heres an odd one

[daz]

I sometimes post just to see what other think when i run into something that i don't understand. So i'm not in need of help with this, i'm just curious. I built a new turret board for my amp because the old one was shot from endless experimentation. I used the same brand and type of board and components except 3 things....wire, PSU caps for the preamp nodes, (main can is te same) and the three .68uf electrolytics for the 3 gain stages cathode bypass'. (yes, i know, but it's still not too bright)

So the amp sounded similar but not quite the same. It seemed darker and less complex and bright. I figured it must be the fact i used three F&T caps for the pre nodes instead of the JJ multi cap can. So i replaced the cap can and it seemed closer but not the same still. Then i replaced the 3 cathode bypass caps on the gain stages and the tone was back ! Both the old and new caps were .68uf polarized electrolytics, but the old ones were a different brand and 100 volts while the new ones were 63 volts and oddly bigger by about 30%. (I would think being smaller voltage with the same capacitance they be smaller)

Just curious.....has anyone experienced this kind of thing? Do you think it was the brand or the voltage or what?

[soundguruman]

I sometimes post just to see what other think when i run into something that i don't understand. So i'm not in need of help with this, i'm just curious. I built a new turret board for my amp because the old one was shot from endless experimentation. I used the same brand and type of board and components except 3 things....wire, PSU caps for the preamp nodes, (main can is te same) and the three .68uf electrolytics for the 3 gain stages cathode bypass'. (yes, i know, but it's still not too bright)

So the amp sounded similar but not quite the same. It seemed darker and less complex and bright. I figured it must be the fact i used three F&T caps for the pre nodes instead of the JJ multi cap can. So i replaced the cap can and it seemed closer but not the same still. Then i replaced the 3 cathode bypass caps on the gain stages and the tone was back ! Both the old and new caps were .68uf polarized electrolytics, but the old ones were a different brand and 100 volts while the new ones were 63 volts and oddly bigger by about 30%. (I would think being smaller voltage with the same capacitance they be smaller)

Just curious.....has anyone experienced this kind of thing? Do you think it was the brand or the voltage or what?

We ignore the Q of the capacitor in these circuits.
But the Q affects the sound of the circuit radically.
There are many specifications of capacitors, that we ignore 100% of the time. And the manufacturers do not tell us anything, normally.
A capacitor analyzer, which measures Q directly (and many other perimeters), is about $4000-$6000. No Bull.

We can make a tone stack, but with typical cheap capacitors, the controls may not work at all, or 'barely.'
With High Q capacitors, the tone stack works killer...but HOW do I sort the crap from the good ones?

The quality factor (or Q) of a capacitor is the ratio of its reactance to its resistance at a given frequency, and is a measure of its efficiency. The higher the Q factor of the capacitor, the closer it approaches the behavior of an ideal, lossless, capacitor.
The Q factor of a capacitor can be found through the following formula:
Where:

• is frequency in radians per second,
• is the capacitance,
• is the capacitive reactance, and
• is the series resistance of the capacitor.

[daz]

We ignore the Q of the capacitor in these circuits.
But the Q affects the sound of the circuit radically.
There are many specifications of capacitors, that we ignore 100% of the time. And the manufacturers do not tell us anything, normally.
A capacitor analyzer, which measures Q directly (and many other perimeters), is about $4000-$6000. No Bull.

We can make a tone stack, but with typical cheap capacitors, the controls may not work at all, or 'barely.'
With High Q capacitors, the tone stack works killer...but HOW do I sort the crap from the good ones?

Interesting, thanks. So if i understand this, while both .68's roll off at the same frequency i'm sure, one may not have a flat curve from that point of within the frequencies that pass and in fact may have a peak somewhere? Is that it ?

[soundguruman]

YES.
I can take a box of 12 identical caps, and analyze them...
And 1 of them might have a Q of 7...
and three of them might have a Q of 3...
So, we should be "pre-screening" caps, just like we select preamp tubes!

[daz]

YES.
I can take a box of 12 identical caps, and analyze them...
And 1 of them might have a Q of 7...
and three of them might have a Q of 3...
So, we should be "pre-screening" caps, just like we select preamp tubes!

This is the first time i ever noticed this to such a degree, but it does make a case for trying many different ones to see if something stands out.

[R.G.]

There's a cheaper way to get to Q.

Actually, what you're looking for is the ESR ("Equivalent Series Resistance"). The actual Q is of some interest, but there are cap meters that can give you a quick, if not super accurate, capacitance value for a low price. The good news is that there are a few ESR meter kits available for well under US$100, some under $100.

I built one for myself from a hacked up circuit diagram, and liked it enough to spend money for one of the kits.

ESR testing is a GREAT way to test electrolytic caps. Old electros have high ESR, and you can often decide which cap needs replaced by a very quick and dirty ESR test.

If you want Q, then the equations hold: calculate the impedance of the cap, divide by the ESR, and you have Q.

... er, at the test frequency or the calculation frequency. Notice that Q is dependent on frequency. ESR is not, nor is capacitance.

Well, usually it's not. There's that thing where all capacitors become inductors if you look at high enough frequencies, but that's something that tube amps don't usually get into.

[trobbins]

I don't think Q is the relevant term to use here as it refers more to the first self-resonant frequency of the cap, which is highly likely to be way out of the frequency spectrum.

ESR may be a concern, but I think when you get actual meassurements of ESR and then look at the equivalent circuit corner frequency, it is again not likely to be noticeable per se.

Where I believe some people have noticed, and made valid measurements (amongst other snake-oil and hype that has permeated some discussions), is the time response of the capacitance and ESR. The electrochemistry of the capacitor appears to impact the dynamic impedance of the cap as signal changes - sort of akin to a battery having a dynamic impedance.

[soundguruman]

There's a cheaper way to get to Q.

Actually, what you're looking for is the ESR ("Equivalent Series Resistance"). The actual Q is of some interest, but there are cap meters that can give you a quick, if not super accurate, capacitance value for a low price. The good news is that there are a few ESR meter kits available for well under US$100, some under $100.

I built one for myself from a hacked up circuit diagram, and liked it enough to spend money for one of the kits.

ESR testing is a GREAT way to test electrolytic caps. Old electros have high ESR, and you can often decide which cap needs replaced by a very quick and dirty ESR test.

If you want Q, then the equations hold: calculate the impedance of the cap, divide by the ESR, and you have Q.

... er, at the test frequency or the calculation frequency. Notice that Q is dependent on frequency. ESR is not, nor is capacitance.

Well, usually it's not. There's that thing where all capacitors become inductors if you look at high enough frequencies, but that's something that tube amps don't usually get into.

The 100$ one does not have a digital mojo indicator, and variable test frequencies...I gotta have that.
The cheap one tests the component at a single frequency.
The expensive one sweep tests it at pre programmed frequencies...and leakage, ESR, Q, and insulation, and a bunch of other stuff and it is tested at a programmed high voltage, instead of just 9 volts. Then it has impedance measurement, Milli ohms, L measurement, etc, etc...and it measures reluctance, and all that cool stuff. USB interface and software, Very large and expensive machine, that goes: "bing" Accurate to within .077 parsecs...calibrated by Voodoo priest.

[nickb]

The primary parameter here is capacitance. ESR is tiny compared to the cathode resistance. The tolerance is 20% for modern electrolytics, that's darn close to half an octave from max to min. It's one of the big amp secrets: tolerances matter.

[R.G.]

The 100$ one does not have a digital mojo indicator, and variable test frequencies...I gotta have that.

Whatever floats your boat. If you want mojo, go for it. It is, after all, only money.

Well, actually the $100 one has a digital readout, but the mojo setting is sadly lacking.

I'm addicted to the 80-20 rule - 80% of the good is gotten with 20% of the effort/money/yada/yada. Actually, with uC based instruments, it's more like 90% of the good is gotten with 1% of the money. The extra decimal points of precision get expensive all out of proportion to their usefulness. But you do get pretty lights, variable-swept-high-def whatsis, automated networked wifi enabled real-time gui guff.

I guess I first got spoiled, then dismissive. For most of my career, I had access to labs with equipment that was regularly and expensively updated to do the latest, fanciest, most esoteric stuff available. I found that I spent a long time every time I used it trying to dumb it down to the tests that I needed most. Well, or dumbing it down to match me, which is another way of looking at it.

And most amp-techies and builders have to be informed what "ESR" means, much less what to do with it.

[madkatb]

So this showed up yesterday at diystompboxes. Purely subjective but would this actually allow you to pick the "better" of two electrolytics?
Checking testing electrolytic capacitors in vintage tube gear using audio - YouTube

[soundguruman]

The primary parameter here is capacitance. ESR is tiny compared to the cathode resistance. The tolerance is 20% for modern electrolytics, that's darn close to a whole octave from max to min. It's one of the big amp secrets: tolerances matter.

Well Said, I say. And the R/C constants matter too.
This One I wanni:
Instek LCR821 High Precision LCR Meter, LCR-821 Instek on Sale at Test Equipment Depot

[R.G.]

... Purely subjective

... oops. "Subjective" means "whatever you thought, whether it matches the real universe or not".

but would this actually allow you to pick the "better" of two electrolytics?

Sure - depending on your mood and how you felt that day, whether you'd been to a concert last night, and how the bean burrito at lunch was.

I'm big into getting some kind of measurement, without going all hifi tweako about how good the instruments are.
(1) As Lord Kelvin said:

"I often say that when you can measure what you are speaking about, and express it in numbers, you know something about it; but when you cannot measure it, when you cannot express it in numbers, your knowledge is of a meagre and unsatisfactory kind; it may be the beginning of knowledge, but you have scarcely in your thoughts advanced to the state of Science, whatever the matter may be."

(2) There are limits of any measurements. Every measurement is limited in its accuracy, resolution, and repeatability, which are all different things.
(3) Effective use of measurement suits the effort expended to the purpose for which the numbers will be used. If you can measure to five decimal places, but only need to know within 10%, you're either wasting time and money, or you have someone else paying for the instruments. If the result of an inaccurate, poor resolution, or non-repeatable measurement is severe, the effort you should expend to make the measurement should go up. Measuring the heat and neutron flux in a nuclear power reactor warrants some serious effort at accurate, timely measurements. If you just want to know which capacitor sounds better in your amp, shoot, that's easy - try them.

Be aware that subjective "measurements" are just that - the next guy, or you the next day may come to a different result.

[soundguruman]

... oops. "Subjective" means "whatever you thought, whether it matches the real universe or not".


Sure - depending on your mood and how you felt that day, whether you'd been to a concert last night, and how the bean burrito at lunch was.

I'm big into getting some kind of measurement, without going all hifi tweako about how good the instruments are.
(1) As Lord Kelvin said:

(2) There are limits of any measurements. Every measurement is limited in its accuracy, resolution, and repeatability, which are all different things.
(3) Effective use of measurement suits the effort expended to the purpose for which the numbers will be used. If you can measure to five decimal places, but only need to know within 10%, you're either wasting time and money, or you have someone else paying for the instruments. If the result of an inaccurate, poor resolution, or non-repeatable measurement is severe, the effort you should expend to make the measurement should go up. Measuring the heat and neutron flux in a nuclear power reactor warrants some serious effort at accurate, timely measurements. If you just want to know which capacitor sounds better in your amp, shoot, that's easy - try them.

Be aware that subjective "measurements" are just that - the next guy, or you the next day may come to a different result.

That's why you need a very large and expensive meter that goes: "boing" Nobody can dispute that measurement.
but subjective can also be a factor that indicates the result of an actual causation. Why "do" we develop the subjective impression?

Is it that we think something sounds "better," and that there is an actual component / construction / material related reason?
Or is it just that we have not created a way to measure mojo more accurately, in the field??

[R.G.]

That's why you need a very large and expensive meter that goes: "boing" Nobody can dispute that measurement.

I always used to love it when I noticed someone telling me about measurements that they didn't understand. The right question at that point would reduce them nearly to tears. I generally never questioned measurements by the ham-radio-since-my-teens guys. They had learned the hard way that it's real or it's not, and the instrument doesn't make it so.

but subjective can also be a factor that indicates the result of an actual causation. Why "do" we develop the subjective impression?
Is it that we think something sounds "better," and that there is an actual component / construction / material related reason?
Or is it just that we have not created a way to measure mojo more accurately, in the field??

Good question. The answer has more to do with the human mind than it does with measurement. The human mind is a pattern matcher without compare, at least for now. The human mind will find patterns in purely generated noise, where there *is* no data. Worse yet, it's been recently found that the human mind will make up the end result in the absence of data.

For a long time, physiologists have been nagged by the difference between the time of human perception of the world and the speed of signals on neurons. For instance, it takes about 200mS for sensory perception from the eyes to get to the conscious levels of the brain. How the devil do we ever catch fly balls, or block incoming roundhouse kicks?

The answer was found in some work on human perception of visual illusions. How can we really, no fooling, see things that are not there? The answer is that our visual systems are always guessing ahead by the nerve transmission time, and INSERTING WHAT IS PROJECTED TO BE THERE for the conscious mind to work on. The data is updated in arrears when it arrives and disagrees with what was proposed. To catch a falling object, your visual subsystems present your conscious mind a picture of where it thinks the falling object will be in one nerve-transmission-and-evaluation-time. If this is very different from where the object actually goes, you missed. For small time lags and low object speeds, this works pretty well. Ask a good boxer.

But for optical illusions designed to trick the subsystems into substituting a familiar, but wrong, picture we get surprised.

The propensity for subjective evaluators of audio phenomena to hear differences that are not there are well documented. If there is no difference, and you're looking for differences, what could be more natural than to hear perceptual differences below the threshold of other, less sophisicated listeners?

That doesn't mean that the human hearing (and vision, and taste, and feel, and...) systems *can't* dig subtle yet real patterns out of noise. But it does mean we should be incredibly suspicious of miraculous feats of perception. They may be real. But it's more likely that they aren't. Humans make things up. This is very good, it makes us imagine a future better than the present. But it's something to be aware of and very, very careful of when we're tying to find out how the universe *really* works.