Dark sounding tubes?

[Gambon88]

I see a reoccurring thread on other sites. “What tube will reduce the high end on X amp”. Of course there are twenty opinions and no consensus. I’m confused, shouldn’t tubes have a relatively flat response across the audio spectrum? Any thoughts on why some tubes would have a different frequency response?

[g1]

Since you only have 2 posts here, I'll have to ask if you are sincere or trying to stir the pot.
This is probably one of the most opinionated topics you can post here.

[Leo_Gnardo]

shouldn’t tubes have a relatively flat response across the audio spectrum? Any thoughts on why some tubes would have a different frequency response?

Yes, and one of our resident experts claimed he measures for instance 12AX7 of all sorts to be ruler flat from DC to beyond 60 MHz, "on a test stand." But we don't use tubes on a test stand, often they're crammed into a relatively small box with a speaker only a couple inches/centimeters away, shaking the bejeezus out of everything. Funny, nobody seems to measure tube frequency responses in this real world setup. In the realm where tubes are subject to such mechanical vibration feedback they reveal what tonal nature they have. At one end of the spectrum, I find chinese Shuguang 12AX7's that ring like a tiny bell and/or rattle along with every pick attack. No problem, some people like that! Me, not so much. At the other end of the spectrum, JJ 12AX7's that don't obviously add any character of their own, simply amplify, and "fade to black" at the end of the note. In between, all the others. Because of this lack of resonances some people like to slag JJ and call them "characterless or lifeless or dull."

If you care to search MEF's back pages you can find discussions that go on for hundreds of entries, page after page of argy bargy that essentially comes down to the s.o.s.

It would be much more satisfying to search & find differences in tone thru a wide choice of speakers available, both current issue and used ones from the good ol' days. No question that you'll find lots of tone choices there.

[Gnobuddy]

Of course there are twenty opinions and no consensus.

Welcome to the wacky world of DIY audio.

Guitar amps are at the intersection of music and engineering.

Music is entirely subjective, and there is always a lot of superstition and unsupported belief involved. For example, Stradivarius violins have been rigorously shown by scientific methods to not be the best sounding ones, but, unsurprisingly, nobody in the classical music world cares about the scientific truth, and Strad's haven't lost any of their silly-stupid value as a result of the finding.

To be fair, superstition and subjectivity is actually necessary to create good music. If *I* don't hear angels and see rainbows in my head when I play my instrument, my audience certainly won't!

I know the angels and rainbows are only figments of my imagination, but I only allow myself to know that after I'm done making music, and am ready to go back to my more left-brained normal state of existence.

I’m confused, shouldn’t tubes have a relatively flat response across the audio spectrum?

Yes, they do. It is possible to concoct 12AX7 gain stages that start to show treble rolloff (due to Miller capacitance) around 15 kHz or so, which would be a problem for Hi-Fi. But I have never seen a guitar speaker capable of reproducing much above 5 kHz, and I have never seen a 12AX7 gain stage that suffers unwanted treble loss before you get to 5 kHz. Ergo, the logical conclusion is that, for sine wave signals, in a guitar amp, one 12AX7 cannot sound any brighter or duller than another.

Remember, that is logic. Logic causes fury among many musicians. It does not inspire nods of agreement. Instead, it often starts flame-wars. This post won't do any better (except I have no intention of participating in any flame wars that might start.)

Any thoughts on why some tubes would have a different frequency response?

Of course you can create different frequency responses on purpose using suitable RC or RL or RLC networks, but I assume you're referring to the natural performance of the tube itself. And in that case, no, there is no reason to expect different frequency responses within the guitar's frequency bandwidth. (There might, of course, be huge differences at, say, 1 MHz.)

I have experienced an odd subjective thing concerning the apparent brightness of an electric guitar, though. If I feed the exact same electric guitar, using the exact same guitar cable, into first an op-amp set up with a 1M input impedance and, say, a voltage gain of 5x, and then into a JFET source follower with the same 1M input impedance and same 5x voltage gain, the guitar signal sounds different. It sounds brighter and harsher through the op-amp. And yet, both op-amp and JFET have a flat frequency response extending far above the guitar bandwidth.

Op-amps are extremely accurate and precise devices, as long as you keep them out of clipping, and within their bandwidth. So I *know* the op-amp is producing an accurate copy of the guitar signal.

The JFET is nowhere near as linear or accurate. I suspect (but have no hard data to back it up) that the big transient voltage at the beginning of a guitar note swings the JFET into a low-transconductance region of its operating characteristics, so the output signal is a little "squished", with the big nasty transient somewhat tamed. I think this is why it sounds less bright to my ears.

Feeding the guitar into a vacuum triode (half a 12AX7) produces a similar effect to the JFET, (after dividing down to compensate for the extra voltage gain). Again, it sounds less bright and harsh than the op-amp. I think this occurs for the same reason as the JFET, i.e. triodes also produce imperfect, "squishy" amplification that reduces the size of the harsh initial transients.

I think this is one of the keys to why so many solid-state guitar amps sound so bad. Accurate reproduction of (particularly solid-body) electric guitars sounds harsh and painfully bright. What sounds good is inaccurate, "squishy" amplification, of the sort that vacuum triodes naturally provide. Subjectively, the reduced harshness sounds like less treble.

My two cents!

-Gnobuddy

[Enzo]

And guitar amps are unlike hifi amps in one important manner: they are never intended to be flat or hifi in the first place.

We don't design amps to run in the linear portion of the tube curves, plus we often overdrive them into distortion. And once there they are not all exactly alike.

Think of cars like your mom drives. Ford Chevy Toyota, they all drive about the same. Now take those cars to the race track and race them. Now they all seem different. You are using those cars in a manner they were never designed for. Same with tubes.

[Chuck H]

Don't care what anyone says (DON'T CARE fingers in ears "LA,LA,LA,LA,LA") the Sovtek 12ax7wa and wb tubes are dark sounding 12ax7's. They're also consistently low in microphonics. They've been tested to a little lower gain than some others, not significant. Good tube for the first slot in modern uber gain stuff that sounds too brash or tends to squeal with a tender tube.

That said, higher gain tubes do tend to "sound" brighter to most people whether they actually test out that way or not. Not sure why this should be. But it would explain the preponderance of opinion on the matter.

For my own part I DO think I can hear a difference in tubes when they're plugged in back to back. And it's not just a difference in gain. Tube rolling has been around long enough that it's somewhat arbitrary to imply that it makes no difference. So there's that.

What makes the difference? How the hell should I know. I build amps, not tubes.

[Gambon88]

Thanks, I’ve no intention of stirring the pot. I’ll search past posts.

[Justin Thomas]

To be fair to those who say tubes are flat across the audio spectrum and beyond, that there's no audible difference between tubes, etc, they DID qualify that statement:

They said that there is no difference between tubes that are of the same manufacturing spec of yesteryear that cannot be made the same, by modifying the circuit. Any differences in tone between tubes is accounted for by manufacturing tolerances and variances, and we all know that NOS tubes were definitely made to much stricter standards than anything today...

The difference to me is, it's a HELL of a lot easier to buy a few different brands of 12AX7 and hear a difference and NOT have to measure my circuits with a spectrum analyzer and other such uber-expensive lab grade equipment to measure every single possible circuit parameter, and tweak the circuit to make the tube sound the same... Tube Rolling is SOOOOOOooooooo much easier and cheaper. Also, maybe if tube manufacturers got their act together (when pigs fly on their own, without Pink Floyd's help) we'll stop arguing over stuff like this...

That said, I'm with Chuck on the Sovteks, and they're kind of my go-to for those reasons. But it also seems that Sovtek has found a way to manufacture a tube consistently for many years, and knows NOT TO @#$% with it if it's GOOD, instead of trying like everyone else to turn it into something questionable.

Justin

[eschertron]

Thanks, I’ve no intention of stirring the pot. I’ll search past posts.

Oh, no you don't! We've just gotten started here

As has been pointed out, guitarists operate tubes in non-linear regions. Also, as has been pointed out, each tube's gain does vary from brand to brand and tube to tube. The amount of distortion produced by a given tube in a given amp, at any given control setting, will likely be different than the amount of distortion produced by the next tube rolled in. Our aural perception of the tube's 'sound' is informed to a degree by the overtones (distortion) generated while operating in the non-linear region. Does that account for all the differences heard by the golden-eared? I dunno, but reason tells me there are real audio-spectrum variations between samples.

If the tubes were operated in RDH4 small-signal don't-even-think-about-exceeding-the-design-center-numbers mode, the valves would indeed test flat and linear. But that's not the real world according to amp aficionados.

[J M Fahey]

Tubes are ruler flat beyond your imagination, just look any tube datasheet (the Manufacturers should know, shouldn´t they )

The circuit surrounding them is not, putting it mildly.

So a tube with different gain or transconductance (which does vary not only brand to brand but individual sample to individual sample) or different (internal) plate impedance (all 3 parameters are related) can and probably will make the external circuit behave differently.

Main problem is that 99% of opinions posted online are based on samples of "one" , as in "I pulled my (ONE) Shuguang tube , replaced it by (ONE) JJ one, and sound is brighter/warmer/meanier/smoother/sexier/duller/acid/sour/sweet/tangy/whatever".
Which of course may be true ... for that ONE Shuguang compared to that ONE JJ.

Problem is that 99% people attribute that to JJ vs Shuguang BRAND, ... all MILLIONS of them ... which is stretching facts a bit.

And it´s compounded by people reading Forum posts and then *expecting¨ it to sound one way or the other, so called "confirmation bias":

That´s why scientifically accepted comparisons are called "double blind tests" , as in: tubes are changed ... behind a curtain, invisible to Musician, and the guy actually switching them (and keeping track which is which at any time) is also invisible to the guy doing the questioning, so he does not transmit info by the tone of his voice or facial expression.

And guess what?
Under that strict test, strongly opinionated people find it hard to pick one from the other, go figure (if they are set to same volume that is, difference in volume is easy to detect by anybody).

[Chuck H]

Agree! I do find it a little arbitrary to say the 'tubes make the circuit behave different" and not attribute that to how the tube sounds though.?. And, fwiw, the Sovtek 12ax7wa and wb tubes are the only tube I'm certain is dark. And I think it's entirely possible that any spec sheet could just be a regurgitated copy of some vintage spec sheet, as is often the case. So how do we know what the actual internal resistances, capacitances and trans figures are? And wouldn't those parameters make it "different". Sure, any tube can produce a flat amplified signal in test equipment, but if this doesn't happen in the real world because of the circuits or otherwise and the change in tube brand makes an audible difference, even a subtle one that is amplified 1000 times, it seems entirely reasonable to me to say that one tube sounds different from another in this circumstance.

A new Chevy pickup might have, say, a given towing capacity, horsepower, load handling, gas mileage, etc. The equivalent Ford product is almost certainly remarkably similar since they are direct competitors and technologies and business models only go so far. Back off a little in perception and they are essentially the same thing. But up close and personal, people will demonstrate a preference and may even site performance differences, however subtle, as being eminently noticeable. Maybe not my best analogy, but I'll stand by it.

[eschertron]

So how do we know what the actual internal resistances, capacitances and trans figures are? And wouldn't those parameters make it "different".

Good point! We know, for example, that certain inter-electrode capacitances combined with external circuitry change the gain in a frequency-dependent way (Miller effect). A significant change in internal capacitance can change the frequency response audibly. Is the actual capacitance value notated on the tube's spec sheet? Maybe not. Probably not, if the spec is a simple copy of a previous manufacturer's data sheet.
To take this a little closer to conspiracy-land, how do we (collectively the techs and engineers who inhabit this board) know that the spec (specific example: grid-to-plate = 1.5pF) that may be true at test voltage, heater power, etc., does not change at higher voltages, temperatures, gassiness? I'm thinking that the tube's vacuum as dielectric material is not constant. A tube heats up, the gas inside ionizes, and now the dielectric is not vacuum (per test) but a plasma. I'm not a physicist, but I think the plasma may have different dielectric characteristics. Of course this falls in line with Juan's "sample of one" argument. But I don't think it's unreasonable to assert different tube plants will have different levels of residual gasses - consistently - in their tubes.
I'm not claiming that plasma causes tubes to sound different. I'm merely suggesting that data sheets don't account for every condition of the tube. Only the ones commonly used for engineering modeling. The ones that are used to prove that a tube is a tube is a tube. :G. Stein: