Fantasy tube
William of Occam suggests that it is dangerous to multiply entities
needlessly. So, in striving for the "Perfect Tube", it would be my
contention that overly complex tubes, and/or tubes required to serve
too many purposes become dangerously complex with a logarithmic
propensity towards error/failure.
I deliberately didn't specify my own idea to start with so that
everyone else could have all the court to shoot from. But it is
amazing how alike we all think, at least as long as we stay inside the
envelope; in fact it will look like I took my idea from John, from
Ronnie, from Chris and from Max!
My idea is a twin triode in an octal base glass tube the shape of and
no larger than a 300B, working off a real life (not max) B+ of 385V
and requiring no more than 60V of signal for full output, which should
be in the order of 10W per side (real output at everyday parameters,
not theoretical output at Pdmax; a WE300B is effectively a 6W SE tube
in real life, not the 10W American ponycar amp makers claim).
Indirectly heated filaments might be nice but I don't consider them as
essential as John does; DH fils have never bothered me much. I don't
care how hot the tube gets if the longevity is reasonable. It seems to
me that the 300B/845/211, that entire lot, are tubes whose performance
is depressed in favour of longevity; I have WE300B getting on for 20K
hours, at least, an entirely unreasonable life expectancy, in effect
indestructable in conservative designs.
A 300B on steroids may be a good idea, save that it leads to the
necessary admission that SE designs have inherent limitations in
headroom that could be solved with a "Bigger Tube".... this leads to
the inevitable Americanism: When all else fails, get a bigger hammer.
The last 80 years of tube design have generally shown that after flea-
power comes PP, and that PP systems may be (to use Mr. Turner's word)
"blameless" with good design.
Two tubes in the same envelope so it can be PSE or PP. Octal for cost.
10W per side so that 20W is easily achieved in PSE or more in PP up to
much more with variation among output classes. 385V is chosen for
component cost; it fits into the standard cap for 230V mains, so the
any iso-transformer will do, and the caps are cheap too. 60V signal is
easily achieved with a single or at most two stages of voltage
amplification.
Now, here is where the laws of physics-on-the-cheap get nudged. Heat
dissipation is a function of design, envelope size (exposure) and
other mechanical factors. This would necessarily be a fairly large
tube, on the order of some of the more significant transmitter tubes
and therefore something of a real-estate hog. Furthermore, that sort
of power-handling would also require the perfect socket with perfect
contact surfaces. So far, though, well within good science and (quite-
costly) available technology. But price would definitely be an
object.
And again, putting both in the same envelope does considerably
multiply entities, especially in a power-tube. Advantages: the
presumption that both sections age equally and together.
Disadvantages: The entire tube becomes useless should that not be so.
I'd also specify a second tube as a driver: two 417A in the same
envelope, which can be bigger than the child's thumb of the present
417A, say up to 6SN7 size. (Actually, I'd ask for two 437A in the same
envelope, and settle for the two 417A if the man said 437A are
impossible. A 417A is good up to 24mA and has a mu of over 40; a 437A
is a 417A on a lifelong diet of steroids.) The reason for asking for
20mA and over is because that is what is necessary on a driver to
overcome Miller in PSE 300B or single 845, which have approximately
the same parameters I specify above. Or, if I can't have any of that,
how about 4x 6SN7 in one envelope, or 2x 6SL7 + 2x 6SN7? But these are
just convenience matters (except for the unobtainable 437A); as Chris
says, small signal and driver tubes are probably a done deal, at a
peak of perfection.
And, since I would have the guy pinned down, still trying to close his
mouth, I'd impress on him the necessity of a new rectifier, twin GZ37
in one tube, i.e. a complete one-tube high voltage high current bridge
rectifier). The reason I don't choose the venerated GZ34 is that the
-37 is a better rectifier in every respect and anyway its size and
shape match the 300B bulb I've already specified above. All that heat
may need the bigger tube too than the GZ34.
The reason for asking for a kilovolt tube bridge rectifier is many a
slip twixt the cup and the lip: the new tubes may turn out to be
rubbish and then we shall all want to go back to 845...
Mpfffff.... OK.... being a crude American, a heavy duty rectifier is
hard enough these days, putting two of them in the same envelope gets
quite close to the physical limitations of the beast... and three (or
more) to make the full bridge becomes needlessly complex.
As to the 417a et.al., I would take a lesson from the venerable7199, a
tube that brackets the extremes of excellence and failure, with the
general propensity towards the latter end of the spectrum. In the
words of George Santyana: "Those who do not remember the past are
condemned to repeat it." I happen to very much like the 7199, when I
have a dozen available to find two or three good ones. But for that
reason, I am staying away from it and that general philosophy with my
homebrew. I have my stock of very good 7199s, but it took me some
years and a few dozen examples to find my six working pairs and six
spares.
I would strive for refinement of existing tube designs, or as the KT90
is an evolutionary improvement to the KT88, strive in those
directions. Tube design has been "perfect" for about 40 years now in
terms of genuine and actual new designs or approaches. Packaging
differences lead to a false sense of simplicity with greater
complexity in the design with little potential for real improvement.
Much as chips offer the same sorts of pitfalls and pratfalls as
compared to discrete components... Lest anyone think I am being a
luddite, not hardly. Cell Phones have the functional equivalent of a
trainload of discrete components, no argument there. But power
amplifiers are not cell phones and overly complicated multi-function
tubes are called Compactrons last I looked.
And, we must all remember, WE was an entity of the Bell System,
concerned with reliable telephone service, movie recording and sound
systems and other broad-brush applications where reliability and
simplicity and longevity and ease of service were driving factors with
'fidelity' being an almost accidental by-product. So their tubes were
to that end and adapted to home audio systems only as latter-day
adaptations. The bandwidth of a 1950s land line was no better than the
typical single-driver horn speaker, so not much was needed of the
electronics other than reliabilty. Similarly all but a very few movie
houses used a couple of massive horn speakers (woofer & tweeter), with
not much expected there either other than within the voice spectrum.
Give me the functional equivalent of a Sylvania Mil.Spec. 5751 at all
levels of the tube spectrum, and I will be a very happy camper. In
general, I am easily pleased. It need only sound good. Looks, cost,
and yiches notwithstanding. That tube is a very fine combination of
price, function, ruggedness and longevity. Would they were all so...
THAT is my fantasy.
To each his own. This is a fantasy throughout, so we are all entitled
to our own without limit.
Peter Wieck
Wyncote, PA
|