Carbon Comp ?

[loudthud]

I must confess I was expecting the CC to show more distortion. The CC resistor looks like an Allen Bradley with no mold marks, but the color is a little on the dark side. I got 'em on ebay, I think they were like stock in a college electronics lab. There were lots of different values in little zip lock bags.

[kleuck]

Then your experiment is flawed, sorry.

WHY ?

You are contradicting yourself: if you didn't know the switch position, how on Earth could you know which is the one you preferred?

...i stopped my friend on my preferred sound, then i looked at the position of the switch, what do you want to compare if you do not know what you are comparing ?

[daz]

I bought some a couple years ago and tried them on the preamp plates. I too don't get it. They weren't noisy, at least not to any problematic degree that i noticed if at all. But tonally they actually sounded worse IMO. But i bet it's very dependent on amp design and i think they probably might benefit a vintage style amp a lot more than one like mine, a master vol marshall type amp. I put the carbon films back in a was pleased at the better tone they provided over the CC's.

[kleuck]

I bought some a couple years ago and tried them on the preamp plates. I too don't get it. They weren't noisy, at least not to any problematic degree that i noticed if at all. But tonally they actually sounded worse IMO. But i bet it's very dependent on amp design and i think they probably might benefit a vintage style amp a lot more than one like mine, a master vol marshall type amp. I put the carbon films back in a was pleased at the better tone they provided over the CC's.

So you heard something I too think that the benefit, if one at all, is more audible in vintage/simple circuit, just like with other components, caps or tubes, and is a matter of taste.Not sure it would be that obvious anyway with lower value plate resistors, circuit comes first, i did not like my Mazda 12AX7 with 100K load (great tube, but more on the Hi-Fi side), it was way better with 220K and the best with 470K, where tubes i did like in casual 100K circuits. I would not use CC in first stages anyway, nor in all gain stages or in stages with nfb (would be stupid), their noise is an issue, but at least with high voltages supply and signal, and high value resistor (i tested with 220K and 470K, not 100K) something happens.

[kleuck]

"...where tubes i did like in casual 100K circuits did not shine." sorry, can't edit.

[Joe L]

My favorite resistors for new builds. 294-100K-RC Xicon | Mouser

Cheap, quiet, 1W,5% tolerance, 500V rating, don't drift.

[bob p]

The graphs were identical at 20Hz, 1kHz and 20kHz. The rise in distortion below about 100mV signal is due to noise from the oscilloscope. This is the 1kHz graph.

I applaud your objective application of the scientific method. I can't see any difference in distortion as a function of voltage.

Of course, you know how this goes -- someone with Golden Ears is going to tell you that your test equipment lacks the resolving power to see what he can hear with his Golden Ears in a non-double-blinded experiment.

[bob p]

I can't see the point in buying 5% tolerance when 1% tolerance is so cheap.

[kleuck]

I applaud your objective application of the scientific method. I can't see any difference in distortion as a function of voltage. Of course, you know how this goes -- someone with Golden Ears is going to tell you that your test equipment lacks the resolving power to see what he can hear with his Golden Ears in a non-double-blinded experiment.

I don't pretend to have golden ears, nor does Daz, but i love the idea of double blind test Another friend of mine, a good guitarist came to play at home.After a few minutes, he said "You modded your amp once again ? I feel it more "organic", touch-sensitive, and loud" I actually did nothing to the amp, but replaced the speaker for an alnico one. I should have said to him "No, you can't say that, you played the amp a week ago for the last time, we should do a blind test with a relay and someone else in another room !You can't hear a difference !" ....

[kleuck]

AND, the fact that the overall distortion level is the same doesn't mean that its harmonic content (what is all about) is the same, CC are supposed to provide even order distortion, maybe it's a trade for uneven ones. And it WAS a double-blind test, i did not know the position of the CC, neither my friend (who is not a musician, nor an electrician)

[J M Fahey]

I don't pretend to have golden ears, nor does Daz, but i love the idea of double blind test Another friend of mine, a good guitarist came to play at home.After a few minutes, he said "You modded your amp once again ? I feel it more "organic", touch-sensitive, and loud" I actually did nothing to the amp, but replaced the speaker for an alnico one. I should have said to him "No, you can't say that, you played the amp a week ago for the last time, we should do a blind test with a relay and someone else in another room !You can't hear a difference !" ....

You must be joking or you are incredibly dishonest.
I vote for the second option because I'm not laughing at the first.

Now you switched mid discussion from a resistor which as proven and measured shows *no* difference to a ***speaker*** , which has different coil/magnet/suspension/adhesives/age/resonance/inductance , but, most important, shows **measurable** differences.

So far I had respected you , just read my earlier posts, but now you are being incredibly dirty.
C'mon, we are grownups !!!

Don't argue like a kid.

[kleuck]

No, you just pretend that nothing is true unless proven by a double-blind test, i pretend 1) to have actually performed one 2) that ears (and fingers) are (sometimes) able to perceive differences even after a few days and without a direct comparison 3) that the experiment performed by Louthdud has a limited interest as datas are limited (no harmonic content) 4) my experiment has been made with higher value resistor and different bias 5) therefore his readings applies only to his amp and experiment condition Actually i pretend nothing, these are simple facts.

[R.G.]

The mojo of CC resistors is said to come from the fact that they change resistance slightly as the voltage changes.

As far as I know, I'm the one that first said that, so I have a dog in the hunt.

If I understand your experimental setup correctly, I would expect the results that you got - name, not any detectable difference.

The signal swing across the CC resistor needs to be BIG. Like 70-150V peak to peak. If I understand your setup right, the plate signal was not necessarily big.

CC's have a specified voltage coefficient of resistance of X% per volt. So at 0V, they're whatever they are. At 100V, they're slightly different. To cause distortion, you need a signal that swings across many volts. The signal changes as the current from a plate sees one resistance at low voltage across the resistor (that is, near plate cutoff) and high voltage across the resistor (that is, maximum peak current in the plate). One end of the swing is larger than the other, effectively compressing either the negative or positive peak. I'd have to go look at whether the voltage coefficient is positive or negative.

In any case, it is possible that newer production methods have brought down the voltage coefficient, and the distortion.

How big was your plate signal swing?

[loudthud]

I did the experiment because I needed to get setup to look at the 2nd stage of a 5F6a. That experiment is complete for now, I could go back and do the CC vs CF thing again. It was done on a little chunk of metal with 4 noval tube sockets on it, not a complete amp. I didn't connect the spectrum analyzer. I could do that but you have to decide what level to set the generator at, probably something just below clipping. I could also scrounge up some 100K resistors out of a vintage amp or other sources.

With only a 10 meg load on the tube, it had a gain close to 60. With 2Vrms input, the output was well past clipping. The problem I had was that I didn't want the scope to clip on the biggest signal so it had to be set to 50V/div. The scope output is fixed at 50mV/div so it was effectively dividing by 1000. You don't want to change the scope attenuator while the generator is sweeping. I did run some sweeps with the scope's sensitivity turned up and the generator turned down. It moved the valley of the distortion plot down and to the left. Not significant because of the interest in big signals.

In general when distortion rises with signal level, that is usually a characteristic of low order distortion in single ended amplifiers or clipping. When distortion rises as the signal is reduced, that indicates that the distortion is dominated by noise or crossover distortion. When I looked at the distortion output of the analyzer (below clipping) on a scope, it was mostly 2nd harmonic with some 3rd.

The notion that distortion might be 2nd harmonic with the CC load and higher order with the CF load yet give the same THD number might be valid at a single amplitude, but with the close tracking over two decades of amplitude change, it is highly unlikely.

[J M Fahey]

No, you just pretend that nothing is true unless proven by a double-blind test,

It's not *me* but the whole bulk of Science.

Some definitions:

1) test (tst) n.
1. A procedure for critical evaluation; a means of determining the presence, quality, or truth of something; a trial:

Double-blind trials
Double-blind describes an especially stringent way of conducting an experiment on human test subjects which attempts to eliminate subjective, unrecognized biases carried by an experiment's subjects and conductors. In most cases, double-blind experiments are held to achieve a higher standard of scientific rigor than blind or non-blind experiments.

So you don't want the experiment to be objective or repeatable?
You want to be able to say "look, *now* sound is horrible, *now* it's great" but at the same time you play in a different way.

When you play with the CCR you play well, when with MF or MO you play bad and call me deaf if I don't hear a difference.

Fact is, I *will* hear a difference, only problem is that it won't depend on the resistor type but on your playing

Now , the beauty of double blind testing lies that you *really* don't know which resistor is connected, so you don't know how and when to play better or worse. You can't bear that.

That would be "blind"; why "double blind" then?
Because in that case the observer in that room does not know either, so you can't guess based in his face, voice, tics or whatever .... because as I said he doesn't know either.
Double blind is a *guarantee* of imparciality , simple as that.

Am I making this out?

No, just see the official definition:

In a double-blind experiment, neither the participants nor the researchers know which participants belong to the control group, as opposed to the test group. Only after all data have been recorded (and in some cases, analyzed) do the researchers learn which participants were which

In this case "the participants" are the resistors and you only are told which is which *after* the control sheets are filled.

i pretend 1) to have actually performed one

Fine, I believe you made that test once.

2) that ears (and fingers) are (sometimes) able to perceive differences even after a few days and without a direct comparison

No. I can't trust your experiments performed days after and without direct comparison. Period.
This has been amply proven in acoustics tests.

To be more precise: if you "want to hear" something, you "will hear" it, no matter what the reality .... but that's something that happens in your mind, not the real world.

3) that the experiment performed by Louthdud has a limited interest as datas are limited (no harmonic content)

How come?
Loudthud's experiment is perfect.

It shows that there is very little distortion (so little that part of it gets buried under the noise, go figure) and the very small amount that exists, comes much more from the tube (a *known* non linear element) than from the resistor used (a known linear element).

Let me add that even if we apply, say, 100V RMS to the resistor, even if it produces a little distortion (this remains to be proven, by the way), the tube will produce gobs more distortion than that, which will absorb and mask it and make it irrelevant.

4) my experiment has been made with higher value resistor and different bias

Irrelevant. Not enough to justify your theory.
And you are contradicting yourself again: at most it would prove that the 12AX7 reacts different with different load and bias, which may be true but is not the point of the experiment.

Loudthud used typical 12AX7 tube gain stage values, made a "clean" (unbiased) experiment, and I can trust results would be the same at some slightly different bias point.

And personally, if I had to choose "just one" set of values, I would also have chosen classic 100K/1K5 because it's the most universally used so it's the most useful.
You are desperate trying to find a justification for the lack of measurable differences in your flawed experiment.

5) therefore his readings applies only to his amp and experiment condition

The way Loudthud performed his experiment make its very trusty.
*Much* more than your flawed/biased one.

Actually i pretend nothing,

żżż??? You look like you pretend a lot, as in disqualifying properly made experiments.
Pity you don't achieve it.

these are simple facts.

Yes, it's a simple experiment, pity you don't want to make it the proper way.