Help me understand Matchless Chieftain preamp?
Mandrake's Rootwerx
I've been trying to study/analyze various amp designs, and have only
encountered one that really has be a bit perplexed. ( Seel Link above
for schematic). I'm hoping one of the knowledgeable folks on this
board will be able to help further my hollow-state education.
1 Regarding the doubled-up preamp tubes: Am I correct in concluding
that this would not affect gain, but would instead increase
current-carrying capacity of the stage, and thus increase clean
headroom? Any other tonal ramifications of this approach?
2. The implementation of the tone stack seems markedly different than
virtually any other amp I've investigated, splitting the controls
between two preamp stages. What's the point/effect of doing it this
way?
3. What about the way the reverb signal is returned to the signal
chain, hooking it into the bottom leg of the phase inverter? Once
again, why do it that way, and what's the effect on tone? Am I correct
in assuming that this would preclude the inclusion of negative
feedback (and if not, how/where would it be introduced)?
Thanks for taking the time to reply.
WB
> http://www.freeinfosociety.com/electronics/schematics/audio/matchlesschieft
> 1 Regarding the doubled-up preamp tubes: Any other tonal ramifications of this approach?
parallel tube reduces noise.
>
> 2. The implementation of the tone stack seems markedly different than
> What's the point/effect of doing it this
> way?
any tone control inserts a loss in signal so you have get back the loss
some how . Fender uses 1 tube after 1 tone stack.
>
> 3. What about the way the reverb signal is returned to the signal
> chain, hooking it into the bottom leg of the phase inverter? Once
> again,
follow the signal path from input to mixer stage ..
the signal gets inverts 180 degress every tube it goes through.
The reverb and plain signal has to in phase to mix.
> Joe
> http://www.rootwerx.com
jh
WB,
whith all due respect, no.
two entirely different signals. The one direct, the other one completely
modulatet due to its travel through the reverb tank. They have no real
phase relationship anymore.
The PI is simply used as a differential amplifier. It amplifies the
difference between the two inputs. IMHO one of the smartest ways to
incorporate reverb in an amp design.
regards
Jochen
jh
WB,
sorry, my tone was not appropriate. would have liked to draw back the
posting....
regards
Jochen
Mandrake's Rootwerx
>> IMHO one of the smartest ways to
>incorporate reverb in an amp design.
Why is this approach superior to the way, say, Fender does it, with
1/2 of the tube as the reverb recovery (not sure if that's the right
term), and the second half mixing the wet/dry signals before they are
routed to the phase inverter?
What about using negative feedback? Is that precluded in this design?
Is it (NFB) something you simply wouldn't want to use in an amp design
like this?
Thanks again,
Stephen Cowell
"Mandrake's Rootwerx" <mand...@rootwerx.com> wrote
> First, thank you both for taking the time to reply.
>
>
>>> IMHO one of the smartest ways to
>>incorporate reverb in an amp design.
>
> Why is this approach superior to the way, say, Fender does it, with
> 1/2 of the tube as the reverb recovery (not sure if that's the right
> term), and the second half mixing the wet/dry signals before they are
> routed to the phase inverter?
You get to avoid using that 3.3M resistor... that thing
is a tone suck, and its bypass cap, the little 10pF'er
(or whatever) is a big microphone... the reverb channel
sound depends on that little circuit, and can sound brittle
as a result. Icepick, anyone?
> What about using negative feedback? Is that precluded in this design?
> Is it (NFB) something you simply wouldn't want to use in an amp design
> like this?
The first gain hole stage has NFB, believe it or not... without
a cathode bypass you're throwing away signal across the
cathode resistor... which appears as negative feedback. You
also lose some bass boost; this can prevent fartiness in
the bass (a Marshall trick, no bypass on the gain-hole cathode).
Amps without negative feedback in the final section sound
rawer, meatier... less hi-fi. You could add NFB to the amp
easily... NFB is normally there to make a marginal amp work,
or to get hi-fidelity response. On first look, seems like a
great amp. I'd love to play one.
__
Steve
.
WB
>
> sorry, my tone was not appropriate. would have liked to draw back the
> posting....
>
WOW .. civil accord ! In AGA ... I am humbled, and respectively, accept
your graceful observation that perhaps your tone was inappropriate , but
I must disagree whole heartily that it was , nor was it received as such.
< flame >
asdjasldjl laksdjalsdj jdlaskjdlasdj as butthead a.smdaldsas.,adm,s
barry a,sd/asd/,as/d,/../,./, obama ,,,./,./ reagan ,./,./,./,.
wacko ././.jher ffsa iran a,sm,d jerk /./,,,,,. your mama ,asl;dka;
McMilf ;askdasl; asdas asxio homo popiopal;sdkl; am,sd/ads,/ and sister
nnkkklzxzxx asd,a, solid state ,a.sd,/ad,./asdm,nmnmm,m, socialist
.a,sdm,.as bvnvmn .
< flame off >
Mandrake's Rootwerx
Joe
RS
<mand...@rootwerx.com> wrote:
>http://www.freeinfosociety.com/electronics/schematics/audio/matchlesschieftan.pdf
>
>I've been trying to study/analyze various amp designs, and have only
>encountered one that really has be a bit perplexed. ( Seel Link above
>for schematic). I'm hoping one of the knowledgeable folks on this
>board will be able to help further my hollow-state education.
>
>1 Regarding the doubled-up preamp tubes: Am I correct in concluding
>that this would not affect gain, but would instead increase
>current-carrying capacity of the stage, and thus increase clean
>headroom? Any other tonal ramifications of this approach?
Hi Joe,
Paralleling the two triodes will result in the same Mu, which is the
gauge of gain for the tube itself. Mu is the tube's transconductance
(in micromhos) multiplied by its internal plate resistance (not to be
confused with the external plate resistor). The parallel
configuration halves the effective plate resistance, but doubles the
transconductance, so the Mu itself is not substantially changed.
However, in this case, the actual gain of the stage will change as
compared to a single stage with the same plate and cathode resistors.
The formula for stage gain is:
A = (Mu * Rp) / (Rp + rp)
where A = gain, Rp = the external plate resistor, and rp = the
internal plate resistance.
If you look at that formula, you'll notice that if you did a
calculus-like convergence by increasing the value of Rp toward
infinity, that it would eventually swamp the value of rp. At that
point, rp becomes negligible, and the equation becomes:
A = (Mu * Rp) / Rp. And that reduces to A = Mu.
What that means is that it takes a very high value plate resistor to
get the stage gain close to the Mu of the tube. Or in other words, a
higher value plate resistor will increase the stage gain. Or a lower
plate resistance will increase stage gain (if transconductance/Mu
stays the same). So, if you were to build a current source (effective
infinite impedance) and use it in place of the plate resistor, you
could achieve gains close to rated Mu (100) with a 12ax7.
But with real resistor values, the gains turn out to be something
like:
Single 12ax7 with 100k plate resistor: Gain is approx 62.
Paralleled 12ax7's w 100k plate resistor. Gain is approx 77.
That's according to the formula above, and is only representative.
Re current: boost, I'm assuming that you're relating that to source
impedance, since the current drain is low. Yes, the reduced plate
resistance will provide lower source impedance, and will also result
in less loss through loading in the subsequent stage.
>2. The implementation of the tone stack seems markedly different than
>virtually any other amp I've investigated, splitting the controls
>between two preamp stages. What's the point/effect of doing it this
>way?
It allows much more control over mids, with less 'insertion loss'.The
conventional Fender-type tone control can't really do effective mid
boost. Best you can get is flat, and even that is with significant
signal loss (insertion loss).
Splitting the tone stack into three separate stages allows much more
control. The tradeoff is that you need to be careful with impedance
scaling from one stage to the next. IOW, the bass control is
contributing to the source impedance as 'seen' by the mid control. The
treble control follows the next gain stage, so it's isolated. It
would be very difficult to put all three in series without having that
intermediate buffer.
Nice concept, but they could have improved it a bit. And there are a
couple odd things, like R8 (220k) which looks like it serves no
purpose.
>3. What about the way the reverb signal is returned to the signal
>chain, hooking it into the bottom leg of the phase inverter? Once
>again, why do it that way, and what's the effect on tone? Am I correct
>in assuming that this would preclude the inclusion of negative
>feedback (and if not, how/where would it be introduced)?
The lack of nfb is an intentional design decision. They could have
still used the same node on the pi to mix the reverb and nfb. But
since it was completely open, it was the obvious and smart thing to
do.
>Thanks for taking the time to reply.
Sure thing, Joe.
RS
>The formula for stage gain is:
>
> A = (Mu * Rp) / (Rp + rp)
>
>where A = gain, Rp = the external plate resistor, and rp = the
>internal plate resistance.
>
>If you look at that formula, you'll notice that if you did a
>calculus-like convergence by increasing the value of Rp toward
>infinity, that it would eventually swamp the value of rp. At that
>point, rp becomes negligible, and the equation becomes:
>
> A = (Mu * Rp) / Rp. And that reduces to A = Mu.
>
>What that means is that it takes a very high value plate resistor to
>get the stage gain close to the Mu of the tube. Or in other words, a
>higher value plate resistor will increase the stage gain. Or a lower
>plate resistance will increase stage gain (if transconductance/Mu
>stays the same).
PS: Clarification: If plate resistance (rp) decreases and Mu stays the
same, that means that transconductance would have increased. That's
the case with the paralleled triodes.
Also, I just noticed that the cathode resistor in your circuit is not
bypassed. That means the equations get a bit more complex, but it's
the same general deal. Easier to see what's going on with the math if
it's kept simple for now.
Phil_S
"RS" <R...@sorrynospam.com> wrote in message
news:9s3gb5db22fklv7ed...@4ax.com...
I've barely finished coffee this morning, so I'm not quite ready for the
math...
I read that increasing the plate resistor increases both gain and noise.
This sets a practical limit that prevents one from milking out all of the
potential gain of the tube. IOW, you could get a 12AX7 to approach gain =
100, but it would be noisy. Is this right?
RS
It's not that tough. Just plug in book values for the tube parameters
and schematic values for the plate resistor, and see what happens.
>I read that increasing the plate resistor increases both gain and noise.
>This sets a practical limit that prevents one from milking out all of the
>potential gain of the tube. IOW, you could get a 12AX7 to approach gain =
>100, but it would be noisy. Is this right?
Increased resistance in the signal path generally does contribute to
noise. I wouldn't say that's a major reason to avoid high plate
resistor values though. Tube response tends to get a bit stiff as the
plate load approaches "ideal current source" territory.
Not to say that it can't benefit in V1a, as the signal will generally
not overload at that point. Some designers use 220k plate resistors
there. Still, you also have the source impedance presented to the
subsequent tone stack to consider.
bloomi...@gmail.com
Thanks much!