Four Interesting experiments

[positron]

For over 50 years I have been performing experiments. Some were blase while others were very interesting and informative. I will attempt to keep the comments simple.

Four experiments have come to mind. The first is current while the second, third, and fourth are from the past.

1) In the first experiment, I have two 10 foot interconnects. One is an RCA and the other is my own DIY. Each IC is connected between the source (dragon fly) and preamplifier of 25k ohms input impedance (Z). The output Z of the DAC is

I am able to switch between the two ICs via the coined silver contact selector switch on the preamplifier. This eliminates the switches contacts as a possible problem. No sweat though.

The RCA IC measures 355pf average and my DIY measures 167pf average.
That is over a two to one ratio. Inductance is neglible when compared to the high preamplifier input Z. The higher capacitance of the RCA would normally cause the loss of highs. However, the low output Z of the source pretty much prevented such a loss.

Interestingly, the higher capacitance RCA IC had tighter bass and as such thinner midrange. The treble was the same, or very close. I was also able to thin the sound of my DIY ICs by simply changing materials in the IC terminations.

2) In the second experiment, I used a pair of speaker wires and altered the termination material/solder and needed to alter the xover to compensate. Clear difference in sound. Could the difference be altering the resistance by type of solder. Yes it could slightly, but maybe enough.

3) The third experiment was the comparison between all copper and silver coated copper ICs. I do not remember the capacitance of either cable. The silver coated ICs had a zing on the high end that the all coppers did not have. The high end exaggeration, however, did not follow through to the midrange. The FR was perceived as a non linear response, even though the capacitance had not changed.

4) I never really thought about ICs sounding different until I heard/read about perceived differences. So I conducted an experiment. I used two identical ICs, same capacitance/inductance, except one "A" was plugged (RCA plugs) between the source/Cd player and preamp "X" (25k input Z), while "B" ic was soldered between the source jack and preamplifier input "Y" jack. Thus I was able to switch (again silver contacts) between. Of course I could reverse the preamplifier inputs "X" and "Y". Contact resistance of the RCA plug was milliohms.

The lesson that I have learned and am reinforced with is that materials do matter in sonic quality. There are more that I do not recall, but I hope this will encourage you to experiment and update your system whenever you want.

Cheers
Pos

 

[JP]

For better understanding, what was the test methodology?

 

[SiliconTi]

Hmm... Interesting. I never gave a though to capacitance in regard to ICs.

 

[mhardy6647]

Re-experiment 1) Unless the only variable is capacitance (everything else about the two cables being the same) it is an uncontrolled experiment. It's anecdotal data ("I tried two different things, and they were different"), like the the famous correlation anecdotally observed by OB nurses between the full moon and the number of births in the maternity ward. The perceived correlation may (or may not) be valid, but as described the experiment does not provide sufficient evidence to even suggest causality.

It doesn't mean it didn't happen; it just means it is uninterpretable.

http://www.dancaton.physics.appstate.edu/Birthrates/Birthrates.htm

Hmm... Interesting. I never gave a though to capacitance in regard to ICs.

Capacitance is a well documented variable of considerable importance in certain circumstances (e.g., transfer of the tiny signal from a moving magnet cartridge to a preamplifier); capacitance should be as low as practically achievable to avoid HF rolloff. CD-4 (Compatible, discrete-4 channel) quad LP records used a multiplex strategy (like FM radio) to encode four channels of data in a record groove. CD-4 records require extremely low capacitance cables (and extremely extended treble response of the cartridge, which was the genesis of all of the exotic stylus geometries beyond the elliptical tip) to pass the ultrasonic pilot signal required to decode the multiplexed 4-channel data with adequate amplitude for the system to work.

 

[JP]

The attacks you levy against anyone who doesn't share your viewpoint are tiresome, as is the condescending manner in which they're delivered. This "public" did not elect you as the gatekeeper of what I'm allowed to see, or what I'm allowed to think.

In the hopes of having an actual conversation rather than continuing with the incessant right-fighting, would you mind addressing the variable confounds with some specificity? You mention this a lot, but I don't recall seeing any black-and-white details. You said your methodology above addressed more variable confounds than a DBT or ABX. Which, and how so?

 

[ConradH]

Capacitance may not be the only variable- did you measure dissipation factor and dielectric absorption? Fortunately I can't hear this stuff, but if you can, my measurements suggest that those things might be as or more important than just capacitance.

What was the outcome of of experiment #4?

I link stuff all the time, never had a complaint. Without additional info, everybody's talking in a vacuum.

 

[JP]

Mhardy may have been mistaken in regards to his comments of your methodology, or not, but I don't see anything in his post that is an attack, or that breaks any rules. I understand how you could feel the subject matter of his link is beneath you, but I can't speak to his motives. Obviously, I didn't see any outward maliciousness.

Aside from my feelings in regards to conduct, which I fully stand behind, I've never proffered a single technical opinion about this thread. I've only asked questions to try and better understand where you're coming from.

 

[mhardy6647]

Capacitance is the only variable as I previously stated, if you understood my posts. As such the results are interpretable. I also addressed more variable confounds than dbt/abx tests. The experiments are accurate, honest, and true.

Cheers
Pos

I'm missing something:

One is an RCA and the other is my own DIY.

EDIT: ah, wait, connector vs. no connector. Wouldn't one expect difference in resistance; metal to metal physical contact vs. solder? All metals the same (elements and alloys)?
Multiple variables to deconvolute.

 

[JP]

For the record, I'm a proponent of people doing their own testing as well.

To summarize, the specific variable confounds you stated you addressed are:

1. Limit AB switches to reduce habituation to stimuli
2. Test at different times of day to reduce environmentally conditioned perceptions
3. Limit SPL to prevent auditory fatigue.

I think that's all of them? The second is using my words to ensure I've grasped the breadth of what you said.

Putting aside any existing DBT, ABX, or what have you and discussing this as just a theoretical exercise, in the context of a person doing a DBT or ABX at home for themselves (ignoring the logistics of such):

For the first confound, I could easily take this in to account doing my own testing. My question is, what is the duration habituation is relevant? Let's say I listen for two hours. Whether I switch from A to B once, twice, or however many times in those two hours, my total exposure time with A and B are the same.

In other situations where I'm aware of auditory habituation occurring, I'm habituated for hours if not days, not minutes. For example, I'm extremely sensitive to sudden, loud, noises that I'm not expecting. My wife loves to slam kitchen cabinets, and travels a lot. When she's home and it occurs on a daily basis I cope, but after she's been gone a few days it's a different story. I'm not sure to what degree these are relevant, but this is how I'm thinking about it.

The second confound, I've no argument. I can't speak to how that shakes out statistically with a large pool, but for a single person at home, I can see the point.

Most of what I can find for cochlea fatigue relates directly to temporary hearing damage, and not the subtleness that I think you're talking about. Regardless, common sense says don't inflict damage upon your ears, so sure, 70dB. Why not?

It seems a person could take all this in to consideration and orchestrate a DBT or ABX that addresses these variable confounds. Taking this to the next logical step, due what we didn't address my DBT or ABX results may not be valid to anyone else, but for me, in my environment, with my system, I'll know exactly what I can actually hear or not.

Do I have that right?

 

[mhardy6647]

Let's check and see what we happens with measurements.

Between solders, let's suppose a very conservative tenth of an ohm variation vs 25k preamplifier input Z. As such we are talking one part in 250,000. That is down -108db or more down. At 50 milliohms, we are talking 1 part in 500,000 (-114db). It was down much less than that, so could not accurately measure into the low milliohms. So we are talking -120db, very very low.

(Solders are combinations of different materials. Some are tin/lead while eutectic solders are combinations of 3 and 4 materials. Resistance wise in a small connection, extremely low resistance, wires being soldered are physically touching or next to each other.)

Next, plugs are composed of different materials as well. I just compared two ICs, one with Vampire gold plated copper plugs, the other with Cardas Rhodium plated plugs. Again the difference between ICs is in the low milliohms, I could not really accurately measure that low.

Between jacks and no jacks, again in the low milliohms.

Now let's plug the ICs into jacks. Now we are measuring the cables, plug solder terminations, contact resistance of the plug/jack mate, and jack/wire solder termination. Conservatively less than 50 milliohms, again difficult to measure accurately. But let's conservatively say 50, so one part in 500,000 (-114db).

What does that leave us with? Well, between ICs to component connections the wire is exactly the same, the DA and DF are the same, the resistance change between solder vs solder, solder vs plugs, plugs vs plugs; the max difference is Ics and plugs/jacks vs ICs and plug/jack are at or less than 1 in 500,000, or -114db.

The only other considerations are different materials used, just mating different material plug to jacks together, rectification effects from mating different materials together.

Cheers
Pos

Sounds good -- sorry, none of that was clear to me from the first post.

 

[ConradH]

Let's check and see what we happens with measurements.

Between solders, let's suppose a very conservative tenth of an ohm variation vs 25k preamplifier input Z. As such we are talking one part in 250,000. That is down -108db or more down. At 50 milliohms, we are talking 1 part in 500,000 (-114db). It was down much less than that, so could not accurately measure into the low milliohms. So we are talking -120db, very very low.

(Solders are combinations of different materials. Some are tin/lead while eutectic solders are combinations of 3 and 4 materials. Resistance wise in a small connection, extremely low resistance, wires being soldered are physically touching or next to each other.)

Next, plugs are composed of different materials as well. I just compared two ICs, one with Vampire gold plated copper plugs, the other with Cardas Rhodium plated plugs. Again the difference between ICs is in the low milliohms, I could not really accurately measure that low.

Between jacks and no jacks, again in the low milliohms.

Now let's plug the ICs into jacks. Now we are measuring the cables, plug solder terminations, contact resistance of the plug/jack mate, and jack/wire solder termination. Conservatively less than 50 milliohms, again difficult to measure accurately. But let's conservatively say 50, so one part in 500,000 (-114db).

What does that leave us with? Well, between ICs to component connections the wire is exactly the same, the DA and DF are the same, the resistance change between solder vs solder, solder vs plugs, plugs vs plugs; the max difference is Ics and plugs/jacks vs ICs and plug/jack are at or less than 1 in 500,000, or -114db.

The only other considerations are different materials used, just mating different material plug to jacks together, rectification effects from mating different materials together.

Cheers
Pos

What you have there is a fairly powerful argument for one of two things. Either some factor has been missed, which IMO is doubtful, or the listening test method has a flaw, IMO, much more probable. Every time I do this sort of thing, and am convinced beyond all reasonable doubt that I hear a difference, it turns out to be a problem with my listening test, not the measurement methods. I always leave the door open that others can hear things I can't (almost certain) but the more rigorous the test, the less evidence I can find that conductor materials and such make any difference. FWIW, I can measure differences between almost anything, including conductor materials and geometries with no problem. I can still hear differences but have a better idea of the "why" of it.

 

[JP]

2) Yes. I didn't say that, but I meant to include it.

In regards to pre-damaged perception, auditory damage, etc. is why I closed with stating the tests may not be relevant to anyone else, or any other environment or system. As a hobbyist, that doesn't bother me as I'm really only concerned with determining what I can actually hear.

Personally, I find testing can be difficult to orchestrate as a lone person as there are so many clues. I can build a competent means to switch between cables, but mechanical clicks, tactile feel, etc. can all be picked up and associated, even subconsciously. The old wife in the kitchen bit. My experience is similar to Conrad's, in that differences tend to disappear as the protocol is tightened up.

I've been pondering how to do some ABX testing of turntable systems. As no two playbacks are identical, if a preference is determined with no obvious reason, how can one determine if the preference was real, or due to audible clues picked up subconciously.

 

[JP]

I think test phobia in the case of testing myself in my space is minimized. I won't say it's not applicable as no one is immune to cognitive bias, but it's as minimized as it's going to get. By tightening up, I mean eliminating variables and eliminating things that can give clues, e.g. that mechanical click, etc.

The wife couldn't care less about hi-fi. I'd be dubious of the results as I know with me, mood affects how I hear to a large degree, and often enough I don't realize I'm not in the right mood until I notice things sound off.

 

[ConradH]

2) ... I've been pondering how to do some ABX testing of turntable systems. As no two playbacks are identical, if a preference is determined with no obvious reason, how can one determine if the preference was real, or due to audible clues picked up subconciously.

Turntables are an incredibly interesting line of investigation, but the difficulties of doing even a moderately controlled test make my head hurt! If you do get a result, connecting it to specific mechanical (or electrical) differences is tough. OTOH, it would put to rest a lot of things people worry about, and probably raise 27 others.

On eddy currents and material combos, those things should show up using lab tests. They aren't mysteries that only arise with music. Based on their absence in spectrum plots and similar measurements, diode effects and noise generation don't seem to be a problem unless something is actually defective, and they rise up above that -??? dB floor. OTOH, my contrary observation is that all systems can be improved 13.8 % by unplugging all the cables and plugging them in again.

 

[JP]

I want to compare an SL-1000MK3 and SL-1000MKII, first on their own and then swapping parts around - EPA-100MK2 arm on the SP-10MKII, EPA-100 on the SP-10MK3, EPA-500, etc. I'll share what I'm doing as well as the recordings for comparison, but I'm doing it to satisfy my curiosity. Trouble is, I know if I can't get rid of the audible cues from vinyl, the entire exercise will be a waste of time aside from any gross differences.

 

[ConradH]

I'd need a lot more evidence to convince me. Of course it's difficult to resolve .0032 dB, but I don't for a second think the ear can do it either. Test equipment-wise, it's easy with a good ratio transformer, in fact some are accurate in parts per billion. Nothing new, that was available back in 1968! As for the cello, something else must have been changing because 32 udB of response change is probably less change then you get from thermal coefficients every time you power up an amp or the room changes by a few degrees. Eddy currents are in fact interesting, but they aren't terribly mysterious, and just show up as frequency dependent (I think) losses. I don't see how they matter in terms of signal transmission. Again, if they had any real effect, it would be easily (OK, maybe not easily, but doable with common equipment) measurable.

 

[ConradH]

Sorry, I'm obviously in the wrong spot. Sometimes I stumble in here by accident when looking at new posts.

 

[Brindabella]

You're a weird (in a good way) and wonderful guy. I really like your posts and threads they really make my brain hurt. Also you have the patience of a saint when faced with fierce opposition. I hope you never change your wicked ways.

Rob

 

[2ndtimelucky]

Well, that's me out. Very interesting read but I understood precisely 0% and somehow my head still aches

Keep up the good work

 

[Brindabella]

That was a lot clearer and confirms to me why I hear more, feel more and actually more observant when I am listening for pleasure. Critical listening I tend to be so focused on trying to hear nuance it actually becomes confusing and find myself switching back and forth from whatever I have introduced that is different/new to duplicate the experiment over and over checking the results ultimately getting less clear in the process.

However, while the music is just on casually in the house I can usually within a few hours of introducing change hear clearly whatever it is that I have done has either improved, worsen the sound, not changed or has changed the sound that is neither better or worse - just different.

Thanks - it's all my stupid brains fault!

Rob