Are carbon films worth the trouble?

[bepone]

Great! I'm glad he was happy with his work.
Unfortunately for the question at hand, his attestation that it was a great improvement is what's called "anecdotal evidence". An anecdote is of substantially no value in proving that something is or is not. This is because people can see and hear things that they just make up inside their heads, and hear them different ways.
This is not to say that he does not hear it as better. He almost certainly does. But does that mean everyone would think it was great? Almost certainly no. So one anecdote is worth about zero at proving things like this. Actually, a thousand anecdotes are worth the same as one anecdote.

Did he make any other changes in his amp while "improving" it?

There is a human bias to think our own work is amazingly good. It's subtle, and all we can do is be aware that we think anything we do while "improving" something is a step in the right direction. And if we believe our bias, it is completely true - inside our own head, but not necessarily outside in the real world.

I know all about my friends build because we were copying one 18W amp and trying to get the same sound as much as possible... in the "orig" version all resistors inside were MOX 2W. I have recommended in the new one to put all low noise MF to the input valve V1 EF86..only to get better signal/noise ratio..
He has finished started the amp, played for a week or two, but reporting some extra "presence" in the sound all the time..voltages were the same in both amps. He was chasing familiar tone, amp which he already had, and playing for many years, and was easy to detect tone of the new amp that is not like that.

So he started to change the differences, MF resistor -> MOX in the first valve cathode EF86.. with that he has killed some amount of the highs and called me that sound lost extra presence and after i was playing that amp too.. so something happened with the highs with that material change.
Did dale just pass all the sound without filtering? Did MOX resistor act like a some very small LP filter and remove some highs that was possible to detect with ears? And why acted like that? Something happened.

This is just one example, i'm experiencing it every day.. so for me there is no doubt, last decade i started to look in the guitar amps tone cooking like a gastronomy, sauce cooking..not anymore like a electrical engineering thing
Then this is about the resistors, if we add capacitors in the formula ...

[bepone]

one important thing to highlight, i will be more happy if i hear no change, for sure

[ChopSauce]

R.G., with due respect, I feel you shouldn't make it so complex.

One side note as an introduction. Let me recall to everyone that scientific advances most often come from improved measurement techniques showing discrepancies with past/present theories.

As long as we can't meet all in person with the amps to test and compare, there's very little to conclude. If we were all in the same place and ready for that, we should be able to design a protocole in order to eliminate the various biases (eg : make the same amp tested twice with different settings while suggesting to the tester that there are two different amps, have enough player test the amps so the results are statistically significant, etc).

My opinion is that the so many reported differences over the internet are too consistent to be just negated on the basis of your contradictory experiments and I look forward to know your thoughts about the possibility that the transient nature of the (superimposed) effects of components could be not fully accounted by your protocole ...

... and I mean that just out of curiosity for I have nothing but doubts about all that matter ...

[LOUDthud]

Would it be possible to construct an amplifier with one type of neutral sounding resistor, then substitute an assembly of resistors selected by a switch for one of the resistors ? This would allow rapid comparisons between different types of resistors. 100K would be my first choice substituting for the first triode's plate resistor, then the cathode resistor driving the tone stack. All the resistors on the switch need to be matched to within 100 Ohms or so.

[martin manning]

I had a plan to to swap the plate resistors on the first tube (V1a and V1b) in my ODS using a DPDT switch. That fell through due to lack of a matching pair of magic resistors to swap for the Dale MF. The idea was to support a mini-toggle switch on the two pars of resistors for rapid A-B'ing.

[nworbetan]

This is not nearly as insurmountable of an obstacle as the long winded essay of low expectations would suggest…

That would be really great!

I'm always eager to learn. Sketch out for me in some detail how it's not nearly as insurmountable.

Let’s pretend that’s not the issue…now go

We can do that - pretend that's not the issue. How does that give us an honest, fair and impartial test?

I think at this point the biggest issue preventing an accurate scientific analysis is that nobody seems to be able to point to any specific make and model of resistor that is going to demonstrate the alleged properties. It's an example of "garbage in garbage out". No matter how well designed the test is, if none of the resistors being tested happen to be the magical chosen ones, the most perfect test possible isn't going to detect the magic.

As far as what the test "should" be, my thoughts are that the test should include multiple gain stages in a row that are designed to maximize the phenomenon being tested and the output should be a spectrum analyzer and distortion meter.

[FUCHSAUDIO]

and said "your amp is very similar to this one" and marked up the schematic with a highlighter to show suggested resistors to change...I found that hilarious....

Second easiest modern amp I ever worked on and I should otherwise keep my opinion to myself!

It wasn't tough to work on, but it's tight in there, and lots (and lots and lots) of hot glue and silicone....I like the sound of Carr amps. The construction style I find a bit bizarre.
There are some interesting choices of power supply filter types and some of the passive are obviously chosen with care, but this is another amp line I question how they make a dime on such a time-consuming build.
Yes, he likes carbon films...

[R.G.]

I appreciate the amount of thinking that's going on.

There is/was a school of thought that there are tiny tiny differences in sound from components that instruments and meters can't detect or can't detect yet. The human ear and mind are viewed as near infinite in their ability to discern effects where soulless meters can't find anything. That quickly gets extended to the idea that any attempt to measure audio phenomena with scientific measurement must therefore be incomplete, and therefore wrong.

The test-the-human-first approach dispenses with the issues of whether instruments can detect phenomena that are heard. It merely asks whether a human, perhaps a select few special humans, can (1) accurately and (2) repeatably detect a difference in sound on repeated tests by ear alone. If the human(s) can, great! We then have verified that we can ignore the scientific instruments and just use the person(s) to tell us which direction improvement in sound lies.

As a side note, instruments need testing and calibration too. At a development site where I worked for years, there was a calibration department that cycled through all of the meters, scopes, and other measurement devices on the site, comparing and adjusting their readings to match reference standards from the National Bureau of Standards, just to keep the meters honest.

If the human(s) can't tell more reliably than coin flipping, where does that leave us in terms of components for better sound? Listening to the "vote" on the internet? I'd be happy to include a blind ABX test on a group of Golden Ears if that would demonstrate that a committee could reliably sense component sound.

The idea of simply live AB testing components is as old as electronic music, and probably as old as musical instrument making. That idea has been reliably proven to be highly prone to reflecting conscious and unconscious biases. People who have pre-selected themselves to be good at selecting tonal qualities do not want to believe that listeners in general and themselves in particular can be mistaken about what they hear. They want to believe in tone magic. I do this myself. But I realize that I am biased and want there to be magic parts. But the testimony of one person just reflects their perception at the moment. Multiply that by billions and all you get is a billion possibly unreliable opinions. Show that one person is reliable at detecting differences outside their biases, and you have a way to start testing components.

This line of reasoning is what leads me to say that you have to prove the listener really can distinguish real-world changes better than coin flipping first before you can make any statement about better sounding parts. I realize that I'm shoveling sand against the tide.

Edit: Let me ask another and more important question: is there a flaw in my reasoning? If so, point it out to me.

[nworbetan]

I appreciate the amount of thinking that's going on.

There is/was a school of thought that there are tiny tiny differences in sound from components that instruments and meters can't detect or can't detect yet. The human ear and mind are viewed as near infinite in their ability to discern effects where soulless meters can't find anything. That quickly gets extended to the idea that any attempt to measure audio phenomena with scientific measurement must therefore be incomplete, and therefore wrong.

The problem is that instruments that measure very precisely and accurately in the gigahertz frequencies exist and electronics that operate very precisely and accurately using gigahertz frequencies are cheap and common. The vast majority of humans can't demonstrate any ability to hear past about ten kilohertz. Cheap and common electronics equipment is on the order of a million times more accurate and precise than human ears. The idea that humans can hear things that can't be measured by any test equipment is absolute bullshit on every level.

[R.G.]

The problem is that instruments that measure very precisely and accurately in the gigahertz frequencies exist and electronics that operate very precisely and accurately using gigahertz frequencies are cheap and common. The vast majority of humans can't demonstrate any ability to hear past about ten kilohertz. Cheap and common electronics equipment is on the order of a million times more accurate and precise than human ears. The idea that humans can hear things that can't be measured by any test equipment is absolute bullshit on every level.

There is a slightly tech-mad part of me that screams that kind of concept. I try to keep that in check with reason and logic.

The instrumentation has gotten very good indeed, yet many people still insist that there are properties of certain simple components that have clearly audible effects on tone and sound. I have thought about that disconnect at some length. The instruments are sensitive, reliable, and more importantly repeatable. I am open to the idea that the human ear and mind can find real world things that instruments cannot (yet) "see". That is, as previously noted, how science advances - find something unknown, then figure out how to measure it.

Humans have been reliably demonstrated to be subject to wide differences in what they perceive based on things like temperature, fatigue, expectation, previous beliefs, auditory illusions, opinions of other possibly less well informed or more opinionated people, and the internal desire to be right, no matter what the instruments say. But I'm open to being convinced that there are as-yet-undiscovered effects in resistors, caps, wire, and so on that have not yet been found in about a century of humans using them.

So the more rational voices inside my head say that if the effect is real, not imaginary or the product of bias or opinion creep, that some person, somewhere, should be able to repeatably do better at finding a difference where one exists than one could do by flipping coins. Until that happens, I believe that component tastings are mildly futile. Yep, tastings are good fun, get lots of people involved and active, but the don't (in my opinion) produce anything definitive.

[nworbetan]

The problem is that instruments that measure very precisely and accurately in the gigahertz frequencies exist and electronics that operate very precisely and accurately using gigahertz frequencies are cheap and common. The vast majority of humans can't demonstrate any ability to hear past about ten kilohertz. Cheap and common electronics equipment is on the order of a million times more accurate and precise than human ears. The idea that humans can hear things that can't be measured by any test equipment is absolute bullshit on every level.

There is a slightly tech-mad part of me that screams that kind of concept. I try to keep that in check with reason and logic.

You try to keep reason and logic in check with reason and logic? Sounds good to me.

I agree completely that if there's any phenomenon that can be heard, all it takes is the correct test setup to demonstrate that it exists. If it can be heard and is a real phenomenon, then it can be measured by the equipment that is on the order of a million times more sensitive and accurate than human ears. A failure to do the correct test doesn't prove anything at all except that possibly the person or people conducting the test don't know what they're doing and can't be trusted to come to correct conclusions.

[martin manning]

Seems like with resistors one could measure them with an LCR meter to see if there is any difference in the macro electrical properties. I've done that with new and old MF, and (no surprise) there isn't any. Next you could set up a simple divider using a reference resistor type, run signal through it, and do a FFT on the output at a few different frequencies. Repeat that with some alleged magic types and see if there is any difference in harmonic content. Beyond those attributes, what else could possibly account for a difference in "tone"?

[R.G.]

There is a slightly tech-mad part of me that screams that kind of concept. I try to keep that in check with reason and logic.

You try to keep reason and logic in check with reason and logic? Sounds good to me.

Just the tech-mad part that wants to scream "If you can't, measure it, it doesn't exist!!" Screaming "just measure it" doesn't convince True Believers. A more rational and inclusive approach is needed.

Seems like with resistors one could measure them with an LCR meter to see if there is any difference in the macro electrical properties. I've done that with new and old MF, and (no surprise) there isn't any. Next you could set up a simple divider using a reference resistor type, run signal through it, and do a FFT on the output at a few different frequencies. Repeat that with some alleged magic types and see if there is any difference in harmonic content. Beyond those attributes, what else could possibly account for a difference in "tone"?

It does seem like that should be sufficient, doesn't it? Sadly, that does not work for people who -want- to believe in magic parts.

[WhopperPlate]

There is a slightly tech-mad part of me that screams that kind of concept. I try to keep that in check with reason and logic.

You try to keep reason and logic in check with reason and logic? Sounds good to me.

Just the tech-mad part that wants to scream "If you can't, measure it, it doesn't exist!!" Screaming "just measure it" doesn't convince True Believers. A more rational and inclusive approach is needed.

Seems like with resistors one could measure them with an LCR meter to see if there is any difference in the macro electrical properties. I've done that with new and old MF, and (no surprise) there isn't any. Next you could set up a simple divider using a reference resistor type, run signal through it, and do a FFT on the output at a few different frequencies. Repeat that with some alleged magic types and see if there is any difference in harmonic content. Beyond those attributes, what else could possibly account for a difference in "tone"?

It does seem like that should be sufficient, doesn't it? Sadly, that does not work for people who -want- to believe in magic parts.

It’s weird . I keep hearing this “wanting to believe” statement….

The only thing I want is to get the sound I want using cheap and easy to find in production parts . Like I have tirelessly emphasized: this has proved challenging …

I actually don’t want to believe there is any difference between anything. It’s far easier and cheaper if there wasn’t …

[R.G.]

It’s weird . I keep hearing this “wanting to believe” statement….

Sorry about that. It's not pointed at you in particular. It's just that this very topic has come up regularly on the net at least since the late 1990s, and in much the same rhetoric. If it seems like I'm preaching to you, I'm not. But the audience is bigger than you, as reading the thread will show. Again - sorry, it's not directed at you, but instead the idea that just interviewing parts can produce solid "these parts are better" without answering the bigger questions: is part tone variation and sound coloring beyond what can be measured real or not?

The only thing I want is to get the sound I want using cheap and easy to find in production parts . Like I have tirelessly emphasized: this has proved challenging …

I actually don’t want to believe there is any difference between anything. It’s far easier and cheaper if there wasn’t …

Ok. Here's how to do that.

Get a disinterested party to run a blind ABX test with you as the subject. If swapping carefully value matched parts in such a blind ABX test shows that you personally can reliably do better than guessing at identifying a part change, and for extra points identify that some brand/material/age of parts is always better, great! You have proven that you personally can identify which parts are better to your hearing. If you prove that you in fact can't do better than random chance, you have actually proved something valuable; you've proved that single part variations do not get detected per se by your hearing, and you can quit searching for mythical parts to make the amps better.

This gets rapidly more complicated if you enlarge the range to "does one amplifier sound better with X parts or Y parts?" or "must I just use X material resistors/caps/wire/solder to get a good sounding amp?"

Amp tonality is much like our perception of actual magic. It sure seems to be there when we see the magician do it, but the harder and closer we look, the more the magic evaporates until it can no longer be found at all. I think this is what why you're finding the process challenging. The closer and more finely you look at parts, the harder is is to say that one part or another is better, and if so, how.

I personally do not believe in magic parts. The physics of electronic components has been too well examined for too long by really Smart People trying to find unexpected effects. I am willing to believe that there is something we can't - yet - measure, and leave that possibility open. I would of course like whomever says there is a difference in parts to be able to prove that they can reliably show they can pick it out.

Here's one other thought for you. If there was a magic resistor, or cap, or whatever, that made amps sound GREAT, the folks with lots of money at risk making entire amps would have long ago figured out, even if crudely, what that different part is/was, and be using only that in their amps. We would not need to be having this discussion, as it would be universally known that Acme Belchfire amps sound magically good, and it has to do with their [fill in the parts list]. This is a big question, along the lines of the Fermi Paradox.

I do believe in magic circuits, and carefully setting gains, noise budgets, frequency rolloffs and the like. And in carefully, carefully, carefully characterizing speakers. It's simply amazing how much difference a different speaker can make to an amp's sound, even one of the same brand and manufacturing date. Speakers are acoustic creators, and have their own voices. This can, and has been measured. I find it strange to find yet another internet discussion on whether carbon comp, carbon film, whatever, has a huge effect on tone, and so little discussion of the known huge effect of speakers and cabs. But that's just me.

Sorry to run on. It's not directed at you, but at the larger reader/responder list on the thread, and the far larger audience of guests perusing this thread.