other than 0.1uF Supreme bypasses awhile ago. Don't know that they did anything at all.

Something I did recently was try a Northcreek Crescendo cap 1.0uF 630V in parallel with the Clarity SA on my Usher 9845 mid dome in my Attitudes just for fun. 5 minutes in they 'clicked'/formed, and they got very fizzily-crappy sounding. This was VERY audible. I tried other 1uF caps I also have of the 'uber quality type', and they did not have this problem. I did not think it would be that noticable, being 5% of the 20uF SA. Also adding 1uF to a 20uF should not excite the dome to be ill or nasty sounding.

I honestly thought it maybe was a damaged cap, since it was second-hand, but I enlisted the Crescendo on the bypass for the BeCu TB tweeter's pad resistor, and it was not as characterized before. It sounds lots better on these.

I'm more inclined to think that caps are application specific, and I never woulda thought that before, but maybe *some* just are.

I don't think I have golden ears, but I can identify things without measurements sometimes.

Later,
Wolf

[jcandy]

Re: Burning in hi-pass capacitors

Are the switches and extra connections going to completely f*** this up?

I would guess that this will not be a problem, especially in auditory tests, as the effective resistance, inductance and 1/capacitance of the switch is going to be exceedingly close to zero.

How are you going to do the test (who will listen and who will switch)?

[Wolf]

Re: Burning in hi-pass capacitors

you don't believe your own test.
Dan

Rhetorical or not, my test proved to me there is a difference.
Wasn't really funny, Dan, but nice try. :rolleyes:
Later,
Wolf

[jcandy]

Re: Burning in hi-pass capacitors

I know your stance, j, but I disagree with that statement anyway. Solens are not metallic, but kinda artificially gritty, unless in the lowpass shunt where I like them. Personally thus far, Sonicaps have been hisssy or a little sssibilant. I would not use them with a bright/articulate tweeter. I've not invested capital in any Mundorf other than 0.1uF Supreme bypasses awhile ago. Don't know that they did anything at all.

I respect your opinion and hope we will all have the patience and objectivity to keep digging deeper into this fascinating subject.

Most of the electrical testing of capacitors is done at the mA level, but listening is done at 1 or more amps. I wonder how much you think driving level has to do with differences in capacitors?

[bkeane1259]

Re: Burning in hi-pass capacitors

How are you going to do the test (who will listen and who will switch)?

I'm going to listen and switch at the same time :p

Seriously, right now I'm just thinking of involving three participants - myself, an old DJ partner buddy, and my 17 yr old son who is very much into this kind of stuff.

The switcher will be hidden from view behind the speaker setup and have access to the crossovers and switch box. The other two participants will listen together (backs to the speakers) and then we will rotate through a few times, each participant having a chance to be the engineer. I plan to keep good notes but I hope my illiteracy doesn't muck things up.

I'm not sold on this plan.......also not in a rush as I have stated before. Comments welcome.

[Wolf]

Re: Burning in hi-pass capacitors

I wonder how much you think driving level has to do with differences in capacitors?

I know that forming caps involves increasing the voltage supplied slowly to do it manually and not send them into overload. Theoretically, most of the caps we all use or test are 250VDC rated or higher. While that's not a current-spec, one could think about the possibilities during transients in music. Since they approach +30dB from reference level as measured by Bob Cordell, we can figure on a lot of power passing through. Suddenly, 400VDC caps don't sound like a bad idea....

Maybe I'm off-base on the above line of thinking, and please correct me if I am, but the math means that an 83dB speaker (assuming 8 ohm so the 1W/m figure will remain to be used) running on 1W, would require close to 1kW of headroom to keep from clipping. +3dB = twice power, ROF. Now; 1000 = I^2*8, so I = ~11A. To resub to find voltage, you get ~90VAC.

With the PSU guys usually factoring in a safe margin by doubling the voltage capacity, that would mean the 200VAC spec to become adequate.

Just spitballing....
Wolf

[Wolf]

Re: Burning in hi-pass capacitors

The other two participants will listen together (backs to the speakers)
I'm not sold on this plan.......also not in a rush as I have stated before. Comments welcome.

You won't be facing the speakers?? I don't think you'll find it as easy to listen cleanly if you are facing away from them.

Later,
Wolf

[bkeane1259]

Re: Burning in hi-pass capacitors

You won't be facing the speakers?? I don't think you'll find it as easy to listen cleanly if you are facing away from them.

Later,
Wolf

Just trying my damndest to separate listeners from the "switcher" but now that I think of it, backs to the sound is not a good idea.

Quick question: I imagine there's going to be an audible click when the switch is thrown, and I don't mean the actual mechanical sound of the switch but a "click" or interruption in the music that will indicate the switch has been thrown. Did you experience this in your tests?

[LSV8]

Re: Burning in hi-pass capacitors

Just trying my damndest to separate listeners from the "switcher" but now that I think of it, backs to the sound is not a good idea.

If the "switcher" is not also posing as a critical listener in these tests.... just put him behind a curtain of some sort. He can sit behind the speakers, with a curtain between the back of the speakers and himself. This way, you and your other listener can face the speakers, but not see the "switcher".

Quick question: I imagine there's going to be an audible click when the switch is thrown, and I don't mean the actual mechanical sound of the switch but a "click" or interruption in the music that will indicate the switch has been thrown. Did you experience this in your tests?

You could use a switch that "makes before breaks". So it's wipers/contacts are still connected to one pole while transferring to the other pole. The transfer happens during the switch motion... so the contact will be made on both poles for only a portion of the throw.

Example: Think of high/low beams on older vehicles with analog/solid state controlled healights. (I.E. not your newer MUX controlled lighting.) When you have you low beams on, and switch to highs.... there is a period between "clicks" that both highs AND lows will be on. Once you get past that "click" your lows turn off and the highs remain on.

The audible downside to this will be the doubling of capacitance during the throw. You wont hear a "click", or dropout in sound... but the crossover frequency will change during that brief transition... I suppose, IMO at least, that an XO position change for a split second would be better than a total dropout in sound or a click of some sort... just my train of thought at least.

EDIT: I suppose another way of using a "make before break" in this situation would be the style of DT switch you use. (Not sure if you're using a SPDT or a DPDT for this test... but either way I'm safe to assume it's still a DT switch.) Anyway, a DT switch can come in a few different ways... with a center position, or no center position. You could have one with no center: (ON-ON) or you could have two types of center position DTs: (ON-OFF-ON) OR (ON-ON-ON)... the latter of the two's center position would have BOTH poles "ON" in it's center position.

An example would be a SPDT on a 2 pickup guitar. It's up position would be the neck pickup only, it's center position would be neck AND bridge pickups, and it's down position would be bridge pickup only. And if it's a quality switch like on my Gibson's for example, the is no "dead" portion during the throw... AND the switch itself (mechanically speaking) is very quiet. In other words, there is no audible mechanical click from physically moving the toggle. (Or very little at least)

[bkeane1259]

Re: Burning in hi-pass capacitors

Crap - just got some bad news:

Sonic Craft will be closed April 18th. through the 29th. Any orders placed during this time will not be processed until May 2nd. Our apologies for any inconvenience.

To prepare for the brand vs. brand test, I ordered additional caps to get the values matching exactly. Sonic Craft is closed, so it's going to be awhile before I get the proper Sonicaps to parallel with the existing 8.2uF's to bring the measured values of all the cap participants to exactly 8.4uF.

I guess we'll skip them till later or just delay the test. Unless somebody has 4 Sonicap Gen I 0.1uF's they want to give me :D

[bkeane1259]

Re: Burning in hi-pass capacitors

You could use a switch that "makes before breaks". So it's wipers/contacts are still connected to one pole while transferring to the other pole. The transfer happens during the switch motion... so the contact will be made on both poles for only a portion of the throw.

Gotcha. I ordered the PE 4PDT switch, without knowing if it will satisfy my needs. I guess we'll just have to see, when it arrives.

I have to use this type of switch for the brand vs. brand cap test in order to flip between the caps with a stereo signal (but correct me if I'm wrong).

Since I do believe that crossovers help to improve things like imaging, sound stage, transparency, etc., we have to conduct the tests with a stereo array, and as such, I can't think of an alternative switching mechanism.

Alas, since I am noobish at this I certainly invite the electrical engineers to chime in with alternatives. Schematics welcome.

[Herman Trivilino]

Re: Burning in hi-pass capacitors

Yes, the engineer will exchange caps and flip switches and keep notes without any of the listeners able to see what he's doing.

That would make the test blind. To make it double blind this engineer (or whoever you get to do it) cannot know, when he's flipping switches or exchanging caps, which cap is which.

This extra layer of ignorance is necessary because experience has taught researchers that there are ways for your engineer to unconsiously communicate with the listeners. Attempts to correct for this have been shown to be problematic, too. Therefore, making the test double blind removes all doubt that a communication has occured.

Please understand that these measures are in place, not necessarily to counter cheating, but to counter a natural human tendency attributed to what we call the placebo effect.

The classic example of a double-blind test is a group of patients being treated for the same or similar malady. All test subjects are administered the medicine and are repeatedly examined by the physician to determine the progression of the symptons.

Half of the subjects receive the actual medicine, the other half a sugar pill, the so-called placebo. The patients don't know whether they are receiving the plecebo or the actual medicine. That makes the test blind. The treating physician also doesn't know which patients are getting the placebo and which are getting the real medicine. That makes the test double blind.

The reason for keeping the doctor ignorant is that he may communicate to the patients that they aren't, or are, getting the real medicine. This is not happening because they are cheating. It happens because they're human. The communication is nonverbal and entirely unintentional. It could be something like a raised eye brow for just a fraction of a second.

What's more, the patient may be looking for these nonverbal clues. He knows there's a 50-50 chance that he's getting the placebo. And he may interpret some gesture or comment made by the doctor to indicate that he's one of the lucky ones getting the real medicine, or that he's not.

Once a patient starts to believe he's getting the real medicine his symptoms may really really improve. This is a very powerful and well-document effect.

[Herman Trivilino]

Re: Burning in hi-pass capacitors

It's just not possible in this manner to do a test and get the results favoring one direction over the other; for a large majority of people.

What researchers do is show that there is, on average, a greater than 95% chance that the demonstrated preference is not due to chance alone.

(In statistics jargon this is known as setting the alpha level to 5% so that the confidence level is 95%).

I'm not claiming that this can or cannot be done in the case you're talking about. Perhaps if your sample was not random, but instead consisted of people who have claimed that they can tell the difference? For me, at least, that would be more meaningful..

[Herman Trivilino]

Re: Burning in hi-pass capacitors

Gotcha. I ordered the PE 4PDT switch, without knowing if it will satisfy my needs. I guess we'll just have to see, when it arrives.

For the first phase of your test (listening to a single speaker) you could use the type of single pole, double throw (SPDT) switch used on a guitar. I bought bought one at a music supply store. When placed in the center position, both devices are "on" as opposed to the typical SPDT switch that doesn't even have a center position.

[carlspeak]

Re: Burning in hi-pass capacitors

That would make the test blind. To make it double blind this engineer (or whoever you get to do it) cannot know, when he's flipping switches or exchanging caps, which cap is which.

This extra layer of ignorance is necessary because experience has taught researchers that there are ways for your engineer to unconsiously communicate with the listeners. Attempts to correct for this have been shown to be problematic, too. Therefore, making the test double blind removes all doubt that a communication has occured.

.

If both the tester (in this case the switcher) and the testee don't know what's being tested or when in this particular DBT scenario, how is a record kept of what was tested when?