wrote:
Just one question - why the battery?
This is the circuit under discussion:
http://members.lycos.co.uk/fiultra/T...trafi-crct.jpg
All components have sonic signature. Some have very little, or are
difficult to use for sound shaping.
There is absolutely no point in building an amp this expensive unless
for some purpose beyond the bragging rights of "I have Western Electric
300Bs in my amp, which of course I built myself." A good purpose is to
take charge of the quality of your sound, rather than leave it in the
hands of some zero-culture, long-since deaf, totally uncivilized,
supercilious, smug silicon slime, of which we can see ample samples on
these conferences. (There are also some very cultured and agreeable
silicon designers but they are successful and don't need my help.)
The WE417A driver tube was chosen for its particular signature. I had
already designed a much more precise reference SE300B amp for Western
Electric tubes with two 6SN7 stages, of which the most popular version
is he
http://members.lycos.co.uk/fiultra/T44bis-'Populaire'-crct.jpg
But in the T39 I was stepping back, building an amp for hedonists, not
soulless technicians. The 417A is very suitable for hedonists, very
linear (but not as linear as a 6SN7), quite a bit warmer in the manner
of the double digit veteran directly heated triodes but much more
widely available. Count what besides the 417A is in that circuit. The
attenuator is a DACT, built by robots on Swiss goldplated switches with
SMD resistors: zero signature, as it should be. I have no belief in
anything more than marginal soundshaping with resistors; Kiwame are
slightly but perceptibly warmer than the common Beyschlagg I also like,
and the rest leave me indifferent to the difference between them, if
any; I believe in overspeccing my resistors to run them cool and so
avoid various kinds of resistor noise which can be important in small
signal circuits.
So that leaves the tube itself, whose sound we can shape by the value
of the resistor in the plate circuit and whatever we decide to put in
the cathode circuit. Taking the plate circuit first, we can lower the
resistor value and thereby make the sound dirtier at the volume
extremes, which to the uninitiated might sound like more bass
(analogous to what you hear on boomboxes on the street or from little
passing hatchbacks owned by wannabe gangsta but of course not degraded
quite that far). That isn't quite my style, so I load the plate up to
the maximum I can within the available power supply, thereby
linearizing the response. I should explain that my style is first to
extract the maximum silence that good engineering allows, which from
tubes is much more impressive than you might imagine when you read the
silicon slime who hang out here to tell us how wrong we are because
they can't get any other employment. After that I back off to a
suitable level of hedonism. This isn't quite euphonious distortion, it
is more like a sense of balance and perspective, and an understanding
of psychoacoustics (I'm by training an economist and psychologist).
At this point we can then choose from four ways to implement a cathode
circuit. One, by constant current sink, I dismiss immediately as too
complicated for an amp announced as KISS (keep it simple, stupid); in
my next project I shall return to CCS because there their complication
is the least of the evils. That leaves three ways of doing it: a
resistor alone, a resistor bypassed by a cap, and battery bias.
Of these, the bypassed resistor is my instinctive fave. It is simple,
it is selfadjusting, and if you spend the time and the money on
development and components, you will eventually choose the right
capacitor; I have long since done my homework and know what I will use.
Open another circuit:
http://members.lycos.co.uk/fiultra/t...17acircuit.jpg
This is a complete amp built only from the first stage of the T39 (in
fact it was made by removing the 300B from a T39). It should now become
clear to you that if I substitute the battery with a resistor and a
cap), the cap becomes the sole determinant of the sonic quality of the
stage. With so little in the circuit, the cap really looms large. An
unbypassed resistor has feedback which changes the sound adversely by
making it harder, more crystalline and by tilting the response towards
the bass when in fact I want to tilt the 417A's "natural" tendency the
other way -- I just want a slightly warm amp, not a hot, gushy amp.
That leaves a battery, which, while not a soundshaping element under my
control (in that there is only one choice of operating conditions for a
417A with battery bias if you already decided the plate voltage), is at
least perfectly neutral. The battery also has a tendency to stabilize
everything around it which is a good thing as I have already paid a
heavy price in efficiency for ballasts and other devices to stablilize
important electrical points and any"free" margin is welcome.
So, by a process of elimination, I am left only with the battery. This
thought process is described in
http://members.lycos.co.uk/fiultra/T...mp%20INDEX.htm
If you study the T68bis "Minus Zero" circuit even cursorily, you will
immediately see that the other big sonic influence, besides a putative,
potential, possible cathode bypass cap, is the power supply. But that
is fully developed and fixed in a desirable sonic already, and is
anyway a large loose cannon on deck if you lose control of it, so you
don't want to mess with success if instead you can do the job by
working with one or at most two cathode circuit components, which
brings us back to the battery decision, which by its impedance in turn
makes any remaining solecism of the power supply a moot point.
All roads lead to Rome. In the T68bis you can see how all currents must
pass through that battery. It is the very dream of every control freak,
though the wannabe control freaks on RAT and UKRA lack the subtlety to
understand what is happening.
OK 2 questions... why the 4 paralleled input resistors? Noise?
The WE417A has wonderful sonics once the designer grasps how to handle
it; in the hands of the usual pretenders it quickly turns to expensive
noise because nobody told them it is a radio frequency tube. Almost all
tubes are, of course, but the 417A is especially efficient in the RF.
It has four grid pins which can pick up radio rubbish, so each one
requires a grid stopper and the signal can be put in to any of them,
though one is better than the others by far for simple reasons of
physics that may be determined by observation.
HTH.
If this is more information than you wanted, next time don't ask such
a(n only apparently) simple question!
Andre Jute
Visit Jute on Amps at
http://members.lycos.co.uk/fiultra/
"wonderfully well written and reasoned information
for the tube audio constructor"
John Broskie TubeCAD & GlassWare
"an unbelievably comprehensive web site
containing vital gems of wisdom"
Stuart Perry Hi-Fi News & Record Review