Why do different brands of the same type of tube sound different?

[Voxbox]

Uumm, the title kind of says it all, really.

I know that they can all be made differently mechanically to some extent, but does this result in a difference in tonality, and if so, why?

If it is not this, then what is making the difference?

Would it be possible to design a tube to sound a particular way?

 

[eleventeen]

Probably the biggest factors are amplification factor (mu) and to a lesser degree, interelectrode capacitance. In guitar amps, we are not necessarily liking perfection as we might in a laboratory instrument. One could make great big plates and increase the interelec capacitance, but at audio frequencies that's probably not much of a factor. Realize that you or I might well prefer the sound of a tube that is markedly inferior in terms of "pure" performance as measured on a bench full of test gear. We might like slightly more noise if it gives us greater dynamic range; or, we might like a compression effect that others would reject. At the same time, if I wanted a huge plate structure for more current-carrying capability, then it's awfully likely that my cathode would have to be bigger and then I would need a bigger filament to heat it up and then the filament current specs wouldn't be the same as 12AX7.

Manufacturers typically designed tubes to be able to run the highest frequencies or the highest currents possible or to withstand the highest plate voltages possible; within the nominal type application. No manufacturer tried to make a 12AX7 run microwave frequencies or handle 200 mils of current or to take 1 KV on the plate. The challenge for Sylvania, for example, when making a type that was brought out by RCA, for example, was to match some levels of minimum performance. Once RCA came out with the type, Sylvania probably wanted to be able to sell tubes into applications designed for that RCA type.

From the sheer number of tubes available, one might surmise that tubes were designed for particular and pretty specific applications. Color TV, for example, suggested (if not demanded) tube types that the manufacturers simply didn't think of building or think they could sell enough of to make it worth it before 1954. If you go back to the 40's, there were many fewer tube types. Over time, introduced were were damper diode types, very tough horizontal oscillator types, teeny little nuvistors to get the electrodes really, really close to improve UHF performance and very small to reduce interelec capacitance. I have 6016 rectifier tubes pretty similar to 5Y3's but they are rated for (I think) 20,000 feet and their filaments take a solid 2.5 minutes to warm up. You can put them into a tube tester and they will show as absolutely dead for about 80 seconds. You're thinking "what the hell can go wrong with a rectifier that's lighting up?" 

I have little doubt that the manufacturers worked long and hard experimenting trying to make new types that they thought they could sell. Strategically, they also probably built tubes of special types to crowd other manufacturers out in specialized applications like aircraft, or, adapted existing types for specific design objectives. The improvements in terms of building the tubes more rugged, or capable of running at higher freqs, or lower voltages, came with time, experience, market perception, customer demand, and better machines to build the things. Ultimately, it was all about whether they could sell the things.

[Voxbox]

Thanks Eleventeen.

It would be nice to look at specs to decide what something sounds like, but life aint like that unfortunately.

[HotBluePlates]

Nobody has a good, believable answer to your question. I've heard some big tone changes with different brands, and sometimes different construction, of tubes. The big problem is that there are too many variables to keep track of.

And one of those variables is the circuit the tube is placed in. I would suggest that a circuit with a large amount of feedback, or with very conservative operation (much headroom beyond what is demanded of the tube), will probably have less "tubey-flavor" and less variation from tube to tube. In fact, there are circuits (not always for guitar amp use) where an overriding design choice is to design with a fairly large amount of feedback to reduce performance variation between tubes or over the life of a tube.

You may have noticed that some amps exhibit a lot of change with a tube swap, and others nearly none. I think you're hearing what I'm describing in that case.

I would suggest that fully bypassed cathode resistors, and the absence of other forms of feedback, are your best chances for hearing variation. That might also mean you need to specially select tubes that sound good.

But if you want to characterize tubes and discover why they sound different, you'll probably need to devise some good tube tester circuits, including curve tracing, and study how the circuit exhibit the tonal change is operating each tube. The curves on the data sheets are really just for a "bogey tube." They could be very different from tube to tube, as well as the characteristics of Gm and rp. Do these changes alter how the stage is working, and how it impact the following stage?

[Voxbox]

Thanks HBP, good point about differences being dependant on circuit design.

I've done brief tests on a cryogenically treated ECC83 from Watford Valves in the UK, but didnt hear any difference, but I only tried it in one amp. Maybe I'll hear a difference in another amp. Maybe I'll go mad before then!
Similarly, I used to try out different speaker cables with the hifi, and managed to stop before insanity set in. (I hope).
Such things can lead to madness......

Cheers

[jjasilli]

By analogy:  different batches of the same paint look different.  Different production runs of the same ammunition may shoot differently; so precision shooters note the batch numbers printed on the boxes.  Electronic components often have wide tolerances, like 20%.  But that's +/- 20% for a whopping 40% margin of error.  It's a wonder any 2 amps sound alike!

Also there are different build techniques for the tube to comply with spec.  This and that type of plate, etc.  Not much talk of that here, but lots on hi-fi forums.

[Ritchie200]

Not to mention mechanical tolerances of individual parts and also assembly.  Then there are metallurgical differences....  All combined, that is a BIG list!

Jim

[theundeadelvis]

Just about everything in manufacturing is like that. It's like buying a medium Fruit of Loom T-shirt versus a Hanes. They're both medium T's, but dammit if one doesn't fit great and the other chokes the heck out of you. 

Of course nowadays it seems like they are making those medium T's a lot smaller than they used to.