How do power tubes affect the sound?

[Joe Bee]

This is actually two questions in one, regarding both different types of tubes (e.g. 6l6 vs el34) and also the condition (emittivity, whatever) of the tube in question. I often hear people talk about how one tube is more mid-rangey than the other, and how old tubes get dull as they age. I am wondering what parameters of the tubes are causing this.

With triodes my first guess would be that the gate-to-cathode capacitance is different for different tubes, and it shorts high frequencies to ground, thereby giving different tubes a different timbre. With power pentodes this should not be the case, since they have a screen to avoid this capacitive coupling, right?

Then, aging of the tube may affect maximum current and gain of the tube, but again I don't see how this effect is frequency-dependent. I am sure the parasitic capacitances don't grow as a tube ages!


Finally, what I am trying to get at is this: I am a fan of small tubes, e.g. pairs of el84 for 20W or so, but since lot's of people say that they sound bright and harsh, and I feel that I am missing out on the smooth sound of 6l6's or whatever. If this is just caused by smaller parasitic capacitances (which would make sense in a smaller tube) then it would be easily cured by just adding a few pF from grid to cathode.

Or is is that the nonlinear characteristic of the tubes are so different that audible differences are mostly due to the harmonic spectrum even at relatively low power levels? This would at least explain the aging effect, maybe tubes become more linear with time

[MWJB]

You are only looking at one perameter to explain the differences between different power tubes, when in fact there might be countless physical differences (some maybe measurable, many more may not), even between tubes of the same broad specification, but from differing manufacturers.

Broadly speaking (VERY broadly) smaller tubes have sweeter, smoother highs, more compression. As you go larger, dynamics get harder and the highs can get harsher and the envelope of attack & decay might be reduced. Bigger amps, with bigger tubes, stiffer power supplies, fixed bias, more efficient speakers & better dynamics are generally more desirable for bigger stages as they are easier to hear from the back line. If you are happy just hearing your amp via monitor, you can just as easily mic up or line out a small amp.

Of course a lot of cheaper/smaller practice amps use smaller tubes because they are cheaper & require cheaper transformers/speakers/less parts etc., these have power tubes that are often "cathode biased" which also affects envelope, as might the type of rectifier the manufacturer has decided will do an adequate job at an acceptable price.

At the end of the day, in a studio, or mic'd up (where outright volume may be of secondary importance) it may be possible for a player to get a similar, benchmark tone from differing amps with differing power tubes. So, to say a particular tube sounds like this, and another sounds like that, is not always relevant. There is an issue of context here.

Also for different tubes to run in optimum conditions, they need to run at very different voltages & currents. They also make differing demands from power supplies and heater circuits. It is very difficult to compare different tubes in the same amp at optimum performance levels...put a 6V6 with a softer rectifier into an amp that happily pushes 6550 or KT88, beyond anything else, the amp will sound different simply due to the different demands made of the power supply and subsequent oversized output transformer.

"Finally, what I am trying to get at is this: I am a fan of small tubes, e.g. pairs of el84 for 20W or so, but since lot's of people say that they sound bright and harsh,"...if you like the sound of them, and they are loud enough for your purposes, why care what other people say?

Tubes wear out and sound different, they may even sound different after a few hours of use, ideally they should stabilise after a period of burning in and remain stable for a reasonable period (say 1000 to 2000hrs?). After which, it's probably best & less aggravation, to replace with a new, efficient pair than to try "stop gap" measures to artificially try and prolong their usefulness. Tubes are designed to be replaced and are relatively cheap in most cases.

There are amps around that run varying tubes at good enough performance levels to make music, that can gve you an idea of how the different types & brands perform in the same circuit...this is perhaps the most practical way to finding the answer to your question.

An analogy - imagine if I said that all beer is made from water, hops, barley malt & yeast...what makes it taste different? Well, you've got 4 answers there already.

[Steve Conner]

I believe EL84s sound harsh because they have high gain. High slope tubes are easier to overdrive, and clip harder when you try to drive the grids positive, generating more high-order harmonics sooner.

I also think the gain of the tube has an important effect in so far as it interacts with the feedback loop round the power amp and O.T. Higher gain tubes will give tighter, thinner bass by damping the speaker resonance more, and the extra loop gain will also make the presence control work more.

I argue that these two effects are what cause the perceived brightness and harshness. So if my theory is right, you can get a more 6L6-like sound by messing with the feedback network to reduce your power amp loop gain, and/or maybe adding unbypassed cathode resistors to the EL84s or whatever.

But remember, EL84s worked pretty well for Brian May! (Doesn't the AC30 have no negative feedback at all, though? You can't get much lower loop gain than that...)


BTW, I don't believe that tubes are responsible for "compression" directly. It's sagging screen and plate voltages that cause compression, and the tube is just a second-order influence in that it draws current to make the plate and screen supplies sag, and that its gain and headroom is modulated by the plate and screen voltages.

I'm also not sure I buy all the stuff about magic harmonics coming from tube curves. (the kind of low-order distortion that folks sometimes call "compression" too.) These add up to maybe only a few percent of THD, which is about audible in a blind test. Guitar tone to me is all about big, drastic effects, things driven into blatant hard clipping, massive peaks and dips in frequency responses and the like.

[MWJB]

"BTW, I don't believe that tubes are responsible for "compression" directly. It's sagging screen and plate voltages that cause compression, and the tube is just a second-order influence in that it draws current to make the plate and screen supplies sag, and that its gain and headroom is modulated by the plate and screen voltages." Hi Steve, I'm not sure that I can agree with this. Take an amp with 6L6/6550/KT88, replace with 6V6 and bias accordingly (I have 3 amps in which I can do exactly this). Even though the 6V6 is drawing less current the amp will still compress more, as the 6V6 IN THE SAME ENVIRONMENT will display a more noticable envelope of attack & decay. There is undoubtedly more going on than demands on plate & screen supplies. Even tubes of the same designation from different manufacturers will display differing envelopes, despite being set up as close as dammit in the same amp.

Even if they were not responsible for compresssion directly, it is still what you hear.

[Steve Conner]

Well, maybe it's the coupling capacitors to the grids recovering from clipping at the grids, then. 6V6s will be clipped much harder at their grids for a given volume, for the reasons I explained above.

I still don't believe that tubes can cause attack and decay by themselves. They don't have any mechanism in them that could modulate the sound with a compressor-like time constant. If you hear an attack and decay, it has to come from RC time constants elsewhere in the circuitry.

I could be completely wrong though! I also have an amp that'll take KT88s or 6V6s, so I'll try to check it out :-)

[MWJB]

"6V6s will be clipped much harder at their grids for a given volume,"...indeed, and take away this "given volume"/db at the speaker and you could possibly get a similar, ballpark tone accross more than one tube type.

At a given volume, what will be the audible effect on dynamic response, note separation and envelope on that 6V6, that's clipping hard, compared to a typical 6L6 or bigger tube that's having more of a walk in the park? Clipping, distortion etc. are broad, generic terms, a distorted tone will have numerous components, all dovetailing to give the final result.

[loudthud]

I tend to agree with Steve to a high degree. I would like to add that I think that clipping behaviour has a strong influence. Screen voltage and screen impedance have a large influence on clipping and voltage gain. Kevin O'Connor claims that increasing the screen resistors on EL84's tames the harshness.

How a tube behaves when clipping into a resistive load, does not predict how it will preform with a speaker as a load. Has anyone ever noticed that a tube amp can supply more voltage to an 8 ohm speaker than it can an 8 ohm resistor? Ever heard of Tube Watts?

[Old Tele man]

...a purely resistive load produces a flat load reflection, but a speaker has both capacitive and inductive reactance and thus reflects an "ellipitical" loading...going both ABOVE and BELOW the equivalent flat load of a pure resistance.

[Sir Cuitous]

Though I have less experience than these esteemed participants I also don't believe the tube is "directly" responsible for the compression of the signal. I believe that the biasing method does have something to do with it. In my opinion, a cathode biased tube can behave as a compressor. I believe that when a note is struck, the power tube(s) draw more current. Some of the extra current passes through the cathode resistor. At a certain point this will cause a more positive voltage at the cathode which then cuts the tube current (compression). As the note decays, less voltage is drawn through the cathode resistor which then causes the tube to draw more current. This is the sustain or "swell" that you can hear in some cathode biased amps. I may be wrong, but your not going to get that from a fixed bias tube. That said, please go easy as you trample my ideas one by one.

[MWJB]

Power tubes will often still sound compressed in a fixed bias amp if hit with a big enough signal. Fixed bias or cathode bias is straying from the point a little...the original post asked how different types of tube sound different, rather than different methods of biasing (which I touched on before). What about fixed bias Class A, how does that sound?

Take an amp, with a speaker & 6550/KT88 run it up to 25-30W and listen, now sub the power tubes for 6V6s (shuguang/TADs would bring home the point best - these can even give a more compressed tone even when compared to other 6V6s), get the amp back up to the same output as before & listen again.

If you have an amp that this tube swap is feasible in, it only takes a few minutes. Better still, do it like I have done it in the past with 2 identical amps (well, as close as is reasonable to expect) for an immediate A/B comparison.

EL84 tubes are often fitted to amps without NFB loops (as Steve pointed out earlier), if you put a tighter/harder/bigger tube in that same circuit, it will probably sound harsh too. Conversely, take away the need for outright volume and put the EL84s in a similar circuit to what you might expect from typical 6V6/6L6 amp and you may well be able to dial in a comparable tone (maybe not identical, but enough to make identifying a particular tube simply from listening to a recording virtually impossible), you might need to alter preamp gain as well.

The only way to really compare what different tubes sound like is to run them under close as possible conditions in the same amp. People often associate a tube with the way that a particular amp (that's fiitted with that particular tube) sounds, without ever having heard the same amp with different tubes. The larger part of the equation is the circuit that the tube is in.

[Joe Bee]

Thanks for your answers guys, I think my initial question has been answered to an extend I had not even hoped for! Looks like I was totally off with my initial guess about parasitic capacitances, because Steves explanation regarding gain and feedback makes a lot more sense. (BTW, hi Steve, I am joe from the 4HV board.)

FYI, the reason I was asking is that I am in the process of modifying a small 4x PL84 Combo with the preamp circuit from a SLO-100, and lots of people were telling me that it would sound very different with the smaller power amp. Therefore I was looking for a way to get the small power tubes to sound more like big ones.

[Joe Bee]

Sorry to double-post, but I have two small questions which really are not worth their own thread, and they "kind of" fit the topic.

First of all, how do I find the maximum anode voltage from a data sheet? Usually there are two figures quoted, Va and Va0, with the latter usually being twice of the former. Which of these two is relevant for me?

The second question is not quite serious, but since my main amp runs on TV-tubes which are available VERY cheaply, and I have an old quad that I want to get rid of, I was wondering if there are any interesting ways to run them way out of spec. Modifying the bias circuit to get them to red-plate intentionally would certainly be an interesting experiment both visually and sonically, but are there other crazy things I could try? I would really like to coax a blue glow out of my power tubes, but since the PL84 run at only 250V anode voltage, I guess this is not very likely to happen?

[Steve Conner]

The maximum anode voltage before anything bad happens (flashover, destruction, thermal runaway, arcing between base pins, etc) is pretty loosely defined. It depends on a lot of factors that change with age of the tube, manufacturing quality, anode temperature, etc etc.

Fender ran 6V6s way over their datasheet maximum voltages in a few amps, and got away with it. I tried running some at 475V for a while with no apparent ill effects. The answer is probably "Whatever you think you can get away with".

If you want a blue glow, just buy brand new tubes, old tired ones don't usually glow.