Slew Rate Current, again, for 845, frightening numbers you shouldn't read unless you are sitting down firmly
Andre Jute
down before I read more than a little of his letter.
SLEW RATE CURRENT
For our purposes, the driver tube is an AC voltage source with its plate
resistance (rp) in series and the driven tube (say the power tube) as a
capacitor connected to ground. This is a series RC circuit.
At low frequencies, little current passes because the capacitor has a
high impedance. At high frequencies, the capacitor has a low impedance
so there is some current being drawn. The driver must supply the current
through the RC circuit at high frequencies.
The slew rate equations calculate this current at the highest frequency
you intend the amp to deliver. Unless you want the high frequency notes
to clip, then your driver must provide this amount of current--in fact
preferably several times this amount of current. We thus have an
arbitrary multiple in the desired slew rate current calculations.
DEFINITIONS
Slew Rate = 2*PI*Bandwith*Vmax in V/uSecond
where
Bandwith = Highest frequency you want to pass, say 20kHz or more.
Vmax is the peak signal voltage the driver must deliver.
Cin = (A + 1) Cgp + Cgk
where
A = the gain of the driven tube
Cgp is its grid to plate capacitance
Cgk is its grid to cathode capacitance
Slew Rate Current = Cin*SR
Desired Slew Rate Current = K*Cin*SR
where K = an arbitrary constant.
Erno Borbely, Jung and the late Ron Gunzler, who taught Gordon Rankin,
all stated that they favour a constant of 5.
PUTTING IN NUMBERS
Desired Slew Rate Current is therefore 5*Cin*SR.
For a 300B,
Cin =(1 + 3.85)*15 + 9 = 81.85pf
For a fullrange amp with 80V signal voltage,
Slew Rate = 2*3.14*20000*80 = 10 048 000
therefore
Desired Slew Rate Current = 5*81.85*10 048 000
and after moving the decimal nine places we get
4.1mA for one tube and 8.2mA for two tubes
And for an 845,
Cin =(1 + 5.3)*13.5 + 6 = 91.05pf
For a fullrange amp with 150V signal voltage,
Slew Rate = 2*3.14*20000*150 = 18 840 000
therefore
Desired Slew Rate Current = 5*91.05*18 840 000
and after moving the decimal nine places we get
8.6mA for one tube and 17.2mA for two tubes. Speaking casually, "nearly
20mA". Well, we can live with that. Take out the box of 417A and pick
them over once more for low microphonics.
MOVING ONWARDS AND UPWARDS TO AUDIO NIRVANA
Okay, but 20kHz is not really good enough, or so some say. In any event,
I have SE output transformers of 5K primary impedance, rated 90W,
suitable for PSE845 among other tubes, and these trannies will go to
43kHz. To discover what current is required to overcome the input
capacitance of PSE 845 to reach 43kHz, we multiply 17.2 by 43/20 and
arrive at 37mA.
Ouch! Still, parallelling drivers will not only make the current, it
will also provide a nice low-impedance driver.
If you find the numbers frightening, try a lower constant and let us
know if the amp works. Me, I'm too old and too scarred to reinvent the
wheel when all the good guys agree it should be perfectly round. If you
still need convincing, study the Yoshino circuit or just the tube
compliment if that is all you can get, and ask yourself why Tim de
Pavaricini, who knows more than anyone on this newsgroup about tube amps
bemoaned the difficulty of reaching even audio bandwidth with an 845.
A COMPARISON OF POWER TUBE SLEW RATE CURRENT REQUIREMENTS
Given a constant of 5, to drive 300B to 100kHz requires 20mA. 845
require twice that much, at 43mA. By way of comparison the Svetlana
SV572-3/10/30/160 require respectively 28/14/45/122mA to reach 100kHz on
the same assumption of a constant of 5. The ease of driving it is one of
the things that makes the SV572-10 such a favourite.
To reach any other frequency level, multiply the milliamps above by
x/100.
HTH.
Andre
--
Andre Jute
an...@indigo.ie COMMUNICATION JUTE
--we support pages for music lovers, writers and audiophiles at
http://indigo.ie/~andre/ComJuteF1.html
George R. Gonzalez
>For a 300B,
>Cin =(1 + 3.85)*15 + 9 = 81.85pf
The plate to grid capacitance can be completely neutralized out with a
bit of positive feedback from the other side of the output transformer.
This was done all the time in old push-pull transmitters.
Is this ever done in audio circuits?
l
Gabe Velez
George R. Gonzalez wrote:
Doesn't a grounded grid amp do it? Or a cascade coupled amp? Or a
differential coupled amp? or a cathode follower? I am not sure, but it
does it for transistors.
--
Have Fun!
Gabe
Remove the *nospam* from my email address.
http://www.trigon.com
http://members.tripod.com/~gabevee
http://www.neutronicstechcorp.com
Andre Jute
>For a 300B,
>Cin =(1 + 3.85)*15 + 9 = 81.85pf
And George Gonzalez suggested:
>The plate to grid capacitance can be completely neutralized out with a
>bit of positive feedback from the other side of the output transformer.
>This was done all the time in old push-pull transmitters.
>Is this ever done in audio circuits?
Yo, George, welcome back from the radio regions. I dunno anything about
transmitters, but Positive Feedback sounds like something to investigate
if it reduces slew rate requirement by a magnitude.
Supposing for a moment Positive Feedback (PFB) doesn't sound as muddling
as NFB often does in SE amps, how would one go about calculating the
requirement and, above all, *controlling* positive feedback? How would
we implement the controls so that they are perfectly failsafe?*
*******
For those without an engineering background lurking here:
Negative feedback is a self-degenerating function: if you apply too
much, the negative feedback consumes all power and the implement shuts
itself down for lack of motive force; no harm is done (except to
biological units of course--they may go into a permanent state of
death**).
Positive feedback is self-feeding: unless controlled, it leads to more
positive feedback, an ever-increasing spiral, until the artifact it is
applied to spins out of band and self-destructs. (You can check this for
yourself if your car has a turbo or rev limiter with electronic control.
Swap the terminals of the electronic control and have your stopwatch
ready: it won't take long for the conrods to come say hello to you.
KIDS, DON'T TRY THIS AT HOME ON DADDY'S CAR UNLESS MOMMY HAS ALREADY
SAID YOU CAN COME LIVE WITH HER.)
********
Andre
No. 21 in the series "Living Dangerously" brought to you by
Gonzalez-Jute Hyper-Excitement Productions and Scrap Merchants
*Bill the Oxy Man, Jeff Rabin and the other grammarian RATs: would you
say "perfectly failsafe" contains a tautological qualification? I mean,
whatever NASA PR says, there is clearly an underlying statistical
assumption to "failsafe", a level of confidence, a case out at
10^almostunimaginable where the presumption breaks down--where, just to
give a stupid example, alien acid spittle eats through the
unobtaniumneverbreaknotoolcanmarkitnochemicalcantouchitnovibrationcansha
keitloose lead on the battery driving the Saviour Device...
**A literary allusion. Title of a novel by John Gardner, who used to
live down the road from me and now lives in Virginia. He is the present
author of the James Bond novels.
Richard Matthews
I have an engineering background (EE for 17 years), involved with
electronics/audio 30 years.
May I recommend a book for you to read: Feedback Amplifiers by Bode.
This book will help you to correct your misunderstandings about feedback,
both positive and negative.
>Yo, George, welcome back from the radio regions. I dunno anything about
>transmitters, but Positive Feedback sounds like something to investigate
>if it reduces slew rate requirement by a magnitude.
>
>Supposing for a moment Positive Feedback (PFB) doesn't sound as muddling
>as NFB often does in SE amps,
Positive feedback INCREASE distortion!
Positive feedback REDUCES bandwidth!
and many other nasty little things that are not agreeable to the repoduction
of music.
how would one go about calculating the
>requirement and, above all, *controlling* positive feedback? How would
>we implement the controls so that they are perfectly failsafe?*
>
>*******
>For those without an engineering background lurking here:
>
>Negative feedback is a self-degenerating function: if you apply too
>much, the negative feedback consumes all power
Negative feedback does not "consume" anything.
and the implement shuts
>itself down for lack of motive force; no harm is done (except to
>biological units of course--they may go into a permanent state of
>death**).
>
>Positive feedback is self-feeding: unless controlled, it leads to more
>positive feedback, an ever-increasing spiral,
Really? Did you know that oscillators are just amplifiers with "enough"
feedback of the right phase to generate there own signals? Sine wave
oscillators, for example, are hardly "out of control".
I'm not interested in any response that could start yet another flame war.
I am interested in any civilized technical discussion.
So save it.
Richard
Grover Gardner
especially in regard to phase splitters and other circuits. What is your
experience with this, Richard?
Richard Matthews wrote:
> Positive feedback INCREASE distortion!
> Positive feedback REDUCES bandwidth!
> and many other nasty little things that are not agreeable to the repoduction
> of music.
--
Grover Gardner
gro...@postoffice.att.net
Proteus
Grover wrote:
>I am given to understand that positive feedback does have some
>implementations, especially in regard to phase splitters and other
>circuits.
There are two primary real-world applications for positive feedback in tube
audio amplifiers. But first, a brief digression: feedback can be positive
or negative, meaning that it tends to make whatever you sense (or measure)
larger or smaller, respectively. Feedback also needs to be categorized
according to what you choose to measure. More on that below (pasted from a
previous post).
The two primary real-world uses of positive feedback are (1) positive
voltage feedback around the voltage amp stage(s) of a preamplifier or power
amplifier, and (2) positive current feedback to lower the output impedance
of a voltage or power amplifier.
In (1), the positive voltage feedback increases the distortion and lowers
the bandwidth of the stages around which it is applied, but this is
sometimes a useful tradeoff. For example, the distortion penalty of a bit
of positive feedback around the low-distortion front-end (voltage amp)
stage of a power amp is small, but it provides a bit more gain which can be
used to reduce the distortion of the inherently much less linear power
stage when overall negative feedback is applied. The distortion of a given
tube lineup can often be reduced by a factor of 2 or 3 this way. This
technique can and will bite back, and really can be applied effectively
only by designers who know feedback theory pretty well. Another
application is more common: The Dyna PAS preamp (like many of its vintage)
uses a resistor betwen the cathodes of the phono preamp tubes for some
positive feedback. This gives more open-loop gain so that the RIAA
feedback doesn't totally run out of steam at the low end (where closed-loop
gain is high and loop gain is low).
In (2), the output current is sensed and fed back so as to increase the
output current. Overall negative voltage feedback is also applied. This
scheme lowers the open-loop output impedance of the amplifier, and
consequently also the final, closed-loop output impedance. Positive
current feedback is also the mechanism behind the relatively low output
impedance of the totem-pole voltage amp (SRPP).
The following is some more general info from a previous post:
To get (just) a little deeper into feedback theory: Either voltage or
current can be measured, and a signal derived from this measurement can be
fed back into the amplifier either to *decrease* the measured quantity
(negative feedback) or to *increase* it (positive feedback). In audio
amplifiers the quantities measured are usually the output voltage and/or
current, but in principle they can be any voltages or currents in the
circuit.
The most common application of *global* feedback in audio amplifiers is
negative voltage feedback: the output voltage is measured, and a signal
derived from this measurement is subtracted from the input signal. This
reduces the output voltage, the distortion, and the output impedance of the
amplifier.
The most common application of *local* feedback in audio amplifiers is
cathode [emitter, source] degeneration using unbypassed resistors. This is
negative current feedback: the current through the amplifying device is
measured, and a signal derived from this measurement is subtracted from the
input signal. This lowers the signal current through the device (by
lowering the transconductance of the composite device), may lower the
distortion (by linearizing the transconductance of the composite device --
the component of the composite transconductance supplied by the resistor is
governed by Ohm's law, not by the power law that governs the device's
transconductance), and raises the output impedance of the amplifier. [I
say *may* lower the distortion, because it is the *transconductance* that
is linearized, and usually the output *voltage* we are interested in.
Whether the increased linearity of the transconductance gives increased
linearity of output voltage depends on the current-to-voltage conversion.
If this is done with a resistor (as it often is), the overall linearity
will be improved. But if it is done with a nonlinear or reactive element
it may not.]
When *output* current is measured (as with a resistor in the speaker lead),
the signal derived from this measurement can either be subtracted from the
input signal (negative current feedback, with results similar to local
degeneration), or it can be *added* to the input signal for positive
current feedback. Positive current feedback lowers the output impedance.
In principle, an amplifier's output impedance can be made as low as one
wishes -- zero, or even negative (i.e., output voltage *increases* when
more output current is drawn). In practice, careful design is required to
avoid problems with instability, operating point drift, and distortion.
Gyrators and negative-impedance-converters are typical circuits using
positive feedback. Some designers (myself included) have used positive
current feedback to achieve lower output impedances. In audio amplifiers
positive current feedback is usually used in conjunction with negative
voltage feedback to stabilize the gain and operating point and to lower
distortion.
The upper tube in a totem-pole (SRPP) stage operates with positive current
feedback. Load current is measured by the upper tube's cathode resistor,
and drives the upper tube's grid to lower the impedance at the output node
and raise the impedance at the lower tube's plate. It is this action that
makes the circuit push-pull.
Proteus (the poster formerly known as Anonymous)
Richard Matthews
experienced but I certainly haven't seen everything!). If you know of such a
circuit I'd like to see it. I avoid positive feedback a.f. amplifier
circuits that design. I believe that finding other ways around this give
more repeatable and predicatable performance. Sometimes, designers use a
"bootstrap" for drivers, for example, that inherently have positive
feedback, they then use global feedback to retain linearity and bandwidth. I
must admit that I am extremely fussy when it comes to circuits for my own
use: I insist that each section or stage is operable and testable on its
own. Using positive feedback can sometimes make that imposible. I have a
philosophical problem with "solutions" that introduce other problem that
have to be fixed
by other solutions and so on. I believe that simpler is better, without
going to the
current trend of reducing to non-function that seems to be pervalent among
hobbiest today (Slagle et al).
My bottom line is this: If you have and use a piece of equipment that uses
positive feedback, and you like the way it sounds, enjoy, listen and don't
worry. If, as an experimentor, you wish to try it, I would certainly say go
ahead learn what works for you and enjoy your experiments! My first post was
to head off the misconception that positive feedback might be a free lunch.
There are no free lunches.
A good designer with honest, realistic goals, and some modest parts can
always provide better results then a bunch of new kids on the block who
think
that they have found the audio rosetta stone and think they can design and
build better than anyone because they use positive feedback or single ended
or silver wire or whatever.
No electronic part or technology has a sound of its own, as parts are only
"audible" when they are in a system and this is why, in my post back to
Andre I DIDN'T say "positive feedback sounds bad".
My sermon is over. Off to work I go
Richard Matthews
PS sorry that I'm such a poor typist
----------
In article <3670A5E1...@postoffice.att.net>, Grover Gardner
<gro...@postoffice.att.net> wrote:
>I am given to understand that positive feedback does have some implementations,
John Byrns
<rich...@worldnet.att.net> wrote:
> I don't think that I have seen a phase splitter using pos. feedback (I'm
> experienced but I certainly haven't seen everything!). If you know of such a
> circuit I'd like to see it.
Hi Richard,
One perhaps trivial example of positive feedback in a phase splitter, is
the canonical split load phase inverter, with capacitor coupling from the
previous stage to the grid. In this case the phase splitter grid resistor
is a source of positive feedback, as it is effectively bootstrapped in
most designs that don't directly couple the grid to the previous stage.
Regards,
John Byrns
[Snip]
pja...@uriacc.uri.edu
Proteus <prot...@nym.alias.net> wrote:
> When *output* current is measured (as with a resistor in the speaker lead),
> the signal derived from this measurement can either be subtracted from the
> input signal (negative current feedback, with results similar to local
> degeneration), or it can be *added* to the input signal for positive
> current feedback. Positive current feedback lowers the output impedance.
> In principle, an amplifier's output impedance can be made as low as one
> wishes -- zero, or even negative (i.e., output voltage *increases* when
> more output current is drawn). In practice, careful design is required to
> avoid problems with instability, operating point drift, and distortion.
Hey Prot, excellent brain dump, I must say. Thanks for the words of wisdom.
I am interested in the negative output Z--how does it sound? I would assume
very tight and controlled. Does it approach a SS sound, or is there still
much of the tube character remaining?
My own personal setup consists of the tried and true "stack," and as such has
8 fairly heavily coned 12" speakers in two sets of parallel 4. This is two
channels of a 2 ohm load, and sometimes, when really stoked up, it can get a
little out of control @ Fs. Natch, a higher DF would be nice.
I would assume if the amp has an output Z of zero the DF (in theory) is
infinite?
Thanks again.
--
Ken Gilbert
Tube Guitar Amp Design/Repair Technician
The Guitarist's Choice
http://www.tgcguitar.com
-----------== Posted via Deja News, The Discussion Network ==----------
http://www.dejanews.com/ Search, Read, Discuss, or Start Your Own
Andre Jute
> There are two primary real-world applications for positive feedback in tube
> audio amplifiers.
Fabulously clear post, Proteus. Thank you!
Turns out I was already using positive feedback and loving it in SRPP
stages and just didn't know what to call it. Fancy that.
Andre
Victor Khomenko
>I don't think that I have seen a phase splitter using pos. feedback (I'm
>experienced but I certainly haven't seen everything!). If you know of such a
>circuit I'd like to see it.
It has been used for many years by us and other companies (I know of at least one
other brand doing this).
Positive feedback in general has a tremendous potential and there have been
articles in foreign publications on its use in audio amplifiers. Just think this
way: if too much of the negative feedback is bad ( a popular sentiments these
days) then less is better, zero is even better. Where do you go from here? You
cross the zero point and introduce some positive feedback. It works sort of like
the guitar body enhancing the string sound. Used in moderation works well.
Regards,
Victor.
George R. Gonzalez
Achhh... I shouldnt have used the term "positive feedback", as this is a
hot phrase.
A better way of putting it:
How about we null out the negative feedback we get from the plate-grid
capacitance by injecting some opposite-phase signal? This opposite phase
signal, if of the proper amount, will cancel out the negative feedback
effect of the grid-plate capacity, restoring the stage to have no
effective feedback.
We are not interested in using too much of this opposite-phase signal, as
that would be starting to use what amounts to positive feedback, which is
not a good thing in an amplifier.
In a push-pull output stage, you can put a pair of variable capacitors
from plate to opposite grid. If you adjust the cap to approx the
grid-plate capacitance, the opposite plate will cancel out each tube's
plate to grid feedback.
We're talking small variable caps in the 1 to 10 pf range, not the big
365pf radio sized ones.
SBench
>It has been used for many years by us and other companies (I know of at least
>one
>other brand doing this).
>
>Positive feedback in general has a tremendous potential and there have been
>articles in foreign publications on its use in audio amplifiers. Just think
>this
>way: if too much of the negative feedback is bad ( a popular sentiments
>these
>days) then less is better, zero is even better. Where do you go from here?
>You
>cross the zero point and introduce some positive feedback. It works sort of
>like
>the guitar body enhancing the string sound. Used in moderation works well.
>
>Regards,
>
>Victor.
One thing to note, Victor. One of the deficits of feedback is the introduction
of some multiplicative effects on some of the distortion products. As it turns
out, positive feedback, with or without negative feedback surrounding it
also has the same problem, so a circuit with some degree of positive
feedback will not necessarily sound like "one with no negative feedback
but moreso".
There was one assumption that was not clearly stated in this thread that
may not be obvious to all readers. Negative feedback traditionally trades
off system gain for increased bandwidth and supposedly lowered distortion.
Positive feedback, as has been noted, has the opposite effect. What was
not noted, is that positive feedback, with the feedback fraction below
unity simply modifies the amplification characteristics. It is not until the
feedback fraction hits or exceeds 1 that oscillation occurs. Bootstrapping
has been noted as a case of less-than-unity feedback fraction that does not
oscillate, but is nevertheless positive feedback, with all its inherent plusses
and minuses.
Best Regards,
Steve
Check my web page .. <A
HREF="http://members.aol.com/sbench101/">http://members.aol.com/sbench101</A>
Remove the .gov to EMail me
Grover Gardner
Now, many Lofton-White style circuits with direct-coupling use a resistor
(usually about 33K) from the cathode of the output tube to the plate supply of
the input tube. To wit:
+--------+
| |
| |
| | ------
/-+-- | // \\
/ \ | / \
| | | | |
------+ | +-----------+ |
| / | | |
\-+-- ++ | |
| Rp || \\ //
| ++ 33K \--+--
++ | +----+ |
Rk || +-----+----+-------+
++ | +---+
| | ++ |
| | Rk || |cap
| | || |
gnd | ++ |
| | |
| | |
| +---+
| |
B+ gnd
Is this an implementation of feedback, and if so, what kind?
>
--
Grover Gardner
gro...@postoffice.att.net
Ron Bales
> Fabulously clear post, Proteus. Thank you!
>
> Turns out I was already using positive feedback and loving it in SRPP
> stages and just didn't know what to call it. Fancy that.
>
> Andre
Oooh yeah! Me tooo! I do that all the time! SRPP oh yeah, that's really
hot. know what I mean, nudge nudge, s e r i o u s posiTIVE feedback.
Nothin' does it like positive feedback - used it since I was in short
pants. Of course we had different names for it then. I mean NObody used
the technical words for these things we just did what felt right.
ROn
Kurt Strain
: +--------+
: | |
: | |
: | | ------
: /-+-- | // \\
: / \ | / \
: | | | | |
: ------+ | +-----------+ |
: | / | | |
: \-+-- ++ | |
: | Rp || \\ //
: | ++ 33K \--+--
: ++ | +----+ |
: Rk || +-----+----+-------+
: ++ | +---+
: | | ++ |
: | | Rk || |cap
: | | || |
: gnd | ++ |
: | | |
: | | |
: | +---+
: | |
: B+ gnd
: Is this an implementation of feedback, and if so, what kind?
As drawn, no feedback. There's no AC at the B+ node, theoretically.
So there's no AC being fed back by the 33K resistor.
Kurt
Gabe Velez
As it stands there, there isn't even DC feedback. It is merely a
fixed bias.
I would think that Andre's analogy of the SRPP is closer, since some
of the ouput from the cathode could seem to get past the resistor and
back to the grid. Interesting that the thing doesn't oscillate since the
bottom half could look to the top as a constant current source. I guess
the fact that the grid of the bottom is not grounded but getting signal
might keep the thing stable. Or that the bottom somehow stabilizes the
top to keep it from happening. One doesn't need amplification to make an
oscilator, just positive feedback (which can be had through a
transformer [whichin itself could merely be a tapped coil to invert
phase] from the plate to grid or an RC from cathode to grid). AM and FM
radios do it similarly all the time.
My $0.02 of ignorance.
Ron Bales
You're missing the two capacitors to ground: the bypass cap on the kathode
of the output tube which would look like a short to ground for any signal
on that side of the 33K resistor - and the other cap to ground that is part
of the B+ supply and probably connected exactly where "B+" is indicated on
the schematic; it would appear as an AC short to ground on the other side
of that resistor.
DC feedback??? That's another story.
Ron
hanc...@econ.sas.upenn.edu
kst...@sr.hp.com (Kurt Strain) wrote:
>
>
> : Is this an implementation of feedback, and if so, what kind?
>
> As drawn, no feedback. There's no AC at the B+ node, theoretically.
> So there's no AC being fed back by the 33K resistor.
>
> Kurt
>
If the size of the cathode bypass cap was small enough that it didn't filter
out everything at 120 Hz, couldn't you call that a form of negative feedback?
Where the feedback is performed only on the unwanted B+ ripple. I guess even
if that was the purpose of the resistor, it wouldn't really be feedback per
se, more like active ripple compensation. What is the purpose of the extra
resistor? Power supply noise rejection is the only thing I can think of, but
that still seems a bit odd.
John
hanc...@econ.sas.upenn.edu
>
> : Is this an implementation of feedback, and if so, what kind?
>
> As drawn, no feedback. There's no AC at the B+ node, theoretically.
> So there's no AC being fed back by the 33K resistor.
>
> Kurt
>
I get it now, it's direct coupled. I missed that the first time around. Now
I'm guessing that the extra resistor is to maintain a stable operating point
for the second tube even as tube parameters change. Am I even close?
Richard Matthews
Hi John;
I not sure I know what you are saying. A split load phase inverter (ala
triode section of a 7199 in the ST70 for example) does not have any positive
feedback. The introduction of a grid resistor and coupling cap does not
change this. If you have a different topology in mind, maybe you could
provide a ref, or better still provide a drawing.
What I am posibly missing is the location of the other (non-grid) end of
the grid resistor.
Best regards
Richard
>> I don't think that I have seen a phase splitter using pos. feedback (I'm
>> experienced but I certainly haven't seen everything!). If you know of such a
>> circuit I'd like to see it.
>
John Byrns
Hi Richard,
In the type of phase splitter I am thinking of, the cathode resistor is
split into two parts, the one connected to the cathode being smaller in
value, typically a kOhm or two, with the larger one being the ground end
and 10's of kOhms in value. The non grid end of the grid resistor
connects to the junction of the two parts of the cathode resistor. This
is essentially a "bootstrap" arrangment, which I think has been mentioned
by others in this thread. A straight cathode follower has the same
connection for the grid resistor, and the same positive feedback. Maybe
this positive feedback is the reason many people find cathode followers to
have an objectionable sound?
Regards,
John Byrns
In article <74sa8e$5...@bgtnsc02.worldnet.att.net>, "Richard Matthews"
John Byrns
[Snip]
>Another
> application is more common: The Dyna PAS preamp (like many of its vintage)
> uses a resistor betwen the cathodes of the phono preamp tubes for some
> positive feedback. This gives more open-loop gain so that the RIAA
> feedback doesn't totally run out of steam at the low end (where closed-loop
> gain is high and loop gain is low).
While I agree as to the ultimate effect of the positive feedback in this
application, I tend to think of it as a money saving device, rather than a
gain increasing device. I ran through the numbers on this once, can't
remember the exact results off hand, but the gain increase was relatively
small, the same or less than you get by putting a capacitor across the
second cathode resistor in the phono preamp. If you look at the history
of some manufacturers phono preamp designs, they started with a
electrolytic capacitor across the cathode resistor of the second stage,
the first stage being unbypassed to accommodate the RIAA feedback
connection. As the years passed, they replaced the electrolytic capacitor
with a resistor between the cathodes. I have always assumed that this was
mainly a cost cutting move?
Regards,
John Byrns
[Snip]
A Dougla6
I believe there's a resistor missing, from B+ to the junction of the other
two.
The original Loftin-White schematic is in the article I wrote for VTV issue #6.
73, Alan
Grover Gardner
rejection. Given the extremely low value of capacitors in the LW amps, hum was a
real problem, as your article pointed out.
Anyway, I have tried this topology in direct-coupled circuits and it does improve
the sound quite a bit--perhaps by stabilizing the circuit...
A Dougla6 wrote:
--
Grover Gardner
gro...@postoffice.att.net
ride...@ride.ri.net
> I get it now, it's direct coupled. I missed that the first time around. Now
> I'm guessing that the extra resistor is to maintain a stable operating point
> for the second tube even as tube parameters change. Am I even close?
John, it looks to me like it's just a voltage divider from B+ to elevate the
DCV on the cathode of the output tube. Sort of a fixed bias thing, as I
think Gabe pointed out.
--
Ken Gilbert
Tube Guitar Amp Design/Repair Technician
The Guitarist's Choice http://www.tgcguitar.com
http://www.geocities.com/SunsetStrip/Garage/5701
ride...@ride.ri.net
jby...@enteract.com (John Byrns) wrote:
>
> Hi Richard,
>
> In the type of phase splitter I am thinking of, the cathode resistor is
> split into two parts, the one connected to the cathode being smaller in
> value, typically a kOhm or two, with the larger one being the ground end
> and 10's of kOhms in value. The non grid end of the grid resistor
> connects to the junction of the two parts of the cathode resistor. This
> is essentially a "bootstrap" arrangment, which I think has been mentioned
> by others in this thread. A straight cathode follower has the same
> connection for the grid resistor, and the same positive feedback. Maybe
> this positive feedback is the reason many people find cathode followers to
> have an objectionable sound?
This is indeed positive feedback, and it is this positive feedback which
makes the input impedance of the boostrapped tube much greater than Rg.
Because the "lower" end of the grid resistor is effectively coupled to a tap
on the cathode resistor, and the signal on the cathode is in phase with the
signal on the grid, there is very little signal voltage across Rg.
Personally I try to avoid all bootstrapping whenever possible. I would have
to agree with John that this may be a cause of the dissatisfaction with CF's.
The CF's I've seen with this "tapped" Rk are often referred to as "self-bias"
types. IMHO I don't see why anyone wouldn't direct couple a CF 90% of the
time. Barring that, it's simpler to add another resistor from the B+ to
provide an elevated voltage on the grid. Same effect, less orders of
complexity precisely because of the positive FB effect.
hanc...@econ.sas.upenn.edu
ride...@ride.ri.net wrote:
>
>
> > I get it now, it's direct coupled. I missed that the first time around.
Now
> > I'm guessing that the extra resistor is to maintain a stable operating point
> > for the second tube even as tube parameters change. Am I even close?
>
> John, it looks to me like it's just a voltage divider from B+ to elevate the
> DCV on the cathode of the output tube. Sort of a fixed bias thing, as I
> think Gabe pointed out.
>
Right. My idea was that the "quasi" fixed bias was somehow important because
of the direct coupling. I know tube parameters can change significantly over
the life of a tube and can differ from tube to tube. For the second tube of
a direct coupled pair, the cathode resistor is large relative to the plate
resistor, so if the DC voltage of the second tube's grid moves about, you can
get some undesirable operating points. Adding the extra resistor from B+ to
the cathode of the second tube reduces the size of the cathode resistor of
the second tube and therefore stabilizes the operating points of the tube to
a certain degree.
Thanks for your response Ken,
John
O'Connor
preamps used a little between the first two gain stages, which had an
over-all negative feed-back loop to set the desired tape/phono/mic
response. The positive feed-back was simply a resistor across the
cathode resistors (as shown in RDH4) that allows the negative feed-back
signals generated across the cthode resistors to be coupled together and
essentially "cancel" each other. This makes both cathodes look like they
are AC ground, boosting the gain of each stage. This then makes the
NFB-loop around the stages more powerful since there is more raw gain
available to sacrifice.
Dynaco also had some positive feed-back in their power amps to extend
band-wicth. Sonic Frontiers does a similar thing in their Power-series
amps.
Robert Danielak used PFB in his direct-drive EL-509 SE amp, in Glass
Audio a few months ago. A very nice circuit that minimises parts count
without minimising performance.
David Funk used positive feed-back in his guitar preamps to get high
gain without cathode by-pass caps. I've tried it out and a little PFB is
undetectable. A bit more gives more gain and too much sounds harsh. A
bit more and you have an oscillator!
Have fun
Kevin O'Connor
Proteus
>I am interested in the negative output Z--how does it sound? I would assume
>very tight and controlled. Does it approach a SS sound, or is there still
>much of the tube character remaining?
>
>I would assume if the amp has an output Z of zero the DF (in theory) is
>infinite?
I'm not very fond of trying to describe sounds of topologies, both because
the sound of a particular implementation is generally much more influenced
by the specific choices of the designer than by the basic topology and
because verbal expression just doesn't communicate the listening experience
well. As the output impedance in lowered, the resonance modes of the
driver is suppressed, to a point. As one goes negative with the output
impedance, the voice coil resistance can be cancelled (again, to a point).
It is predominantly the voice coil resistance that makes damping factors
higher than 200 or so [others believe, 20 or so] immaterial. In effect,
the negative output impedance alters the mechanical driver parameters.
Several practical points: this works best with multi-amped systems (one amp
per driver, or type of driver if several drivers are in parallel), because
the synthesized output Z is hard to tailor for several drivers at once; the
desired negative impedance for most real drivers will not be resistive --
it will be complex (have negative inductance and capacitance); and you have
to watch your phase margins, including the contributions of the load, very
carefully, because you can create resonances very easily. For all these
reasons, negative output Z is usually confined to built-in subwoofer amps.
And yes, zero output impedance produces infinite damping factor.
John Byrns notes:
>While I agree as to the ultimate effect of the positive feedback in this
>application, I tend to think of it as a money saving device, rather than a
>gain increasing device. I ran through the numbers on this once, can't
>remember the exact results off hand, but the gain increase was relatively
>small, the same or less than you get by putting a capacitor across the
>second cathode resistor in the phono preamp.
I don't know what actually motivated them, but you are right that bypassing
the second cathode is just as productive (certainly, if you partially
bypass the first cathode as well). Saving money is a very plausible reason
for them to have done it with PFB.
John wrote:
>Adding the extra resistor from B+ to the cathode of the second tube reduces
>the size of the cathode resistor of the second tube and therefore
>stabilizes the operating points of the tube to a certain degree.
Ummmm, actually, the current through the second tube will be more stable
when its cathode resistor is _larger_. Most likely, the mystery resistor
is supposed to pull the cathode up while the filaments are still cold, to
prevent full B+ from appearing between the grid and cathode of the second
tube at power-up.
Richard Matthews
>connection for the grid resistor, and the same positive feedback. Maybe
>this positive feedback is the reason many people find cathode followers to
>have an objectionable sound?
>
>
>
>John Byrns
>
Er , ah John..
Cathode followers do not have positive feedback.
Richard
John Byrns
<rich...@worldnet.att.net> wrote:
Er Richard,
Sure they do, at least the non direct coupled ones of the type I
described. Think about it a little, why is the input resistance of a
cathode follower at low frequencies many times higher than the value of
the grid resistor? There are many subtitles in tube audio design that
even the guru's sometimes miss, so don't feel bad.
Regards,
John Byrns
hanc...@econ.sas.upenn.edu
Proteus <prot...@nym.alias.net> wrote:
>
> Ummmm, actually, the current through the second tube will be more stable
> when its cathode resistor is _larger_. Most likely, the mystery resistor
> is supposed to pull the cathode up while the filaments are still cold, to
> prevent full B+ from appearing between the grid and cathode of the second
> tube at power-up.
>
OK, thanks proteus. I think I'm straight on this now. I was making this more
complicated than it actually is. Intuitively, for the operating points of
the second tube to be stable as the grid voltage changes, you want a large
change in cathode voltage from a small change in bias current and you want a
small change in anode voltage from a small change in bias current. The
"quasi" fixed bias arrangement actually "destabilizes" the operating points
since it decreases the change in anode voltage from a change in bias current.
Now that I have that intuition under my belt, it's also clear to me that
choke loading the second tube of a direct coupled pair would be desireable
since this would make the operating points more stable by reducing the dc
resistance on the anode.
Thanks again, John
Richard Matthews
John you need to do some reading. Try RDH.
But I will explain......
First consider with one end of Rg going to ground
Rin is simply RG (for low frequencies, say less than 25kHz)
If cathode followers had positive feedback the the gain would be HIGHER than
for plate coupled amplifiers with the same load and tube. The cathode
followers gain approaches unity as Rk in parellel with the load approaches
infinity. Consider the load an open circuit:
Av=mu*Rk/(Rp+(mu+1)Rk)
In cathode followers the negative feedback is SERIES voltage-feedback.
Series voltage feedback INCREASES Rin!!!!!!! (Don't confuse this situation
with that of an unbypassed cathode resistor)
Take the simple case of Rg connected right at the cathode (addmittedly a
non-practical case)
Vin' = Vin - Vout (because Vout = VRk)
Notice how the voltage is lower as Vout approaches Vin.
The input current, therefore, is not greater than Vin'/Rg
And THIS is why Rin is bigger than Rg
So John : What were you saying about "feeling bad"?
I am going to post a rigorous analysis of this on my website in a few days.
I'll post the link when I'm finished
Richard
----------
>From: jby...@enteract.com (John Byrns)
>Newsgroups: rec.audio.tubes
>Subject: Re: Can Positive Feedback (PFB) be a Good Thing in Tube Amps (was
Re: Slew Rate Current, again, for 845, frightening numbers you shouldn't
read unless you are sitting down firmly)
>Date: Sat, Dec 12, 1998, 6:27 PM
John Byrns
<rich...@worldnet.att.net> wrote:
> John you need to do some reading. Try RDH.
OK, I like to read, but I don't know if the RDH will have the answer to
this question. The RDH is singularly poor at explaining things in many
cases, it is good at giving final answers, but poor on how to get there.
> But I will explain......
>
>
> First consider with one end of Rg going to ground
>
> Rin is simply RG (for low frequencies, say less than 25kHz)
>
>
> If cathode followers had positive feedback the the gain would be HIGHER than
> for plate coupled amplifiers with the same load and tube. The cathode
> followers gain approaches unity as Rk in parellel with the load approaches
> infinity. Consider the load an open circuit:
>
>
> Av=mu*Rk/(Rp+(mu+1)Rk)
>
> In cathode followers the negative feedback is SERIES voltage-feedback.
> Series voltage feedback INCREASES Rin!!!!!!! (Don't confuse this situation
> with that of an unbypassed cathode resistor)
>
> Take the simple case of Rg connected right at the cathode (addmittedly a
> non-practical case)
>
> Vin' = Vin - Vout (because Vout = VRk)
>
> Notice how the voltage is lower as Vout approaches Vin.
>
> The input current, therefore, is not greater than Vin'/Rg
>
> And THIS is why Rin is bigger than Rg
>
> So John : What were you saying about "feeling bad"?
>
> I am going to post a rigorous analysis of this on my website in a few days.
> I'll post the link when I'm finished
I have only skimmed your comments above, but they appear to be essentially
correct, the only problem is they don't address the issue of positive
feedback, or the lack thereof. I agree that the cathode follower uses a
lot of negative feedback, but I believe that when a grid resistor is used,
there is positive feedback too. I am sure your rigorous analysis will be
correct, but will it prove that there is no positive feedback present? I
think the problem is that we first need a rigorous definition of what we
mean by positive feedback, before we can begin to prove or disprove it's
existence. I will leave it to someone else to provide a rigorous and
complete definition of positive feedback. Once we have that, I will
attempt to prove my claim that positive feedback is present in cathode
followers using a grid resistor and driven from a finite source impedance.
Regards,
John Byrns
John Byrns
<rich...@worldnet.att.net> wrote:
> John you need to do some reading. Try RDH.
>
Hi Richard,
On second thought why put it off, we might as well get right into it.
Consider the following equivalent circuit for a cathode follower.
Rg
+--/\/\/\/----+
| |
| |\ |
| | \ |
Rs | | \ |
| | \ |
Ein o---/\/\/\/---+---| +A >----+------o Eout
| /
| /
| /
|/
Where:
Ein = The equivalent generator Voltage of the preceding stage.
Rs = The source resistance of the stage driving the Cathode follower.
Rg = The cathode follower grid resistor.
A = The "gain", this is a positive value for a cathode follower stage.
Eout = The cathode follower output Voltage.
Is this not positive feedback? If not, what is it?
Regards,
John Byrns
ride...@ride.ri.net
"Richard Matthews" <rich...@worldnet.att.net> wrote:
> If cathode followers had positive feedback the the gain would be HIGHER than
> for plate coupled amplifiers with the same load and tube. The cathode
Whoah, not necessarily. Why can't they have unity gain or less WITH positive
feedback? They are not mutually exclusive.
> In cathode followers the negative feedback is SERIES voltage-feedback.
> Series voltage feedback INCREASES Rin!!!!!!! (Don't confuse this situation
> with that of an unbypassed cathode resistor)
Whoah again. The NFB present in the cathode loaded stage is CURRENT
feedback. Remember the tube is a delta V in, delta I out. Transconductance.
You need some other components in series with the tube to generate a voltage
output, like inductors or resistors.
As far as confusing this situation with that of an unbypassed Rk... how the
hell else do you get output from a CF? If the cathode is bypassed there wont
BE any output!
--
Ken Gilbert
Tube Guitar Amp Design/Repair Technician
The Guitarist's Choice http://www.tgcguitar.com
http://www.geocities.com/SunsetStrip/Garage/5701
-----------== Posted via Deja News, The Discussion Network ==----------
ride...@ride.ri.net
> attempt to prove my claim that positive feedback is present in cathode
> followers using a grid resistor and driven from a finite source impedance.
The only thing I might add to this (a point of clarification) is that not
only does it have a grid resistor, but also that it is referenced to a
tapping on the cathode load. It is fairly easy to grasp that, in this case,
as the tube turns ON harder (as the grid goes positive) the voltage at that
arbitrary tapping point on the cathode resistor will rise, due to the tube
passing more current. If that tapping point, which is the lower end of the
grid resistor, is rising in voltage, then the grid circuit will naturally be
"loaded" less by the voltage dividing action of Rg.
I know this is exactly what you're talking about, John.
Richard Matthews
Your problem is that you DO NOT HAVE THE CORRECT EQ CIRCUIT FOR A CATHODE
FOLLOWER!!!!!!
I have to say that I am tiring of this.
I will not post in response to your questions on this again, except to show
the link to my site with my analysis.
Your diagram is wrong. It is the diagram for the type of FB used with op
amps. Which by the way does not even support your view that it has pos FB
In my first post I said that I've been an EE for 17 years. I already went to
school for this.
Forgive me if I get just a little annoyed about this back and forth, from
someone the I presume is a hobbiest. What do you do for a living?
Whether you understand cathode followers or anything else does not in any
way effect my life. I will post on RAT as long as it is enjoyable. It's
getting to the point that this isn't.
Listen - then try to understand.
Richard
I appologize to anyone that feels my tone is curt.
----------
In article <jbyrns-1312...@pm3-2-43.chi-focal.enteract.com>,
jby...@enteract.com (John Byrns) wrote:
>In article <74vfn7$s...@bgtnsc02.worldnet.att.net>, "Richard Matthews"
><rich...@worldnet.att.net> wrote:
>
>> John you need to do some reading. Try RDH.
>>
>
>Hi Richard,
>
>On second thought why put it off, we might as well get right into it.
>Consider the following equivalent circuit for a cathode follower.
THIS IS NOT THE EQ CIRCUIT OF A CATHODE FOLLOWER
FIRST OF ALL THE INPUT SIGNAL OF A CATHODE FOLLOWER IS CONNECTED TO ONE SIDE
OF THE GRID RESISTOR!!!!!
John Byrns
<rich...@worldnet.att.net> wrote:
> John:
> Your problem is that you DO NOT HAVE THE CORRECT EQ CIRCUIT FOR A CATHODE
> FOLLOWER!!!!!!
Richard, please don't shout, keep your voice down, we can hear you just
fine without the shouting. OK, help me out, what is wrong with my "EQ"
circuit?
> I have to say that I am tiring of this.
> I will not post in response to your questions on this again, except to show
> the link to my site with my analysis.
Why didn't you just include the URL here, then you wouldn't have to post again.
> Your diagram is wrong. It is the diagram for the type of FB used with op
> amps. Which by the way does not even support your view that it has pos FB
Since we don't even have a working definition of Positive feedback yet, I
thought it might be helpful to draw it in the form of an op amp, which
seems perfectly legitimate to me. If my diagram "does not even support"
positive feedback, then pray tell what is positive feedback?
> In my first post I said that I've been an EE for 17 years. I already went to
> school for this.
What school did you say you went to? I would like to know in case my son
wants to study to be an engineer someday, I will know where not to send
him.
> Forgive me if I get just a little annoyed about this back and forth, from
> someone the I presume is a hobbiest. What do you do for a living?
What exactly is wrong with being a "hobbiest"? (SIC) Did God only hand
the one true word on tube circuits down to engineers? How many of the
courses you took for your engineering degree covered the operation of tube
circuits?
> Whether you understand cathode followers or anything else does not in any
> way effect my life. I will post on RAT as long as it is enjoyable. It's
> getting to the point that this isn't.
Fine, that's OK by me, I'm not going to tell you what to do, or what to enjoy.
> Listen - then try to understand.
I'm listening, but you aren't making it easy to understand, guess I should
read a few more pages from the RDH.
> Richard
>
> I appologize to anyone that feels my tone is curt.
>
>
> ----------
> In article <jbyrns-1312...@pm3-2-43.chi-focal.enteract.com>,
> jby...@enteract.com (John Byrns) wrote:
>
>
> >In article <74vfn7$s...@bgtnsc02.worldnet.att.net>, "Richard Matthews"
> ><rich...@worldnet.att.net> wrote:
> >
> >> John you need to do some reading. Try RDH.
> >>
> >
> >Hi Richard,
> >
> >On second thought why put it off, we might as well get right into it.
> >Consider the following equivalent circuit for a cathode follower.
>
> THIS IS NOT THE EQ CIRCUIT OF A CATHODE FOLLOWER
>
> FIRST OF ALL THE INPUT SIGNAL OF A CATHODE FOLLOWER IS CONNECTED TO ONE SIDE
> OF THE GRID RESISTOR!!!!!
And isn't that exactly what my circuit shows? The only thing I left out
is a coupling capacitor in the lead from Rs that connects to Rg and the
input to the amplifier block. That was because I only meant it to be an
AC equivalent circuit. True, I didn't fill in the gain block, but if you
are really the engineer you claim to be, you ought to be able to drop a
triode, and two resistors in there, and add power supply connections of
course. Now it begins to look sort of like a cathode follower doesn't it?
Regards,
John Byrns
SpeedCom
Richard,
I would like to see the analysis, but didn't see the link. Would you find it
interesting <VBG> to post that again or email it to me?
Don Carron (who does understand cathode followers)
Cheers/
Donald Carron,
Preserve endangered species; collect a complete set
Richard Matthews
The continuing saga of the misunderstood cathode follower.
I will explain.
Let
Vin = input voltage
Vin¹ = the voltage between grid and cathode
By kirchoff¹s voltage law, Vin¹+Vout=Vin
(remember for a cathode follower Vout = the voltage across Rk)
A cathode follower, can be thought of as a ³normal² amplifier with 100
percent feedback. (All of the output voltage is fed back). It¹s not too much
of a leap to see that the ³feedback signal² in a cathode follower is Vout
(The voltage across Rk)
Remember:
(a) Cathode follower voltage gain Av = mu*Rk/(rp+(mu+1)Rk)
and
(b) Av=muRp/(rp+Rp) is the voltage gain for a stage without any
feedback (plate loaded)
Now remeber...
(c) Av¹= Av/(1-AvB)
Where:
mu = amplification factor
Rp is the ³load resistor²
rp is the ac plate resistance of the tube
B = beta whichis the feedback network transfer function, which in this case
is 1, since allthe output voltage is fed back.
Now if you guys do the algebra you will see that if you put equation (b)
into equation (c) , you¹ll have two equations that have the same numeric
answer when Rp of (b) is equal to Rk of (a). In other words a plate coupled
amplifier with 100 percent feedback has the same gain as a cathode follower.
With just a little thinking this should show you that cathode followers
COULDN¹T have any positive feedback.
Please: Do the work first - argue second.
Next......
It is apparent that many do not know the definition of ³current feedback².
Below are 3 quotes and there sources:
1) Radiotron Designers Handbook, F. Langford-Smith 4th edition page 307:
³...or it (feedback) may be proportional to the current though the load,
when it is called current feedback²
2)Electronic Circuits and Tubes by Cruft
page 413:
³The variational voltage fed back may be proportional to the voltage across
the load or to the current through the load, and accordingly feedback may be
defined asvoltage controlled or current controlled...²
3)Electronic Circuits by Edwin C. Lowenberg
page 242:
.²...where the signal voltage which is proportional to the current in the
load...²
I take it that you gentlemen will agree that the sources above are more than
adequate.
All three sources are saying that with current feedback the feedback voltage
is proportional to the load.
Now consider this:
In a cathode follower, as the load is made smaller (meaning lower in
resistance) the load current increases but the feedback signal DECREASES. So
in a cathode follower the feedback voltage is INVERSELY proportional to the
load. Hence, a cathode follower, DOES NOT HAVE CURRENT FEEDBACK!
Take the case of an unbypassed cathode resistor. As the load resistance gets
smaller, the plate current (which is equal to the cathode current)
INCREASES. Since cathode current flows thru the cathode resistor the
feedback signal INCREASES, which agrees with what all three gentlemen cited
above say is current feedback.
I hope this convinces you.
If not go to school as I did 20 years ago.
Richard Matthews
----------
Andre Jute
> John:
> I have to say that I am tiring of this.
> I will not post in response to your questions on this again, except to show
> the link to my site with my analysis.
> In my first post I said that I've been an EE for 17 years. I already went to
> school for this.
> Forgive me if I get just a little annoyed about this back and forth, from
> someone the I presume is a hobbiest. What do you do for a living?
> Whether you understand cathode followers or anything else does not in any
> way effect my life. I will post on RAT as long as it is enjoyable. It's
> getting to the point that this isn't.
>
> Listen - then try to understand.
>
> Richard
>
> I appologize to anyone that feels my tone is curt.
Are you the same Richard Matthews who in his first post to me explained
that he had been an EE for all of 17 years and then added this piece of
christian good fellowship?
>I'm not interested in any response that could start yet another flame
>war. I am interested in any civilized technical discussion. So save it.
>Richard
Andre
*****
"I'm saving my virginity for an EE with manners."
"Honey, you gonna die a spinster."
*****
--
Andre Jute
an...@indigo.ie COMMUNICATION JUTE
--we support pages for music lovers, writers and audiophiles at
http://indigo.ie/~andre/ComJuteF1.html
ride...@ride.ri.net
"Richard Matthews" <rich...@worldnet.att.net> wrote:
OK Richard, here we go. Listen very closely. I will quote directly from the
RDH v4.
With a CF the load impedance is connected between cathode and earth instead
of between plate and earth in the more conventional arrangement. It is
therefore frequently referred to as "cathode loading" in distinction to the
conventional "plate loading." As a result of the 100 per cent. negaive
voltage feedback inherent in a CF, both the distortion and the output
impedance may be very low.
Obviously I was erroneous in stating the CF had current FB. It is voltage FB
due to the fact that the FB signal is not derived from a SERIES resistance (in
which the amount of FB would be proportional to the output current) but rather
from a resistance in parallel with the load.
> Let
> Vin =3D input voltage
> Vin=B9 =3D the voltage between grid and cathode
>
> By kirchoff=B9s voltage law, Vin=B9+Vout=3DVin
> (remember for a cathode follower Vout =3D the voltage across Rk)
>
> A cathode follower, can be thought of as a =B3normal=B2 amplifier with 100
> percent feedback. (All of the output voltage is fed back). It=B9s not too muc=
> h
> of a leap to see that the =B3feedback signal=B2 in a cathode follower is Vout
> (The voltage across Rk)
>
> Remember:
>
> (a) Cathode follower voltage gain Av =3D mu*Rk/(rp+(mu+1)Rk)
>
> and
>
> (b) Av=3DmuRp/(rp+Rp) is the voltage gain for a stage without any
> feedback (plate loaded)
> Now remeber...
>
> (c) Av=B9=3D Av/(1-AvB)
>
> Where:
> mu =3D amplification factor
> Rp is the =B3load resistor=B2
> rp is the ac plate resistance of the tube
> B =3D beta whichis the feedback network transfer function, which in this case
> is 1, since allthe output voltage is fed back.
>
> Now if you guys do the algebra you will see that if you put equation (b)
> into equation (c) , you=B9ll have two equations that have the same numeric
> answer when Rp of (b) is equal to Rk of (a). In other words a plate coupled
> amplifier with 100 percent feedback has the same gain as a cathode follower=
> .
> With just a little thinking this should show you that cathode followers
> COULDN=B9T have any positive feedback.
Here is where I have issue, Richard. We aren't talking about your average
CF. John and I were discussing the SPECIFIC case in which the grid resistor
is attached to a point somewhere on the Rk. On page 322 in the RDH v4,
section 7.2H, we see a variety of grid circuit arrangements. One of them,
fig 7.10, is what John and I were talking about, and I quote:
The circuit of fig. 7.10 does not apply the full degree of feedback (as in fig
7.8) unless the output resistance of the preceeding stage is very much smaller
than Rg. In other cases the feedback is reduced by a factor R0/(Rg+R0) [where
R0 is the complete plate resistance value for the preceeding stage].
[note: fig 7.8 is a plain old CF with Rk cap coupled. There is no grid
resistor.]
So Richard it would seem that we are talking about TWO different cases.
And if the degree of NFB is _decreased_ to any extent, isn't that a way of
saying some level of PFB has been applied? It is very clear in my mind that
applying an IN PHASE signal must do something to that effect.
I didn't go to school for this, and I don't want to. It really has no bearing
on what the circuit sounds like.
I don't really give a rat's ass about this pontificating anyway. I never use
my CF's like that!
Richard Matthews
I lost my temper and have already apologised to John privately for that.
I did say I was an EE for 17 years, but that does mean that I am not a
normal human being too, subject to the faults that humans are subject to.
I live, I laugh, I love, and sometimes I lose my temper too.
If I have offended, you Andre, or anyone else, I now take this opportunity
to oppologise publically.
Best regards and happy holidays!
Richard
>
----------
In article <1dk0equ.1ht...@ts01-028.arklow-new.indigo.ie>,
Bob Danielak
i have used positive current feedback with some success lately.
it was applied in the driver stage of my EL509 amp that was
featured in Glass Audio a couple issues back.
should also show up on the svetlana site (app note #33) real soon...
in this arrangement the 1st stage is direct coupled to the 2nd stage
cathode follower. the cathode follower load resistor is returned
to the cathode of the 1st stage instead of ground. the 2nd stage
cathode current flows thru the unbypassed cathode resistor of the 1st
stage. this positive current feedback increases the stage gain to
nearly the value that would be achieved by bypassing the 1st
stage cathode with a capacitor (eliminates an ugly cap).
it also reduces the output Z of the driver.
it also provides a very convenient node for the application of
global NFB.
the circuit is VERY simple. (who said this had to be complicated).
the circuit is completely DC coupled, and the topology, while
providing positive AC feedback, also provides negative DC feedback
which stabilized the operating point!!!
a total of 6 resistors and one cap (output cathode bypass)
for a three stage screen driven
EL509 amp.
i applied the same concept to a common split-load phase splitter.
dc coupled one stage to the split load circuit. returned the split-
load cathode to the cathode of the 1st stage. same effect.
increases gain without needing a bypass cap for the first cathode.
the ac balance seems to be improved, output Z is lowered,
and i think the effect of mismatched output Z between outputs
is also mitigated (although i haven't done the analysis).
Grover: you can let these guys off the hook now <grin>
and post
the ascii schematic if you still have the one i sent you.
(i trashed my SENT ITEMS folder before my vacation).
i can e-mail .gif schematics on an individual basis to those
who are interested....
take care!
bob.d.
Grover wrote:
> >I am given to understand that positive feedback does have some
> >implementations, especially in regard to phase splitters and other
> >circuits.
>
Proteus wrote:
> There are two primary real-world applications for positive feedback in tube
> audio amplifiers.
<snip>
and (2) positive current feedback to lower the output impedance
> of a voltage or power amplifier.
>
>
Grover Gardner
.
.
. +-----------------+------ B+
. | |
. Rp1 Rp2
. | |
. +-------+ +-----C1----- to grid 3
. plate | plate
. in--+--grid +-----grid
. | cath cath
. | | |
. Rg1 + ----Rk2------+-----C2-----to grid 4
. | |
. | Rk1
. | |
. +---+---------------------------
. |
. GND
.
Thanks for posting, Bob!
Bob Danielak wrote:
> i applied the same concept to a common split-load phase splitter.
> dc coupled one stage to the split load circuit. returned the split-
> load cathode to the cathode of the 1st stage. same effect.
> increases gain without needing a bypass cap for the first cathode.
> the ac balance seems to be improved, output Z is lowered,
> and i think the effect of mismatched output Z between outputs
> is also mitigated (although i haven't done the analysis).
>
> Grover: you can let these guys off the hook now <grin>
> and post
> the ascii schematic if you still have the one i sent you.
> (i trashed my SENT ITEMS folder before my vacation).
--
Grover Gardner
gro...@postoffice.att.net
Grover Gardner
up and you'll have to figure it out from Bob's text. Damn ASCII! :-)
> .
> . +-----------------+------ B+
> . | |
> . Rp1 Rp2
> . | |
> . +-------+ +-----C1----- to grid 3
> . plate | plate
> . in--+--grid +-----grid
> . | cath cath
> . | | |
> . Rg1 + ----Rk2------+-----C2-----to grid 4
> . | |
> . | Rk1
> . | |
> . +---+---------------------------
> . |
> . GND
--
Grover Gardner
gro...@postoffice.att.net
Andre Jute
[an apology]
Don't judge us by the tensions the minority of malcontents with an
agenda create with malicious deliberation. The majority of people on
this newsgroup have enormous goodwill, and many are cultured and witty
and very entertaining on diverse subjects if they get a chance to be
heard. Sometimes even electrics, and we used to mention tubes too, once
in a while...
Have a great holiday.
Andre