Svetlana 6D22S voltage drop
André Jute
Can anyone tell me what the centretapped secondary voltage requires to
be before a pair of Svetlana's 6D22S halfwave rectifiers if I want to
draw 1084V at 281mA after the rectifier tubes and before the pi filter.
Full wave, cap input, no info on tranny winding resistance so assume
some normal value--that's the reason I don't just run a test, that the
info is required to specify a mains tranny not yet in existence.
1. Svetlana has no curves or any other hard info.
2. Duncan Munro--are you there, Duncan?--ran tests but only up to 20V
forward voltage and with no load resistance, under which condition the
6D22S dropped about 12V at 150mA and 20V at 300mA. That's a long way
from nearly 1100V forward in a notoriously non-linear relationship.
3. Eric Barbour published a circuit (Class AB1 PP SV572-3, on the
Svetlana site) with 2x 6D22S in a full wave cap input which produces
around 670V at about 165-175mA after the rectifier from 800V centre
tapped (Hammond 278X).
Any information, help or hints will be appreciated.
Andre
--
Andre Jute an...@indigo.ie COMMUNICATION JUTE
--see our pages for music lovers, writers and audiophiles at
http://indigo.ie/~andre/ComJuteF1.html
http://www.foundmark.com/ComJute/ComJuteF1.html
Ron Bales
André Jute wrote:
>
> Can anyone tell me what the centretapped secondary voltage requires to
> be before a pair of Svetlana's 6D22S halfwave rectifiers if I want to
> draw 1084V at 281mA after the rectifier tubes and before the pi filter.
> Full wave, cap input, no info on tranny winding resistance so assume
> some normal value--that's the reason I don't just run a test, that the
> info is required to specify a mains tranny not yet in existence.
I see by your address that you are in Ireland. There's a fellow in Bandon
you might want to get in touch with, claims to have written a book on
designing vacuum tube amplifiers. Says he designs from first principles
and has licensed dozens of designs which are produced in the thousands and
to have sold dozens of his own extremely ultra-fi amps. He should be able
to figure something like this out pretty easily. That is if he really has
earned the credits he's claimed - otherwise he's just a dumbfuck lying
webwanker making up impossible stories about himself.
> Andre Jute an...@indigo.ie COMMUNICATION JUTE
> --see our pages for music lovers, writers and audiophiles at
> http://indigo.ie/~andre/ComJuteF1.html
> http://www.foundmark.com/ComJute/ComJuteF1.html
Oh yeah, that's him. Amazing fellow, time after time he says that he has
made things up as he went along that were superior to what others had taken
decades or even centuries to work out. Times must be bad though, maybe
some sort of economic downturn since he has been forced by circumstance to
form a partnership with a make-believe friend (well sure, who the fuck else
is gonna put up with his crap) and review manuscripts for money. At least
it beats his old job at a bus terminal glory hole.
So why not ask him.
ROn
fdeu...@blackrock.com
In article <1d8qucc.19f...@ts02-40.cork.indigo.ie>,
an...@indigo.ie (=?ISO-8859-1?Q?Andr=E9_Jute?=) wrote:
> Can anyone tell me what the centretapped secondary voltage requires to
> be before a pair of Svetlana's 6D22S halfwave rectifiers if I want to
> draw 1084V at 281mA after the rectifier tubes and before the pi filter.
Surely this is a job for the Master Designer at RealMcCoy Aduio -- imaginary
makers of invisible audio equipment! (Interesting set of specs, too: rather
non-standard, but then, I suppose that's a clue of a superior design; guess we
can all learn something here!)
> Full wave, cap input, no info on tranny winding resistance so assume
> some normal value--that's the reason I don't just run a test, that the
> info is required to specify a mains tranny not yet in existence.
Surely you would have a 2kV power supply or at least a power transformer
hanging around, no? What about your friend, the power design engineer -- oh,
all the lab equipment is in the shop, and you need this now? Let's see what
we can do....
> 1. Svetlana has no curves or any other hard info.
Hey, its a high vacuum diode; what more info do you need? Oh, right, you need
quite a precise set of numbers to give your tranny winder....
> 2. Duncan Munro--are you there, Duncan?--ran tests but only up to 20V
> forward voltage and with no load resistance, under which condition the
> 6D22S dropped about 12V at 150mA and 20V at 300mA. That's a long way
> from nearly 1100V forward in a notoriously non-linear relationship.
That's the problem with all this vacuum stuff; its so dang non-linear. But,
having invested years of effort learning some really hard math, and building
up some rather impressively complex models from "first principles", I am now
equipped with a spread sheet that can do the tough stuff -- so let's see if I
can help!
Looking at Duncan's numbers, and your requirements, I entered it all into my
magic spreadsheet. The spreadsheet crunch engine -- which I tied to a
monte-carlo chaos simulation engine, running on my company's CM-5 (a little
outdated, to be sure, but hey, its the backup) -- took several hours to run
these numbers -- the non-standard nature of the specs was aparently rather
taxing for the eninge -- but finally produced the answer: you need a power
tranny with a 2206.41 V center-tapped secondary, at 0.0 Ohms winding
resistance. (This shouldn't be a problem for a serious transformer winder; be
advised, though, that small deviation from these numbers will completely
invalidate the analysis -- you're on your own if your winder can't maintain a
0.01% tollerance on the winding resistance, which, after all, shouldn't be all
that hard, as even 1% of 0.0 is still just 0.0.) For this run, I did have the
quadratic resampling feature enabled, to get the run down to a reasonable
time, so while convergance was improved, these numbers might be off by a few
hundredth's of a percent; let me know if you would like me to produce a more
accurate answer.
Please also take into consideration that this analysis depends on the
assumption that you not short the B+ to ground, and that the transformer
secondary is also not shorted to the center tap; if instead, for design
reasons perhaps, you intend on shorting the powersupply in the final design
(an approach I hear RealMcCoy is infamous for), these numbers will be
substantially altered; you will need a much larger power tranny, preferably
with negative resistance. Further, I would have to caution you that you might
want to choose a different diode; the 6D22S would not be a recommended device
for such a power supply. But then again, this is an engineering perspective;
an "artistic" approach to amplifier design might dictate otherwise.
> 3. Eric Barbour published a circuit (Class AB1 PP SV572-3, on the
> Svetlana site) with 2x 6D22S in a full wave cap input which produces
> around 670V at about 165-175mA after the rectifier from 800V centre
> tapped (Hammond 278X).
Ahhh, Barbour is one of those "engineering types" -- I'm sure you can build
something much better just through application of artistic design principles,
an approach I hear RealMcCoy holds rather dear (especially as aparently noone
at RealMcCoy has any engineering knowledge)....
> Any information, help or hints will be appreciated.
OH, THE BUFFOONERY! Just when I run out of ideas for what to have my favorite
buffoon do/say next, Andre saves the day, and entertains us all on his own!
Now *THAT* is the mark of a true buffoon: the ability to originate
entertainingly stupid behavior all by one's self! Bravo, Mr. Jute!
> Andre
> --
> Andre Jute an...@indigo.ie COMMUNICATION JUTE
> --see our pages of buffoonery for people who don't know better at
> http://indigo.ie/~andre/ComJuteF1.html
> http://www.foundmark.com/ComJute/ComJuteF1.html
>
-frank
-----== Posted via Deja News, The Leader in Internet Discussion ==-----
http://www.dejanews.com/ Now offering spam-free web-based newsreading
Henry Pasternack
Ron Bales (rba...@gte.net) wrote:
: So why not ask him.
Indeed. Or maybe get a copy of the RDH and go to the section on power
supply design and figure out the relevant details. It would be necessary
to contact Svetlana to get some information on the dynamic characteristics
of the tubes in question. And some elementary graph reading and algebra
would also be required -- hardly beyond the grasp of a World Famous Irish
Amp Designer (tm).
Ah, but this would be engineering, and we all know that Engineering is
irrelevant to amplifier design.
-Henry
--
ATTENTION! Reply to h...@nortel.ca (hen...@nortel.ca won't work).
Henry Pasternack
fdeu...@blackrock.com wrote:
: > 2. Duncan Munro--are you there, Duncan?--ran tests but only up to 20V
: > forward voltage and with no load resistance, under which condition the
: > 6D22S dropped about 12V at 150mA and 20V at 300mA. That's a long way
: > from nearly 1100V forward in a notoriously non-linear relationship.
: That's the problem with all this vacuum stuff; its so dang non-linear. But,
: having invested years of effort learning some really hard math, and building
: up some rather impressively complex models from "first principles", I am now
: equipped with a spread sheet that can do the tough stuff -- so let's see if I
: can help!
Perhaps you should point out to the Master Amp Builder of Ireland
(MABI -- as in MABI he's for real and MABI he isn't) that the non-linear
relationship between voltage drop and current flow in a diode does not
depend on the absolute DC potential at which the diode operates. And
so, all other things being equal, it matters not whether we are building
a 110V power supply or an 11kV power supply.
Perhaps you should also point out to the MABI the difference between
forward voltage drop and DC output voltage.
In fact, maybe you should just erase his so-called "brain" and install
some circuit analysis software (if it will run on such out-moded hardware)
so he can estimate the answer for himself.
Come to think of it, having built twenty or thirty amps a week for the
past five years at least, you would think the MABI would have an intuitive
feel for such things. But I'm hardly the first to point out that seeming
contradiction.
I did enjoy your post, though, Frank.
Henry Pasternack
Henry Pasternack (hen...@nortel.ca) wrote:
: [Deletia]
I take it by your silence, Andre, that I am now in your "kill file"
(meaning you read what I post, but pretend not to notice)?
Sheldon D. Stokes
> Perhaps you should also point out to the MABI the difference between
> forward voltage drop and DC output voltage.
And forward voltage drop isn't really all that big a deal if you ask me.
So you loose 30 volts, so what? Your b+ is then 470 instead of 500
volts. You can adjust the operating points of your tubes slightly if you
think it makes a difference. It's the power for an amp, not a voltage
reference.
Sheldon
--
Remove SPAM_BE_GONE from my address to reply to me.
Henry Pasternack
Sheldon D. Stokes (stokes@SPAM_BE_GONEexis.net) wrote:
: And forward voltage drop isn't really all that big a deal if you ask me.
: volts. You can adjust the operating points of your tubes slightly if you
: think it makes a difference. It's the power for an amp, not a voltage
: reference.
We all lose when electrons get loose.
Consider, too, that the output of an unregulated 1100V power supply
changes 10V for every 1V change in AC line voltage. If the power company
delivers 115V in the early evening while Andre is cooking his Crubeens
and brown soda bread, but 122V at 2:00AM while he is composing chapters
for his authoritative primer on tube audio design, the B+ voltage in his
amplifier will vary by about 70V. For this reason it's just a wee bit
silly to specify as he has done a supply voltage to four significant
places.
As well, Andre should realize that the plate curves he is using to
design his amplifier are only representative of the performance he will
actually get with representative tubes. A certain amount of experiment-
ation will be required; it is much easier just to build the power
supply around given iron and then tweak the bias and dropping resistors
to get the desired performance.
Andre writes:
: Can anyone tell me what the centretapped secondary voltage requires to
: be before a pair of Svetlana's 6D22S halfwave rectifiers if I want to
: draw 1084V at 281mA after the rectifier tubes and before the pi filter.
: Full wave, cap input, no info on tranny winding resistance so assume
: some normal value--that's the reason I don't just run a test, that the
: info is required to specify a mains tranny not yet in existence.
It's odd, too, that Andre specifies the supply voltage "after the
rectifier tubes and before the pi filter". If we disconnect the pi
filter from the rectifier, of course, the voltage at this point will
be worthless pulsating DC. The pi filter is an integral part of the
power supply. It is the smoothed DC output voltage that needs to be
specified, not the magnitude of the pulsating rectified wave. This
cannot be done without setting constraints on the size of the input
capacitor and the DC resistance of the choke.
The secondary winding resistance is not the only factor to consider
in the specification of the power transformer. The primary resistance
and core losses also contribute to the effective resistance seen looking
back into the secondary. A good transformer design balances losses in
the windings (called "copper losses") and losses in the core to achieve
a given degree of regulation and allowable temperature rise at a given
current rating. While there is a range of "typical values", the design
of the transformer will depend on the required power supply performance,
any constraints on physical size, and cost considerations.
The power transformer winder should have the expertise in house to
design a unit to perform adequately in the target application. The
amp builder would normally go to the winder with a description of the
circuit, including required DC supply voltage, load current, type of
rectifier and filter, duty cycle, temperature rise, shielding, cost
limits, and so on. This will give the winder enough information to
come up with a design. A prototype will be built to verify the design,
which should come in within a few percent of the target.
I've never actually ordered a custom power supply transformer. The
information given here comes from reading the catalogs of transformer
winders that do custom work, and from my general knowledge of power
supplies and electronics manufacturing. If I can learn so much about
this process just from reading a few brochures, how is it that Andre,
OEM designer of dozens of amplifiers for well-known manufacturers,
hasn't a clue?
The mind boggles.
André Jute
Andre Jute wrote:
>Can anyone tell me what the centretapped secondary voltage requires to
> be before a pair of Svetlana's 6D22S halfwave rectifiers if I want to
> draw 1084V at 281mA after the rectifier tubes and before the pi filter.
> 1. Svetlana has no curves or any other hard info.
Henry Pasternack <hen...@nortel.ca> wrote:
> Or maybe get a copy of the RDH and go to the section on power
> supply design and figure out the relevant details. It would be necessary
> to contact Svetlana to get some information on the dynamic characteristics
> of the tubes in question. And some elementary graph reading and algebra
> would also be required -- hardly beyond the grasp of a World Famous Irish
> Amp Designer (tm).
> -Henry
Are you malicious, Henry, or merely thick? I told you in my question
that Svetlana have no "information on the dynamic characteristics", that
there are no graphs to use. Why don't you take a reading comprehension
course?
The signal to noise ratio in this thread is appalling:
SIGNAL = 0.16
Andre Jute asking a perfectly reasonable question.
NOISE = 0.84
a) Bales pressing on us for the umpteenth time the same stupidity about
amps he knows nothing about. No useful information but adds 16% to
noise.
b) Pasternack being vicious and disingenuous. No useful information but
adds 16% to noise.
c) Deutschmann spending an hour composing some lumbering "wit". No
useful information but adds 16% to noise.
d) Pasternack congratulating Deutschmann on being "witty". No useful
information but adds *another* 16% of noise.
e) Pasternack wanting to know whether I will rise to these stupid
provocations. No useful information but adds * yet another* 16% of
noise.
Wow, Henry, you're responsible for 50% of the unproductive noise in this
thread all by yourself. This attempt of yours to inhibit the free flow
of information demonstrates well why so many of us consider you a petty
disgrace to Stanford, Bell and the human race.
If you three aren't interested in anything except regurgitating old
designs, why don't you just butt out and let those of us who are
interested in doing something different get on with it?
You're so desperate for attention that you comment on your own posts:
Henry Pasternack <hen...@nortel.ca> wrote:
> Henry Pasternack (hen...@nortel.ca) wrote:
>
> : [Deletia]
>
> I take it by your silence, Andre, that I am now in your "kill file"
> (meaning you read what I post, but pretend not to notice)?
Naw, Henry, I just got bored with the predictable results of you
repeatedly sending the over-inflated balloon of your self-importance out
into the cactus field. Polemically speaking, you're a gnat. Boringly
predictable.
But that's a good idea, to put you in my killfile with the rest of the
useless Joenet scum: Bales, Deutschmann, Bubba C Bubba, Pasternack,
crows of a feather. Good riddance.
Andre
--
Andre Jute an...@indigo.ie COMMUNICATION JUTE
--see our pages for music lovers, writers and audiophiles at
http://indigo.ie/~andre/ComJuteF1.html
http://www.foundmark.com/ComJute/ComJuteF1.html
Sheldon D. Stokes
> Sheldon D. Stokes (stokes@SPAM_BE_GONEexis.net) wrote:
> : And forward voltage drop isn't really all that big a deal if you ask me.
> : So you loose 30 volts, so what? Your b+ is then 470 instead of 500
> : volts. You can adjust the operating points of your tubes slightly if you
> : think it makes a difference. It's the power for an amp, not a voltage
> : reference.
>
> We all lose when electrons get loose.
Another case of my brain getting away from my fingers. I should
proof-read my posting a little better.
> Consider, too, that the output of an unregulated 1100V power supply
> changes 10V for every 1V change in AC line voltage. If the power company
> delivers 115V in the early evening while Andre is cooking his Crubeens
> and brown soda bread, but 122V at 2:00AM while he is composing chapters
> for his authoritative primer on tube audio design, the B+ voltage in his
> amplifier will vary by about 70V. For this reason it's just a wee bit
> silly to specify as he has done a supply voltage to four significant
> places.
Maybe Irish power is accurate to 4 significant places. :)
> As well, Andre should realize that the plate curves he is using to
> design his amplifier are only representative of the performance he will
> actually get with representative tubes. A certain amount of experiment-
> ation will be required; it is much easier just to build the power
> supply around given iron and then tweak the bias and dropping resistors
> to get the desired performance.
Tell me about it, if you plot the curves on an amperex 6922 and then the
curves on a sovtek 6922, they aren't even close.
> It's odd, too, that Andre specifies the supply voltage "after the
> rectifier tubes and before the pi filter". If we disconnect the pi
> filter from the rectifier, of course, the voltage at this point will
> be worthless pulsating DC. The pi filter is an integral part of the
> power supply. It is the smoothed DC output voltage that needs to be
> specified, not the magnitude of the pulsating rectified wave. This
> cannot be done without setting constraints on the size of the input
> capacitor and the DC resistance of the choke.
You could specify an RMS voltage there, but even that wouldn't really be
all that meaningful, as you stated, due to the losses in the filter.
>
> I've never actually ordered a custom power supply transformer. The
> information given here comes from reading the catalogs of transformer
> winders that do custom work, and from my general knowledge of power
> supplies and electronics manufacturing. If I can learn so much about
> this process just from reading a few brochures, how is it that Andre,
> OEM designer of dozens of amplifiers for well-known manufacturers,
> hasn't a clue?
That's a personal issue, obviously andre doesn't know as much about this
stuff as his web page would have the reader believe.
Custom Transformer digression:
I've wound a toroid transformer for one of my DACs. Toroid corp of
maryland sells toroid kits where the primary is already wound, and you
wind the secondaries yourself. It gives you all the math you need to
calculate the number of turns and gauge for a desired voltage and
current. I wound one with a 250 volt b+ (50 mA), two 9 volt filament
windings (2A each), and two center tapped 15 volt windings for the low
voltage stuff (250 mA). It was pretty easy and came out quite nicely.
The reason I did it, is partly for fun, and to save money. an 85 VA
toroid kit only costs $27. Add about $30 for spools of secondary wire,
and you've got a transformer that is exactly what you want for like $50.
I'm going to wind another one for my pre-amp design, and I've already got
the secondary winding wire so, it will be pretty cheap. Winding the
toriod is a bit of work (particularly the high voltage winding), but it's
not very hard, and quite rewarding to have yet another aspect of your
design be yours.
Sheldon D. Stokes
In article <1d8x622.1iu...@ts02-49.cork.indigo.ie>,
an...@indigo.ie (=?ISO-8859-1?Q?Andr=E9_Jute?=) wrote:
> Andre Jute wrote:
>
> >Can anyone tell me what the centretapped secondary voltage requires to
> > be before a pair of Svetlana's 6D22S halfwave rectifiers if I want to
> > draw 1084V at 281mA after the rectifier tubes and before the pi filter.
>
> > 1. Svetlana has no curves or any other hard info.
>
>
> Henry Pasternack <hen...@nortel.ca> wrote:
>
> > Or maybe get a copy of the RDH and go to the section on power
> > supply design and figure out the relevant details. It would be necessary
> > to contact Svetlana to get some information on the dynamic characteristics
> > of the tubes in question. And some elementary graph reading and algebra
> > would also be required -- hardly beyond the grasp of a World Famous Irish
> > Amp Designer (tm).
>
> > -Henry
>
> Are you malicious, Henry, or merely thick? I told you in my question
> that Svetlana have no "information on the dynamic characteristics", that
> there are no graphs to use. Why don't you take a reading comprehension
> course?
And what Henry was alluding to in his reply is that you really don't need
the dynamic characteristics of the rectifier for an audio amplifier power
supply. The filter design will effect the output voltage as much or more
than different rectifiers could. You can estimate the voltage drop across
the rectifier to be 30 volts or so, and call it a day, I'd bet you'd be
pretty close.
low voltage supply design is similar, a rule of thumb is that the diode
bridge will drop the voltage 1.2 volts or so (from it's theoretical
rectified voltage). But I really don't care if it's 1.1 or 1.3 volts,
cause the resulting raw DC supply precise voltage doesn't matter much, if
you need a precise voltage, you'd want to regulate it anyway. There's
also the issue of different power line conditions changing the DC supply
values. That's why regulation is so key in precision circuits. But an
amplifier is not a precision circuit.
If you really want to know the characteristics because your design relies
on it, why not plot them yourself? It's easy with a regulated b+ supply
and an XY plotter. I've done this with my HP plotter and it's a fun thing
to do once.
> The signal to noise ratio in this thread is appalling:
>
> SIGNAL = 0.16
> Andre Jute asking a perfectly reasonable question.
>
> NOISE = 0.84
> a) Bales pressing on us for the umpteenth time the same stupidity about
> amps he knows nothing about. No useful information but adds 16% to
> noise.
> b) Pasternack being vicious and disingenuous. No useful information but
> adds 16% to noise.
> c) Deutschmann spending an hour composing some lumbering "wit". No
> useful information but adds 16% to noise.
> d) Pasternack congratulating Deutschmann on being "witty". No useful
> information but adds *another* 16% of noise.
> e) Pasternack wanting to know whether I will rise to these stupid
> provocations. No useful information but adds * yet another* 16% of
> noise.
Your confrontational approach to many situations is partially to blame for
that. Not that everybody else is not without blame (including myself).
> Wow, Henry, you're responsible for 50% of the unproductive noise in this
> thread all by yourself. This attempt of yours to inhibit the free flow
> of information demonstrates well why so many of us consider you a petty
> disgrace to Stanford, Bell and the human race.
See, note my above comment. You get what you ask for.
> If you three aren't interested in anything except regurgitating old
> designs, why don't you just butt out and let those of us who are
> interested in doing something different get on with it?
Measure the voltage drop across the diode and use that in your design.
IT's an easy circuit to build up. You can even hand plot the curves if
you don't have a plotter.
fdeu...@blackrock.com
In article <6j9n2a$2...@bmtlh10.bnr.ca>,
h...@nortel.ca wrote:
> I take it by your silence, Andre, that I am now in your "kill file"
> (meaning you read what I post, but pretend not to notice)?
Either that, or MABI's revving up to throw another ripping tantrum, yet
another all-out hillarious assault on how "Andre Anartist" is the World's
Foremost Amp Designer, not to be trifled with by mere "engineering-types" such
as yourself, Bombthrower Bales, and I....
I wait in gleeful anticipation for our come-uppance; its certain to be most
entertaining!
Henry Pasternack
=?ISO-8859-1?Q?Andr=E9_Jute?= (an...@indigo.ie) wrote:
: Are you malicious, Henry, or merely thick? I told you in my question
: that Svetlana have no "information on the dynamic characteristics", that
: there are no graphs to use. Why don't you take a reading comprehension
: course?
Agreed, there are no graphs on the Web site. But the data sheet
states that the 6D22S is similar to a 6DL3 damper diode, which you
certainly could look up. You could also contact Eric Barbour or
another Svetlana application engineer and ask for assistance. Or
you could make a general estimate based on your previous experience
with other low-impedance rectifier tubes of this ilk. You do have
experience, don't you?
You could also take my last suggestion, which is to find a reputable
transformer winder with the wherewithal to solve the problem for you.
As I said, any decent winder should be able to tell you, the customer,
what design suits your requirements. That's his business, after all,
and you'd be a fool to give money to a winder so ignorant that he
couldn't assist you in this regard.
: Wow, Henry, you're responsible for 50% of the unproductive noise in
: this thread all by yourself. This attempt of yours to inhibit the free
: flow of information demonstrates well why so many of us consider you
: a petty disgrace to Stanford, Bell and the human race.
I have by no means attempted to inhibit the free flow of information
on the newsgroup. In fact, it is my strong desire to inform the public
of the risks of taking bad advice from pretenders such as yourself, and
to do so by providing superior information whenever I am in a position
to do so.
Oh, yes, I do take some juvenile pleasure in rubbing your nose in
your own ignorant snot. If you can organize your imaginary throng of
detractors to raise a public (or private) outcry against me, I will
gladly back down. In the mean time, I consider my persecution of you
to be perfectly legitimate and well within the time-honored bounds
of acceptable usenet sport. Heaven knows you've done enough to bring
this upon yourself. And certainly you wouldn't play these games if
you didn't find them a sporting recreation yourself. Bear in mind
I've been around the net since before there was a net, so I have a
pretty good idea of how these things play out.
By the way, I don't work for Bell. Try again, Dr. Knowitall.
: If you three aren't interested in anything except regurgitating old
: designs, why don't you just butt out and let those of us who are
: interested in doing something different get on with it?
Well, you see, this is where I have to disagree with you. In the
field of tube amplifier design there really isn't much new under the
sun. You seem divided on this question yourself, criticising, for
instance, Mike LaFevre for building "obsolete" transformers, while at
other times touting the virtues of "classic" designs like the Quad.
The fact remains that there is nothing new or original about building
an 1100V, capacitor-input power supply with vacuum rectifiers. This
type of power supply works exactly the same today as it did in 1943,
and the proper way of designing it is, if you prefer to see it that
way, nothing but a "regurgitation" of time-tested methods.
I am perplexed by your reference to this problem as "something
different". Something different (or at least uncommon) would be to
design an amplifier with a regulated solid-state switching supply.
Please do not insult me and the newsgroup by pretending as you do
that the circuits you build are "new and different".
You share this brand of historical myopia with Herb Reichert, who
also ascribes to himself and a small band of acolytes the same false
originality you claim. It seems that what you have in common with
Reichert is a glaring lack of awareness of the electronics design
legacy, and so everything you build does, in fact, seem fresh and
new. One thing is for sure though: rehashing the same old circuits
with new parts is not innovation. Iterating through endless
combinations of passive components to achieve a desired sound is
not art.
: You're so desperate for attention that you comment on your own posts:
You've got me on the ropes now, Andre.
: Naw, Henry, I just got bored with the predictable results of you
: repeatedly sending the over-inflated balloon of your self-importance
: out into the cactus field. Polemically speaking, you're a gnat.
: Boringly predictable.
Rhetorically (as opposed to polemically) this is a weak tactic
for it invites the obvious response, "If the debate is so boring,
who do you continue it?" You are then forced to claim that there
is some higher principle at stake, such as the upholding of truth
and so on. Then we come back to the facts and I trample you yet
another time.
Better simply to admit that you find the argument stimulating
and admit that on the basis of technical competence and general
intellect (not to mention honesty and integrity), I am a serious
threat to your scheme of establishing yourself as the grand poo-bah
of tube lore on the usenet.
Actually, a correction: Andre Jute is the most serious threat
to your scheme, the way he keeps discrediting you.
: But that's a good idea, to put you in my killfile with the rest of the
: useless Joenet scum: Bales, Deutschmann, Bubba C Bubba, Pasternack,
: crows of a feather. Good riddance.
I am a free agent, Andre. And I am not a "Joenetter". I think
Bob C. is your best ally on the net, because he actually makes you
look good through his maniacal efforts to discredit you. In spite
of occasional dalliances with the "Sound Practices" mailing list, I
find it annoying and mostly avoid it like the plague. As far as I
know, I was not even subscribed during the time of your fabled and
star-crossed tenure there. I keep telling you this, but you don't
seem to listen, or you have a problem with long-term memory retention.
And anal retention, come to think of it.
Anyway, you know, and I know, that you will keep reading my postings
even while pretending not to do so if only because there is a great deal
of useful information in the things I write, but also because you cannot
rein in your curiosity (and vanity).
fdeu...@blackrock.com
In article <1d8x622.1iu...@ts02-49.cork.indigo.ie>,
an...@indigo.ie (=?ISO-8859-1?Q?Andr=E9_Jute?=) wrote:
> Are you malicious, Henry, or merely thick? I told you in my question
> that Svetlana have no "information on the dynamic characteristics", that
> there are no graphs to use. Why don't you take a reading comprehension
> course?
Jute, you are, plain and simple, a buffoon; you delight at every turn with
your predictably moronic statements, and what's more, you seem to not even
realize just how well you are playing your part!
In your first post on this subject you said:
> 2. Duncan Munro--are you there, Duncan?--ran tests but only up to 20V
> forward voltage and with no load resistance, under which condition the
> 6D22S dropped about 12V at 150mA and 20V at 300mA. That's a long way
> from nearly 1100V forward in a notoriously non-linear relationship.
This is, in fact, ALL OF THE INFORMATION THAT YOU NEED, in addition to the
most basic details which are already available on the Svetlana web site!
Here, let's spell it out for you:
Though as I pointed out in my earlier post you have chosen a wacky approach to
specifying your PS (you specify voltage/current "after the rectifier tubes and
before the pi filter"), we will still look at your design as needing 281mA at
1084V. (We'll discuss the absurdness of your spec momentarily.) Anyhow, the
rectifier tube DOES NOT SEE the 1084 V as a forward drop; instead, as Henry
pointed out, the rectifier tube only sees the current in the forward
direction. So, if we assume you are drawing 281 mA DC (absurdities of this
asside), we see that this is less than the 300mA that Duncan ran his second
test at, and therefore, the rectifier tube will drop LESS than the 20V that
Duncan found!
Notice, the "notoriously non-linear relationship" didn't even enter into this;
there isn't even really a need to interpolate Duncan's numbers to get
quasi-improved accuracy, as power line fluctuations and the non-DC nature of
this circuit will contribute far greater error.
Now, as for the absurdity of the spec: the power transformer is putting out a
large sine wave, which is being rectified and feeding a capacitor. After the
initial inrush current (you do realize that the 6D22S has a max transient peak
current limit of 1A, beyond which it will be damaged, don't you?), the cap is
charged, and the signal between the rectifier and cap is pulsating, neither DC
nor AC -- in other words, the current drawn through the rectifier tube is NOT
DC, and so the voltage drop varies. Consequently, a spec such as, "I want to
draw 1084V at 281mA after the rectifier tubes and before the pi filter" is
essentially meaningless; instead, you need to have someone who actually knows
what they are doing look at your desired output voltage and filter
type/rectifier, and make some basic calculations.
That said, anyone with even just a little experience under their belt would
have a very intuitive feel for this situation, and could readily consult a
reference table (like that found in so many readily available texts, from
modern EE texts to RDH, etc), immediately recognizing a reasonable number for
a power transformer. Any inaccuracies in the estimation are far, far
outweighed by the line voltage fluctuations, the variation in the diode (these
are vacuum diodes, after all), and variations in transformer losses: in other
words, close is good enough for this design problem; after the transformers
are in hand, the output of the PS can be trimmed with bleeder resistance, or
the exact operating point trimmed.
Of course, the above does not apply to you: you, fool, are a clueless newbie,
and you are now aparently dabbling with 1kV circuits, and we can only hope you
survive this little experience: regardless of your attrocious attitude, it
would still be a tradegy if you had to pay the ultimate price simply because
you are too damn stupid/hard-headed to do some basic reading and learning.
It is, of course, also plainly obvious to all now that your "reports" of great
success with 845's at 1kV are nothing more than lies, the fantasies of a lazy
little man who longs for greatness but who is unwilling to work for it.
Piss off, liar.
-frank
(Oh, buffoonery is just so much fun; what will Jute do next to entertain us?!)
Henry Pasternack
Sheldon D. Stokes (stokes@SPAM_BE_GONEexis.net) wrote:
: you don't have a plotter.
The problem, and to be fair it is a valid one, is that Andre wants
to specify the design of a power transformer in advance with reasonable
assurance that the resulting B+ voltage will be nearly spot-on to his
requirements the first time around. It's just a sad truth of power
supply design that you pretty much have to build it and measure it if
you really want to know what the DC output voltage will be. This is
exactly what you're saying, of course.
We ridicule Andre because he ought to know all of this by now,
given the dozens of amps he designs and builds each month. But
having been so aggressive with him, my attitude is softening.
Just to show that I hold no grudge against my Irish friend, I will
continue to instruct him in the design of his power supply. The
reason for my charity, regardless of the mistreatment I have received
from him, is that I recognize in Andre the plight of a neophyte with
no ready source of information. I feel morally bound to help beginners
whenever I can.
There is nothing short of regulating the supply that Andre can do
to stabilize it against AC line voltage fluctuations. Given that
the wall voltage may vary by 10% or more, he will just have to live
with the same percentage variation in his B+ voltage and design
the amp accordingly. By the way, I absent-mindedly referred to 115V
AC mains in a previous posting. Of course, the European electrical
grid runs on 220VAC, and I assume the same is true in Ireland. The
problem remains regardless of the value of the nominal line voltage.
In a Class A design, the average DC load current is nearly constant,
diminishing the importance of static load regulation. By varying the
value of the filter input capacitor over a range of, say, 10uF to 40uF,
Andre can tune the B+ voltage at the output of the filter. This is
because the output capacitor will tend to charge to the average value
of the rippled DC on the input cap. Making the input cap smaller
increases the magnitude of the ripple, lowering the average DC voltage.
Load regulation decreases with smaller input caps, but again, this
shouldn't be a problem in a SE triode amplifier.
In fact, the value of the filter input cap is the single biggest
determinant of the supply output voltage. It is more significant
than the diode voltage drop (which is going to be about 30V anyway),
or the secondary-referred transformer resistance. This is true up
to a point, above which the ripple voltage becomes so small that
even with large increases in capacitance the supply voltage changes
very little. It really is instructive to look at the curves
(attributed to a fellow named Schade) published in the RDH and
most tube textbooks to get an intuitive feel for how this all works.
Theory says the maximum output voltage from the supply can be no
more than 1.414 times the seconday AC voltage. In practice the
multiplication factor is unlikely to be less than 1.0. A good rule
of thumb is to estimate 1.1 for a small input cap, and perhaps
1.2-1.3 for a big one. Therefore, a 1000-0-1000V tranny would be
about as high as one would expect to go, and 850-0-850 would be
about as low. If I had only one shot at it and no other design
data, I'd go for 1800VCT @ 300mA. That's one big tranny.
I don't really like giving out these rules of thumb because we
can do better with more rigorous analysis, and there is always the
risk that people will cling to these rules as though they are iron
clad. I'd prefer to see Andre do a little homework and establish
for himself an objective and intuitive grasp of the concepts
involved. But I recognize that Andre, like most non-technical
neophytes, is intimidated by even simple circuit analysis. And
who is to blame him? This is scary stuff for those who haven't
any background in electronics. That's fine just so long as Andre
is upfront in his writings as to where he is coming from (or where
he isn't coming from, as the case may be) technically.
Tighter regulation and closer estimates of output voltage can
be had with a choke-input supply. The output voltage under load
is always close to 0.9 times the secondary voltage minus 25V or
so for the rectifier drop and another 20V for the choke. So call
it 2450VCT to put you in the ballpark with a choke-input supply.
Choke-input supplies are my favorite. They are a bit bulkier and
require a mechanically solid choke if buzzing is to be avoided.
Everyone knows by now Andre gets edgy when he hears buzzing
in the vicinity of a transformer. A decent choke such as one of
those supplied by MagneQuest shouldn't have this problem.
One has to be careful, though, to account for the tendency of
the transformer secondary voltage to run high at less than the
rated load voltage. As well, many transformers seem to be rated
for lower-than-average mains voltages (due, I suppose, to the
nasty tendency of power companies to let its lines run more than
a bit on the high side to give a safety margin against brownouts
under peak load conditision). This means the supply voltage will
likely turn out to be up to 10% higher than anticipated. This
is why, when sourcing a custom transformer, it makes sense to
choose a supplier who knows what he is doing and takes the time
to work through all the details with you.
In the end, to build it and measure it is good advice.
Paul Gustafson
Hi Frank,
I laughed out loud again !! You have a great dry sense of humor.
Can I have a copy of that spreadsheet ??
Paul G.
fdeu...@blackrock.com wrote:
> That's the problem with all this vacuum stuff; its so dang non-linear. But,
> having invested years of effort learning some really hard math, and building
> up some rather impressively complex models from "first principles", I am now
> equipped with a spread sheet that can do the tough stuff -- so let's see if I
> can help!
>
Sander deWaal
On 12 May 1998 20:18:17 GMT, hen...@nortel.ca (Henry Pasternack)
wrote the following:
<parts snipped>
> Tighter regulation and closer estimates of output voltage can
>be had with a choke-input supply. The output voltage under load
>is always close to 0.9 times the secondary voltage minus 25V or
>so for the rectifier drop and another 20V for the choke. So call
>it 2450VCT to put you in the ballpark with a choke-input supply.
>Choke-input supplies are my favorite. They are a bit bulkier and
>require a mechanically solid choke if buzzing is to be avoided.
Thanks for an excellent post, Henry.
One question though: These high voltages may cause
problems when the choke is in the B+ line.
Is it possible to have the choke in the negative, or ground circuit,
too?
And are there any drawbacks in doing this, like the danger of
cathode stripping of the rectifier?
Thanks in advance,
_
Sander deWaal
postm...@pegasus.demon.nl
www.pegasus.demon.nl
_______________________________________________
Jim S. (In Houston)
Henry Pasternack <hen...@nortel.ca> wrote in article
<6j9uc7$b...@bmtlh10.bnr.ca>...
> Sheldon D. Stokes (stokes@SPAM_BE_GONEexis.net) wrote:
> : And forward voltage drop isn't really all that big a deal if you ask
me.
><snip>
>
> The power transformer winder should have the expertise in house to
> design a unit to perform adequately in the target application. The
> amp builder would normally go to the winder with a description of the
> circuit, including required DC supply voltage, load current, type of
> rectifier and filter, duty cycle, temperature rise, shielding, cost
> limits, and so on. This will give the winder enough information to
> come up with a design. A prototype will be built to verify the design,
> which should come in within a few percent of the target.
>
> I've never actually ordered a custom power supply transformer. The
> information given here comes from reading the catalogs of transformer
> winders that do <snip>
Thanks for the insight on amps. Good post. Like most newbies on amps I soak
it up like a damp sponge.
About transformer winders. As a custom xfrmr designer and builder for over
thirty years, please don't bring be tube amp circuit drawings. Just tell me
how much power you need, your service factor, duty cycle, enclosure
requirement, minimum impedence requirements, and degree of voltage
regultion (if known), and of course, input and output voltages, and input
and output circuit configurations (single phase, pahse shifting, three
phase wye or delta, number and voltage of taps). Jim Shumberg
Jim S. (In Houston)
PS: I'm talking pure custom work here. (Meaning: If you can find what you
need on the shelf somewhere, buy it.) Minumum charge for one audio custom
transformer is about $4000.00 plus the transformer (say $125.00). That
means the first one cost you about $4125.00. The second one about $2125.00
and so on.... An order of 10 would be a total of $5250.00 or $525.00 each.
Its' sad but true...... Jim S.
Jim S. (In Houston)
What is "choke-input"? Are you talking about "input" as in in-between
transformer output and amp input (inductor type filter in other words)? I
guess I'm having truble understanding where you are talking about placing a
choke. Thanks, Jim S.
Sander deWaal <au...@pegasus.demon.nl> wrote in article
<3559c054....@news.demon.nl>...
> On 12 May 1998 20:18:17 GMT, hen...@nortel.ca (Henry Pasternack)
> wrote the following:
>
> <parts snipped>
>
> > Tighter regulation and closer estimates of output voltage can
> >be had with a choke-input supply. The output voltage under load
> >is always close to 0.9 times the <snip>
André Jute
Thanks, Sheldon. Information, at last.
Various snips, two letters conflated:
Sheldon D. Stokes <stokes@SPAM_BE_GONEexis.net> wrote:
>The filter design will effect the output voltage as much or more
> than different rectifiers could.
My 1084V allows for the filters that follow. Once you get in the kV
range, if you want to KISS your choice of filter components is
drastically reduced, so you choose them first and then worry about the
mains tranny. You just about have to get the mains trx on the mark first
time. Looking for just the right 1600-2000V cap to adjust the B+ by
exactly the right voltage can cause severe hair loss.
>You can estimate the voltage drop across
> the rectifier to be 30 volts or so, and call it a day, I'd bet you'd be
> pretty close.
Unfortunately, it's not so close. To get an idea of rectifier drops at
really high voltages (for audio, the hams wouldn't necessarily agree), I
once spent a day swapping rectifiers in an 845 amp and the variations
were surprising. Chelmer also ran some tests for me with a variety of
rectifiers on a lower-voltage amp; result ditto. You don't actually have
to do tests; there's a bunch of graphs in the RDH4, p1171ff, which will
tell you at a glance that a rectifier isn't a generic device. Some of
the underlying assumptions described in the accompanying text complicate
matters further.
> if
> you need a precise voltage, you'd want to regulate it anyway.
I build SE linestage amps. The few times I tried it, regulated supplies
gave me a whited-out studio type sound. On the other hand, I have never
heard a phono stage without regulation that sounded as good as one with
regulation. I conclude that the smaller your signal, the more aurally
important regulation becomes. In HV amps, regulation can glare at you
rather nastily. (In my experience anyway. I notice that Mattijs de
Vries, who sometimes comes here, is big on regulation. Presumably he
likes the sound or maybe he designs better regulators than I do.)
> There's
> also the issue of different power line conditions changing the DC supply
> values. That's why regulation is so key in precision circuits. But an
> amplifier is not a precision circuit.
And from your other letter:
>Maybe Irish power is accurate to 4 significant places. :)
Theoretically, by law, I am told, accurate to within 5V either side of
230V.
>
> If you really want to know the characteristics because your design relies
> on it, why not plot them yourself? It's easy with a regulated b+ supply
> and an XY plotter. I've done this with my HP plotter and it's a fun thing
> to do once.
I thought about this. But I already have a 20-30V general drop guideline
from the factory, an example 670V design even if at about half the
current, and Duncan's low-voltage results, so I was trying to avoid
drawing curves with kilovolt lashups.
Also from your other letter, re tranny specification:
>That's a personal issue, obviously andre doesn't know as much about
this stuff as his web page would have the reader believe.
Never claimed to know about tranny innards. Don't intend to become an
expert now when real experts are plentyful and I can never know as much
as they already do. My web page specifically hotlinks the more than
casually interested reader to no fewer than three winders of proven
reliability. (It's a fallacy that anyone can know everything about
anything, or even needs to. All that is required is to know what you
don't know--and to know the right people who do know.) Before I've
always done the smart thing and simply shown my amp and filter design to
the winder, told him how closely it must be regulated (i.e. the
acceptable drop from full load to half load) and let him spec the mains
tranny. But my present design was iffy if the voltage dropped by too
much, so instead of passing the buck I thought I would take a closer
interest in the mains tranny and maybe learn a little something.
But I was so pissed off with the dumb interference here for two days
that I simply designed a regulated supply and, in case that too sounds
whited-out like previous regulated HV supplies, in addition redesigned
the amp to work without regulation even if the mains voltage varies 20
per cent, not as big a job as it sounds, fortunately. Thus the 6D22S
voltage drop and the exact B+ with virtually any filter components or
mains input variation becomes irrelevant, which I take to be your
general subtext.
Oh well, the curious may propose but as always the god of conservative
engineering disposes. The redesign for 20 per cent variation was an
extremely interesting exercise, so there is a net gain in the pleasure
of achievement, especially since the amp will now work reliably in any
of the countries for which the Toroid Corp lists mains voltages. From
adversity, profit.
*****
I read with interest your description of rolling your own tranny. Hmm.
An old ham showed me how to take an inexpensive standard RS (European
equivalent to Mouser or Digikey, brands some stuff with their own label,
usually saves a few bucks over a brandname made in exactly the same
factory to exactly the same spec) tranny and *unwind* some turns to get
precisely the voltage required for e.g. a filament. Cheaper than buying
the kit, the wire and so on. Faster too if probably not as satisfying as
rolling your own. The question is, How, without a machine, do you keep
the tension constant when you wind from scratch?
*****
> > The signal to noise ratio in this thread is appalling:
etc.
>
> Your confrontational approach to many situations is partially to blame for
> that. Not that everybody else is not without blame (including myself).
I don't write to these people first, I don't contribute to threads they
start, and I don't try to influence others to deny them information.
They practice no such restraint. If they continually put themselves in
my face, they, and others, mustn't blame me when they get clipped in the
ear for their impertinence. But I think we can leave that there. Since
in the six months I have been here they've proven to be of no technical
use to me--quite the contrary, and are too dull to be entertaining, I no
longer read their posts.
*****
Thanks again, Sheldon.
SBench
Sander asks...
>On 12 May 1998 20:18:17 GMT, hen...@nortel.ca (Henry Pasternack)
>wrote the following:
>
><parts snipped>
>
>> Tighter regulation and closer estimates of output voltage can
>>be had with a choke-input supply. The output voltage under load
>>so for the rectifier drop and another 20V for the choke. So call
>>it 2450VCT to put you in the ballpark with a choke-input supply.
>>Choke-input supplies are my favorite. They are a bit bulkier and
>>require a mechanically solid choke if buzzing is to be avoided.
>
>Thanks for an excellent post, Henry.
>One question though: These high voltages may cause
>problems when the choke is in the B+ line.
>Is it possible to have the choke in the negative, or ground circuit,
>too?
>And are there any drawbacks in doing this, like the danger of
>cathode stripping of the rectifier?
>Thanks in advance,
>
>_
>Sander deWaal
The advantage of putting the choke in the negative side is
lower voltage on the choke, potentially helping any breakdown
issues. The circuit is still pretty much the same, cap input
if you return the input cap to the CT/choke junction, choke input
if you return it to ground.
There are some potential capacitance issues that *usually*
are not important.
The disadvantage? Potentially electrically hazardous if the
choke opens, as there will be voltage where you don't
expect it.
Had no problem with an 813 running 2500 volts and the choke
on the "low" side. Worked just fine.
Best Regards,
Steve
Adam Stouffer
fdeu...@blackrock.com wrote:
>
> In article <1d8x622.1iu...@ts02-49.cork.indigo.ie>,
> an...@indigo.ie (=?ISO-8859-1?Q?Andr=E9_Jute?=) wrote:
> > Are you malicious, Henry, or merely thick? I told you in my question
> > that Svetlana have no "information on the dynamic characteristics", that
> > there are no graphs to use. Why don't you take a reading comprehension
> > course?
>
> Jute, you are, plain and simple, a buffoon; you delight at every turn with
> your predictably moronic statements, and what's more, you seem to not even
> realize just how well you are playing your part!
>
> In your first post on this subject you said:
> > 2. Duncan Munro--are you there, Duncan?--ran tests but only up to 20V
> > forward voltage and with no load resistance, under which condition the
> > 6D22S dropped about 12V at 150mA and 20V at 300mA. That's a long way
> > from nearly 1100V forward in a notoriously non-linear relationship.
>
> This is, in fact, ALL OF THE INFORMATION THAT YOU NEED,
Lets see you figure it out, go head I'm waiting.
>
> That said, anyone with even just a little experience under their belt would
> have a very intuitive feel for this situation, and could readily consult a
> reference table (like that found in so many readily available texts, from
> modern EE texts to RDH, etc),
if you can't answer this question then dont fukin post
>
> Of course, the above does not apply to you: you, fool, are a clueless newbie,
you never asked a question?
>
> Piss off, liar.
>
> -frank
> (Oh, buffoonery is just so much fun; what will Jute do next to entertain us?!)
>
eat me hoebag
--
-----------------------------------------
| Public enemy #10 |
| Remove the NO-SPAM |
-----------------------------------------
Kirk R. Patton
That subject line just doesn't doo (sic) justace to this thread . . .
Man, some of you guys could really stand to get into a few more
barfights . . . really, the cost of a heated entanglement on Usenet is
such a low price that people will get going over anything . . . If every
time one of these things cropped up all participants woke up with
bruises, black eyes, broken ribs, hangovers . . . not to mention waking
up in a jail cell. . . maybe it would educate them on when the point at
hand is WORTH arguing about . . .
Or maybe somebody needs to develop a few gigs of bloatware to
systematically compile a specific individual's postings from Usenet,
analyze the entirety for any improprities of logic or deviations from
the Sacred Will of all Engineering and Audio muses, and emulate the
style of eighteenth-century poets in hourly postings reminding the
billions of readers of R/A/T of the ultimate Universal placement of the
individual in question.
. . . . it's nice to get a bit heated up about somebody spewing some
drivel on the net, it provokes thought and clarifies ideas, and compells
their expression and communication. Pasternak on Psychoacoustics was an
example of this. But what's with all this other shit, guys? Why is
there this extraenous fallic-mass insecurity and inferiority crap mixed
in? Yes, even in the same post!
You insult the intelligence of all the readers of this newsgroup by
constantly reminding us that you don't like Andre, that you feel he's
full of shit. I'm sick of having to sort through post after post of
what really is the same thing, over and over . . . or "killfile" people
and miss the shining rare gem of an interesting post.
I am more than intelligent enough to form my own opinion of the regular
posters of this NG . . . and not so insecure about my ability to do so
that I have to put it out for everybody to see (because it's really
nobody's goddamn business) . . . or jump onto this whole "with me or
against me" horseshit.
Listen guys, if your tongues spew forth in the same manner as your
keyboards, watch it, because you won't have the benefit of an ocean
separating you from the guy you may make a habit of verbally insulting
on a regular basis . . . and if it was me caught slightly drunk and
irate at the end of a hard week, I'd have been happy to haul your sorry
ass outside and break a few ribs by now -- but I guess that's the only
gesture some people can understand . . . .
Best Regards,
Kirk Patton
har...@q3-consulting.com
In article <355913...@NO-SPAM.sgi.net>,
Adam Stouffer <te...@NO-SPAM.sgi.net> wrote:
> > In your first post on this subject you said:
> > > 2. Duncan Munro--are you there, Duncan?--ran tests but only up to 20V
> > > forward voltage and with no load resistance, under which condition the
> > > 6D22S dropped about 12V at 150mA and 20V at 300mA. That's a long way
> > > from nearly 1100V forward in a notoriously non-linear relationship.
> >
> > This is, in fact, ALL OF THE INFORMATION THAT YOU NEED,
>
> Lets see you figure it out, go head I'm waiting.
Easy-peasy. Diodes follow Child's Law to the letter (as it were), at
least at the kind of currents we are talking here. (It gets iffy
below 1mA). I assume everyone here is familiar with (IAEHIFW) Child's Law?
You really only need one data point, although more is better. You
calaculate the perveance of the tube (IAEHIFW perveance?) by
K=I/(V^1.5) for any given measured I and V. This shows that for this
tube K is 3.35-3.61, depending on which point you believe. Any other
point can be worked out down to 1mA or so and up to the point at which
temperature saturation becomes an issue, 5-10 times the rated max
current typically.
I've resisted joining this slanging match, but now that I'm here...
by far the best way to answer this kind of question is with Spice...
Microsim PSpice is FREE for the asking at their website. Whip up a quick
model for the diode (probably Duncan already has one) and feed in the
other circuit parameters including resistance and anything else you
want, and look at the answers. MUCH easier (and safer and cheaper) than
trial and error with the real thing; not a substitute for doing the
real thing in the end of course, but a pain-free way to eliminate
all sorts of things that DON'T work.
John
har...@q3-consulting.com
In article <35592D...@sprintmail.com>,
kirkp...@sprintmail.com wrote:
>
> That subject line just doesn't doo (sic) justace to this thread . . .
>
> Man, some of you guys could really stand to get into a few more
> barfights . . . really, the cost of a heated entanglement on Usenet is
> such a low price that people will get going over anything . . . If every
> time one of these things cropped up all participants woke up with
> bruises, black eyes, broken ribs, hangovers . . . not to mention waking
> up in a jail cell. . . maybe it would educate them on when the point at
> hand is WORTH arguing about . . .
>
HEAR HEAR! What I find truly amazing is that under all the childish
jibe-throwing (by all parties; I take no sides, although this doesn't
stop me having opinions that I keep to myself) there is an interesting
and insightful technical discussion going on. Maybe some people are
just Usenet-masochists, and can only get excited if they are being
beaten.
Ned Carlson
On 12 May 1998 14:41:14 GMT, hen...@nortel.ca (Henry Pasternack)
wrote:
>Henry Pasternack (hen...@nortel.ca) wrote:
>
>: [Deletia]
>
> I take it by your silence, Andre, that I am now in your "kill file"
>(meaning you read what I post, but pretend not to notice)?
One may wish to check out the following URL:
http://www.geocities.com/Nashville/Opry/3904
to see a truly sarcastic web page. Someone may take the
same pity on Andre as Dr Stereo, one never knows.
Ned Carlson Triode Electronics,2225 W Roscoe Chicago, IL, 60618 USA
ph 773-871-7459 fax 773-871-7938
12:30 to 8 PM CT, (1830-0200 UTC) 12:30-5 Sat, Closed Wed & Sun
http://www.triodeel.com
Text file catalogs:Catalog 'Bot at cat...@triodeel.com
Sander deWaal
On 12 May 98 22:02:19 GMT, "Jim S. (In Houston)" <j...@drivesys.com>
wrote the following:
>What is "choke-input"? Are you talking about "input" as in in-between
>transformer output and amp input (inductor type filter in other words)? I
>guess I'm having truble understanding where you are talking about placing a
>choke. Thanks, Jim S.
Choke input means, that the cathode of the rectifier goes directly
into a choke (often called "swinging choke"), and after that, a cap
to ground.
This approach gives some voltage regulation with a varying load,
like a class AB amp.
As opposed to a capacitor input filter, where the cathode is
followed by a cap to ground, and then to a filter choke
or resistor.
This cap can't be too big, like 50 uF max, and for some
tubes even less.
Too much capacity at the cathode can lead to cathode
stripping, because of the high current, drawn by empty
caps, and higher charge currents.
Like Cary does in some models, a CV378 (GZ37) is followed
by 1200 uF (with the optional powerbank even a tenfold of that).
Good example of bad engineering.
Choke-input filters give us some voltage loss, and the
choke has to be insulated really good from the core, to avoid
breakdown.
That's why I asked if placing the choke in the ground circuit,
would make any difference re: capacitor load at the cathode.
har...@q3-consulting.com
In article <355b86fb....@news.demon.nl>,
au...@pegasus.demon.nl (Sander deWaal) wrote:
> Choke input means, that the cathode of the rectifier goes directly
> into a choke (often called "swinging choke"), and after that, a cap
> to ground.
In fact, a swinging choke is a type of choke rather than a way to use
it. It is a choke with a very small air gap, such that the inductance
falls as the DC current rises (I think I have this the right way round).
It avoids the problem with choke-input filtering that under no load,
the choke value isn't high enough and it becomes in effect a cap-input
filter, with B+ rising up to 1.4 Vrms instead of 0.7 Vrms.
If you don't use one of these (which according to the Morgan Jones
book aren't made any more, although you could probably get someone
to build them for you), you either have to arrange for the current never to
drop too low (bleed resistor) or over-value the choke, or some combination
of the two.
Sheldon D. Stokes
In article <01bd7dd5$76700200$0203a8c0@jenny>, "Jim S. (In Houston)"
<j...@drivesys.com> wrote:
Toroid corp of maryland does custom design work on smallish toroids for
like $250 if memory serves me right. But those are just power
transformers and are like a lay-up in the world of transformers. I'd
expect a custom wide bandwidth transformer design to cost you much more.
Sheldon D. Stokes
In article <1d8xpgl.17p...@ts01-06.cork.indigo.ie>,
an...@indigo.ie (=?ISO-8859-1?Q?Andr=E9_Jute?=) wrote:
> Looking for just the right 1600-2000V cap to adjust the B+ by
> exactly the right voltage can cause severe hair loss.
For voltages over 450 volts I have always ended up putting caps in series
with bleeder resisitors. There is a small range of high votlage caps, but
they are expensive.
> I build SE linestage amps. The few times I tried it, regulated supplies
> gave me a whited-out studio type sound. On the other hand, I have never
> heard a phono stage without regulation that sounded as good as one with
> regulation. I conclude that the smaller your signal, the more aurally
> important regulation becomes. In HV amps, regulation can glare at you
> rather nastily. (In my experience anyway. I notice that Mattijs de
> Vries, who sometimes comes here, is big on regulation. Presumably he
> likes the sound or maybe he designs better regulators than I do.)
For a single ended amp, the power supply is more critical than a push pull
one (do the large PSRR of PP). Regulator design is an art and science
like everything electronic. (don't read this wrong, science is critical)
> Theoretically, by law, I am told, accurate to within 5V either side of
> 230V.
That's like 2% variation. So for a 1 KV supply, that is a change of +-20
volts or so. Not an issue for any audio cicuits I know about.
> I thought about this. But I already have a 20-30V general drop guideline
> from the factory, an example 670V design even if at about half the
> current, and Duncan's low-voltage results, so I was trying to avoid
> drawing curves with kilovolt lashups.
100 volts will kill you just as quickly if you aren't careful. high
voltages shouldn't be toyed with. I'm really lucky that I'm still here.
> I read with interest your description of rolling your own tranny. Hmm.
> An old ham showed me how to take an inexpensive standard RS (European
> equivalent to Mouser or Digikey, brands some stuff with their own label,
> usually saves a few bucks over a brandname made in exactly the same
> factory to exactly the same spec) tranny and *unwind* some turns to get
> precisely the voltage required for e.g. a filament. Cheaper than buying
> the kit, the wire and so on. Faster too if probably not as satisfying as
> rolling your own. The question is, How, without a machine, do you keep
> the tension constant when you wind from scratch?
I've done that as well. Some cheap transformers are wound on two bobbins,
one for primary, and one for secondary. That makes it super easy to wind
whatever you want for a secondary.
The transformers are toroids (doughnuts), and I pull each winding tight by
hand. They suggest that you wind the largest guages closest to the core
and wind the finer guages on top (if oyu want multiple windings). I
beleive that this methode tends to hold the larger wire in place with the
smaller wire.
My DAC transformer has some heavy guage windings for the heaters first,
then some medium windings for the low voltage analog and digital supplies
next, then a ton of very fine 32 guage windings for the b+. The 32 guage
wire totally covers the others. With the final wrap, the transformer is
dead quiet and has quite good regulation. A friend of mine did this first
(and prompted me to try it), he's supposedly working on a description and
tests of the resultion transformer. I told him I'd put it on my web site
when he's done. He reported that he got 2% regulation from 20% to 100%
load on his windings if memory serves.
Henry Pasternack
Jim S. (In Houston) (j...@drivesys.com) wrote:
: As a custom xfrmr designer and builder for over thirty years, please don't
: bring be tube amp circuit drawings.
Thanks. When I mentioned "the circuit", I really meant just the info
relevant to power supply specifications. It doesn't really matter to the
guy building the tranny what kind of driver topology you're using...
Henry Pasternack
=?ISO-8859-1?Q?Andr=E9_Jute?= (an...@indigo.ie) wrote:
: My 1084V allows for the filters that follow. Once you get in the kV
: range, if you want to KISS your choice of filter components is
: drastically reduced, so you choose them first and then worry about the
: mains tranny.
What's so hard about three 500V caps in series?
: You don't actually have to do tests; there's a bunch of graphs in the
: RDH4, p1171ff, which will tell you at a glance that a rectifier isn't
: a generic device. Some of the underlying assumptions described in the
: accompanying text complicate matters further.
I would appreciate it if you would refrain in the future for abusing
me for reading RDH. The answer to your comment is this: Rectifiers come
in two basic flavors, low-impedance and high-impedance. Because of the
pulsatile nature of the capacitor charging current, the effect of the
series impedance of the rectifier is greater than would be expected from
simply applying Ohm's law to the DC load current. If you stick to tubes
from the low-impedance class of rectifiers (which I believe is what the
6D22S is), you will get basically the same results and the aforementioned
30V drop.
: I build SE linestage amps. The few times I tried it, regulated supplies
: gave me a whited-out studio type sound.
No one said you have to like regulation. I mention it because it's
the only way to stabilize the B+ voltage against AC mains variations.
: Never claimed to know about tranny innards. Don't intend to become an
: expert now when real experts are plentyful and I can never know as much
: as they already do.
Well thank you very much. If you would then refrain from condemning
trained experts for what they know, we will all get along very much
better.
: Oh well, the curious may propose but as always the god of conservative
: engineering disposes. The redesign for 20 per cent variation was an
: extremely interesting exercise, so there is a net gain in the pleasure
: of achievement, especially since the amp will now work reliably in any
: of the countries for which the Toroid Corp lists mains voltages. From
: adversity, profit.
You're welcome, Andre.
Henry Pasternack
Kirk R. Patton (kirkp...@sprintmail.com) wrote:
: Or maybe somebody needs to develop a few gigs of bloatware to
: systematically compile a specific individual's postings from Usenet,
: analyze the entirety for any improprities of logic or deviations from
: the Sacred Will of all Engineering and Audio muses, and emulate the
: style of eighteenth-century poets in hourly postings reminding the
: billions of readers of R/A/T of the ultimate Universal placement of the
: individual in question.
Sure, I appreciate your concern. I don't believe in dumping on
clueless newbies and die-hard subjectivists the way certain hard-core
technical types have done for years on rec.audio.high-end. But I'm
damned sore at Andre for what I perceive to be extraordinary abuses
of this newsgroup and its reader community. You're just gonna have
to deal with my being pissed off at him.
Sorry. I don't believe in being a saint.
Henry Pasternack
har...@q3-consulting.com wrote:
: ...with B+ rising up to 1.4 Vrms instead of 0.7 Vrms.
Very true, but just to note that the correct value is 0.9 Vrms.
Chris Broadbent
Henry Pasternack <hen...@nortel.ca> wrote in article
<6jaaq9$s...@bmtlh10.bnr.ca>...
> ...<ONE JIANT SNIP>
> This
> is why, when sourcing a custom transformer, it makes sense to
> choose a supplier who knows what he is doing and takes the time
> to work through all the details with you.
>
> In the end, to build it and measure it is good advice.
>
> -Henry
>
> --
> ATTENTION! Reply to h...@nortel.ca (hen...@nortel.ca won't work).
Both to re-enforce Henry's comment and as an aside to the thread, may I add
a recommendation for Electra-Print of Las Vegas, NV and Jack Elliano(?)
(forgive me if I get his second name wrong, I'm working from memory).
I am in the middle of building a prototype SE amp based on the SV811-3.
After working up the design (this is my first SE tube amp design), I
ordered both the output and power transformers from Electra-Print. The
output transformers are off the shelf. The power transformers are custom
wound. Not only did Jack spend much time on the phone giving this newbie
much useful advice, he showed me a way to reduce the number of secondaries
I need, thus saving me money. The power supply is now built (the B+, C and
main filament supply are all choke input). All the voltages are, for all
practical purposes, right on the money.
So I second Henry's advice and I thank Jack for his assistance.
I have no connection with Electra-Print or Jack other than being a very
satisfied customer.
Cheers,
Chris.
NOTE: Remove the x in my reply address (xc...@bga.com) if you wish to email
me.
Chris Broadbent
har...@q3-consulting.com wrote in article
<6jcbbs$vst$1...@nnrp1.dejanews.com>...
> ...<SNIP>
> In fact, a swinging choke is a type of choke rather than a way to use
> it. It is a choke with a very small air gap, such that the inductance
> falls as the DC current rises (I think I have this the right way round).
> It avoids the problem with choke-input filtering that under no load,
> the choke value isn't high enough and it becomes in effect a cap-input
> filter, with B+ rising up to 1.4 Vrms instead of 0.7 Vrms.
>
> If you don't use one of these (which according to the Morgan Jones
> book aren't made any more, although you could probably get someone
> to build them for you), you either have to arrange for the current never
to
> drop too low (bleed resistor) or over-value the choke, or some
combination
> of the two.
>
> John
>
>
Electra-Print makes them. For my SV811-3 SE tube amp design, I use choke
input supplies almost everywhere (including the main filament supply). For
my B+ supply, I am using a 15H swinging choke.
However, I also use bleeder resistors, both to provide a minimum draw and
as a voltage divider network to obtain other HT voltages I need (for the
input and driver stage B+ supply and a filament floating voltage for the
driver tube to keep the heater-cathode potential from being too high).
Note: given the high potentials in many of my supplies, I have the chokes
in the -ve line. Electra-Print ensures me that the chokes can take the
potentials, but I figure always better safe than sorry.
Cheers
Chris.
Chris Broadbent
Sheldon D. Stokes <stokes@SPAM_BE_GONEexis.net> wrote in article
<stokes-1305...@backstep.larc.nasa.gov>...
>
> For voltages over 450 volts I have always ended up putting caps in series
> with bleeder resisitors. There is a small range of high votlage caps,
but
> they are expensive.
>
For my supply (choke input, approx. 700V DC out), I use the Axon caps -
20uF@1200V. Currently, I have four in parallel. They are not polarized.
Although they work fine in the supply, I haven't finished the actual amp
stage yet, so I don't know how they'll sound. They cost me about $12 each.
Just FYI.
Cheers,
Chris.
<SNIP>...
SBench
>Sheldon D. Stokes <stokes@SPAM_BE_GONEexis.net> wrote in article
><stokes-1305...@backstep.larc.nasa.gov>...
>>
>> For voltages over 450 volts I have always ended up putting caps in series
>> with bleeder resisitors. There is a small range of high votlage caps,
>but
>> they are expensive.
>>
>
>
I missed this one. Bleeder resistors in SERIES with your caps?
Maybe I'm misreading this, but.... don't you mean capacitors
in series with paralleled bleeder resistors shunting each of them??
Steve
Sheldon D. Stokes
In article <199805131900...@ladder01.news.aol.com>,
sbe...@aol.com (SBench) wrote:
Right, caps in series with bleeder resisitors. I didn't say anything
about the resisitors, or at least I didn't mean to. I just mentioned that
I used bleeder resistors, which I thought was obvious that they would be
in parallel with the caps.
Henry Pasternack
Sander deWaal (au...@pegasus.demon.nl) wrote:
: Choke-input filters give us some voltage loss, and the
: choke has to be insulated really good from the core, to avoid
: breakdown.
Bear in mind that although the choke-input filter gives less
DC output voltage than the cap-input filter, it actually is more
energy efficient. This is because of the low peak value of the
current drawn from the transformer secondary. Resistive losses
go up with the square of current, so all other things being equal,
it is more efficient to deliver current in long, steady waves
than short, peaky bursts.
Using a choke-input filter will result in less heating of the
power transformer and less wear on the rectifier tube. But the
filter choke works hard with a significant AC flux in the core.
This is a source of additional loss, and the reason we want to
use a quality part at this point in the circuit. Cheaply wound
and stacked chokes will buzz, and the buzzing will get worse as
time goes on.
Example: A pair of 35W monoblocks in continuous service for
a year and a half, using Triad 10H, 200mA, 150 Ohm filter chokes.
There has been some low buzzing from day one which was partially
silenced by mounting the choke and power transformer on rubber
washers. The buzz hasn't gotten any worse with time. I've
collected quite a few military surplus chokes in steel cans,
and these should be very sturdy and quiet in choke-input
service.
I really like choke-input power supplies. It's easier to
predict what the output voltage will be, the regulation is
better, and the sound is supposed to be better, to boot. The
main drawback is the extra space consumed. This shouldn't be
a problem for the home constructor who can afford to build a
slightly larger chassis.
Choke-input filters are particularly synergistic with tube
rectifiers. Most old tube textbooks, written before the advent
of silicon diodes and large, reliable electrolytic caps, state
that choke-input filters are preferred for push-pull amp design.
This is primarily because of the improved regulation.
I encourage people to try choke-input filters. And Reverse
Polish Notation calculators.
SBench
>
> I encourage people to try choke-input filters. And Reverse
>Polish Notation calculators.
>
>-Henry
I'll agree with one out of two of those suggestions. I still
have to use my left hand, bent over halfway, looking in a
mirror, crosseyed to deal with RPN.
I guess it goes back to early training.... "Whats 2 times PI equal?"
was asked much more often than "2 enter PI, now multiply,
you idiot" was commanded.
Steve
Steve
George R. Gonzalez
SBench wrote in message
<199805131900...@ladder01.news.aol.com>...
>>Sheldon D. Stokes <stokes@SPAM_BE_GONEexis.net> wrote in article
>><stokes-1305...@backstep.larc.nasa.gov>...
>>>
>>> For voltages over 450 volts I have always ended up putting caps in
series
>>> with bleeder resisitors. There is a small range of high votlage
caps,
>>but
>>> they are expensive.
>>>
>>
>>
>
>I missed this one. Bleeder resistors in SERIES with your caps?
>Maybe I'm misreading this, but.... don't you mean capacitors
>in series with paralleled bleeder resistors shunting each of them??
>
>Steve
No, no, he's just thinking ahead. Capacitor internal series resistances
have been dropping steadily thourhg the years. He's just thinking ahead
for the time not too far off when the resistances will go negative and
we'll have to start adding external positive resistance stabilize things.
fdeu...@blackrock.com
In article <1d8xpgl.17p...@ts01-06.cork.indigo.ie>,
an...@indigo.ie (=?ISO-8859-1?Q?Andr=E9_Jute?=) wrote some more stupid
stuff:
> My 1084V allows for the filters that follow.
WRONG! (How is it that every time you start typing, you manage to say
something even dumber than the last time?)
A spec of "1084V" all on its own, without specifying the waveform, is
meaningless. In this situation, as the waveform is highly irregular, being
effectively pulses with ramps between them, there is no meaningful way to
specify this independant of capacitor size and transformer losses; in other
words, this waveform has a large crest factor, and the exact waveform is
highly dependant on the circuit values. Consequently, the whole question is
meaningless from the start.
> Once you get in the kV
> range, if you want to KISS your choice of filter components is
> drastically reduced, so you choose them first and then worry about the
> mains tranny. You just about have to get the mains trx on the mark first
> time. Looking for just the right 1600-2000V cap to adjust the B+ by
> exactly the right voltage can cause severe hair loss.
But you are free to parallel or series combine caps....
> Unfortunately, it's not so close. To get an idea of rectifier drops at
> really high voltages (for audio, the hams wouldn't necessarily agree), I
> once spent a day swapping rectifiers in an 845 amp and the variations
> were surprising.
The variations were NOT AT ALL SURPRISING, once you understand the
relationships inherent in the circuit; yes, if you look at a PS as a simple DC
circuit, you will be completely lost and mystified by the "surprising"
variations just from rectifier changes. However, if you do the learning, and
realize how the capacitor size, current draw, output voltage, rectifier
impedance, transformer losses, and transformer secondary voltage all interact,
the variations are no longer surprising, and are downright predictable. Of
course, this is all covered in even basic electronics texts, and is mandatory
for even a novice amp designer.
> Chelmer also ran some tests for me with a variety of
> rectifiers on a lower-voltage amp; result ditto. You don't actually have
> to do tests; there's a bunch of graphs in the RDH4, p1171ff, which will
> tell you at a glance that a rectifier isn't a generic device. Some of
> the underlying assumptions described in the accompanying text complicate
> matters further.
I don't have a copy of RDH here at work, but I can say with utmost certainty
that the rectifier is indeed a "generic device", and I am quite positive that
the RDH would not in any way state anything counter to this! Rectifiers, of
all sorts, obey very simple laws, and are readily modelable to extremely high
accuracy; however, for power supply design, even the most basic approach to
rectifiers produces useable results: there is no rocket science here, and
claiming the matter as unduly complicated when it is really quite simple
benefits noone.
> rather nastily. (In my experience anyway. I notice that Mattijs de
> Vries, who sometimes comes here, is big on regulation. Presumably he
> likes the sound or maybe he designs better regulators than I do.)
Regulator design is a science; given your expressed disdain for and ignorance
things engineering, I am not at all surprised that you are incapable of
designing a decent voltage regulator. (Of course, the first step in designing
a linear voltage regulator is designing an unregulated supply -- which you
also seem to have rather a lot of difficulty doing, what with shorts and
"unpredictable and highly non-linear" rectifiers and such....)
> I thought about this. But I already have a 20-30V general drop guideline
> from the factory, an example 670V design even if at about half the
> current, and Duncan's low-voltage results, so I was trying to avoid
> drawing curves with kilovolt lashups.
OK, now this is really, really stupid: you have just informed the entire world
that you have not a clue about how a diode works!
Duncan tested the 6D22S at 150mA and 300mA, half its continuous current
rating, and its full continuous current rating, respectively. These are the
only data points you need (I say for the third time).
So, as Duncan did all the testing needed, we really have to wonder what you
would mean by "drawing curves with kilovolt lashups"? I sincerely hope that
you were not intending to place 1kV+ DC across this diode in a "test
lashup"?!!! Perhaps the testing you refer to would instead be testing with a
1kV+ sinewave -- then again, how would you generate that, especially at the
current that would be required? Persumably, the only way to do a "kilovolt
lashup" would be with a kV power transformer, which you alread said you don't
have....
So, the testing bit is nonsense, and you aparently have no clue as to what a
diode does or how it behaves.
> Never claimed to know about tranny innards. Don't intend to become an
> expert now when real experts are plentyful and I can never know as much
> as they already do.
Ahhh, but you *DID* calim to be an expert; after all, you said, "I was asked
to study the mains trx spec and some people would sell their wives for a
tranny like that." Sure sounds like you are claiming to be an expert, and an
expert who claims to be involved in consulting, no less....
> But I was so pissed off with the dumb interference here for two days
> that I simply designed a regulated supply
Um, as the first part of a linear regulated supply is an UNregulated supply,
this then implies that you have solved your original problem.... But of
course, you are lying here: you have never designed a regulated supply in your
life!
> and, in case that too sounds
> whited-out like previous regulated HV supplies, in addition redesigned
> the amp to work without regulation even if the mains voltage varies 20
> per cent, not as big a job as it sounds, fortunately.
One wonders how you managed this, when you just a day or so ago had no notion
of transformer losses, rectifier voltage drops, etc, etc: are you implying
that your amp works reliably with a +/-20% variation in B+? Or are you
claiming that you engineered your amp to take into account the passive
regulation aspects and you can tell us what the amp B+ will be for a +/-20%
line voltage variation? Somehow, I think we all know that the answer is
neither: you simply made up a convenient, though rather transparent, lie.
Way to go, Buffoon!
-frank
Sander deWaal
On Wed, 13 May 1998 14:39:56 GMT, har...@q3-consulting.com wrote the
following:
>In article <355b86fb....@news.demon.nl>,
> au...@pegasus.demon.nl (Sander deWaal) wrote:
>> Choke input means, that the cathode of the rectifier goes directly
>> into a choke (often called "swinging choke"), and after that, a cap
>> to ground.
>
>In fact, a swinging choke is a type of choke rather than a way to use
>it. It is a choke with a very small air gap, such that the inductance
>falls as the DC current rises (I think I have this the right way round).
>It avoids the problem with choke-input filtering that under no load,
>the choke value isn't high enough and it becomes in effect a cap-input
>filter, with B+ rising up to 1.4 Vrms instead of 0.7 Vrms.
>
>If you don't use one of these (which according to the Morgan Jones
>book aren't made any more, although you could probably get someone
>to build them for you), you either have to arrange for the current never to
>drop too low (bleed resistor) or over-value the choke, or some combination
>of the two.
Forgive me for asking, but doesn't every choke have an air gap?
I mean, there's DC running through it, just like in a SE output
transformer, and wouldn't the core reach saturation too quick,
if there wasn't any?
Sander deWaal
On 13 May 1998 19:00:58 GMT, sbe...@aol.com (SBench) wrote the
following:
>>Sheldon D. Stokes <stokes@SPAM_BE_GONEexis.net> wrote in article
>><stokes-1305...@backstep.larc.nasa.gov>...
>>>
>>> For voltages over 450 volts I have always ended up putting caps in series
>>> with bleeder resisitors. There is a small range of high votlage caps,
>>but
>>> they are expensive.
>I missed this one. Bleeder resistors in SERIES with your caps?
>Maybe I'm misreading this, but.... don't you mean capacitors
>in series with paralleled bleeder resistors shunting each of them??
There was a comma missing, I suppose.
"Putting caps in series, with bleeders"......
Sander deWaal
On 13 May 1998 16:54:06 GMT, hen...@nortel.ca (Henry Pasternack)
wrote the following:
> Sure, I appreciate your concern. I don't believe in dumping on
>clueless newbies and die-hard subjectivists the way certain hard-core
>technical types have done for years on rec.audio.high-end. But I'm
>damned sore at Andre for what I perceive to be extraordinary abuses
>of this newsgroup and its reader community. You're just gonna have
>to deal with my being pissed off at him.
>
> Sorry. I don't believe in being a saint.
You're entitled to your opinion, but meanwhile,
a very interesting thread was generated, and I
learned some things here.
I'd like to thank you, as well as all others who
participated, for your information.
The Man
I have a stupid question relating to this. Would it be easier/better to
regulate the A.C. before it gets to the HT trans, so that the tranny always
sees the same ac voltage. Could you do it with some kind of triac setup maybe
feeding a choke/cap to round off the squareness?? Please be kind I'm just a
rank amateur.
Paul G.
Henry Pasternack wrote:
> Consider, too, that the output of an unregulated 1100V power supply
> changes 10V for every 1V change in AC line voltage. If the power company
> delivers 115V in the early evening while Andre is cooking his Crubeens
> and brown soda bread, but 122V at 2:00AM while he is composing chapters
> for his authoritative primer on tube audio design, the B+ voltage in his
> amplifier will vary by about 70V. For this reason it's just a wee bit
> silly to specify as he has done a supply voltage to four significant
> places.
>
Sander deWaal
On Wed, 13 May 1998 16:01:58 GMT, The Man <go...@jail.com.for.15.years>
wrote the following:
>I have a stupid question relating to this. Would it be easier/better to
>regulate the A.C. before it gets to the HT trans, so that the tranny always
>sees the same ac voltage. Could you do it with some kind of triac setup maybe
>feeding a choke/cap to round off the squareness?? Please be kind I'm just a
>rank amateur.
Hi Paul,
This is certainly possible, but such a system is too slow to
react to fast current demands.
I tried this once in a SS amp.
The trick was, to wire a bridge rectifier in series with the
AC mains input voltage, and load the DC side of it with a suitable
powertransistor, driven by a comparator circuit via an optocoupler.
The main reason why this isn't very good, is
A. slowness (smoothing caps have to be charged and discharged
to regulate B+, this takes time),
and B. the comparator reacts inly to the one secondary voltage that
is used to compare.
There's also the risk of burning out the bridge or parallel regulator,
leaving you with a too high primary and secondary voltage.
But still, this is the way, some switched mode power supplies operate,
so it can be done.
Not suitable for (tube) amps, IMO.
André Jute
Sander deWaal <au...@pegasus.demon.nl> wrote:
> meanwhile,
> a very interesting thread was generated
Glad to oblige, Sander.
Andre
--
Andre Jute an...@indigo.ie COMMUNICATION JUTE
--see our pages for music lovers, writers and audiophiles at
http://indigo.ie/~andre/ComJuteF1.html
http://www.foundmark.com/ComJute/ComJuteF1.html
Bob Fitzgerald
Henry Pasternack wrote in message <6jcq0o$e...@bmtlh10.bnr.ca>...
> I really like choke-input power supplies. It's easier to
>predict what the output voltage will be, the regulation is
>better, and the sound is supposed to be better, to boot. The
>main drawback is the extra space consumed. This shouldn't be
>a problem for the home constructor who can afford to build a
>slightly larger chassis.
I agree too, and will be playing with some old chokes to prove it (to
myself) as well.
BTW, another small disadvantage is there are very few transformers available
at the plate supplies I need (500V out of the choke) that have the 5V and
6.3V filament power as well. (At least not in the Hammond = reasonable
line). This has prompted me to consider a Hexfred rectifier feeding the
choke.
>
> Choke-input filters are particularly synergistic with tube
>rectifiers. Most old tube textbooks, written before the advent
>of silicon diodes and large, reliable electrolytic caps, state
>that choke-input filters are preferred for push-pull amp design.
>This is primarily because of the improved regulation.
I am drawn to them for this reason as well, plus I like overkill. If I NEED
200ma, I prefer to use a 300-400ma t-x, and the big Hammond plate supplies
are the only way to get this reserve. Ahhhhh....., the sweet smell of
engineered, cool running overkill power supply for my new monoblocks.... :~)
>
> I encourage people to try choke-input filters. And Reverse
>Polish Notation calculators.
Check - check.
Bob
Ron Bales
fdeu...@blackrock.com wrote:
> > Never claimed to know about tranny innards. Don't intend to become an
> > expert now when real experts are plentyful and I can never know as much
> > as they already do.
>
> Ahhh, but you *DID* calim to be an expert; after all, you said, "I was asked
> to study the mains trx spec and some people would sell their wives for a
> tranny like that." Sure sounds like you are claiming to be an expert, and an
> expert who claims to be involved in consulting, no less....
He was checking for bees.
> > But I was so pissed off with the dumb interference here for two days
> > that I simply designed a regulated supply
>
> Um, as the first part of a linear regulated supply is an UNregulated supply,
> this then implies that you have solved your original problem.... But of
> course, you are lying here: you have never designed a regulated supply in your
> life!
Hey Frank be fair, just because he can't design a simple unregulated supply
doesn't mean he isn't competent to design a regulated kilovolt supply.
Sorry, just thought I'd try that one out, sings like a duck. Nope, the
boy's hopeless, just as you said.
I'll give him this, he won't stop swinging. First he's the great and
magnificent Jute and then he's humbly begging for help from those he
insulted mere days ago. Is there a technical term for that?
ROn
Bob C
Ron Bales wrote in message <6jdep5$3ul$1...@gte1.gte.net>...
>fdeu...@blackrock.com wrote:
>
>Hey Frank be fair, just because he can't design a simple unregulated supply
>doesn't mean he isn't competent to design a regulated kilovolt supply.
>Sorry, just thought I'd try that one out, sings like a duck. Nope, the
>boy's hopeless, just as you said.
>
>I'll give him this, he won't stop swinging. First he's the great and
>magnificent Jute and then he's humbly begging for help from those he
>insulted mere days ago. Is there a technical term for that?
>
>ROn
The technical term for him is... GIANT <> ASSHOLE!!!
Bob
har...@q3-consulting.com
In article <355b0d1a....@news.demon.nl>,
> Forgive me for asking, but doesn't every choke have an air gap?
> I mean, there's DC running through it, just like in a SE output
> transformer, and wouldn't the core reach saturation too quick,
> if there wasn't any?
I said "a very small air gap". Normal chokes have a larger (though still
small in absolute terms) gap. (Of course, there are plenty of chokes
with no air gap at all, but we don't use them for smoothing DC).
John
Bob Fitzgerald
Sheldon D. Stokes wrote in message ...
>I've done that as well. Some cheap transformers are wound on two bobbins,
>one for primary, and one for secondary. That makes it super easy to wind
>whatever you want for a secondary.
Actually, as far as I know this is not a "cheap" design, but rather
intentionally done to minimize the capacitive coupling between the two
sections. It is important to minimize HF noise transfer between the
primary/secondary. The best way to get the same effect on a toriod is to use
an electrostatic shield, which costs very little extra if you are specifying
a design, or doing it from scratch as below. I've not seen it used much in
audio, but it is seen frequently in other more critical industries (i.e.
medical electronics, etc)
Bob
Henry Pasternack
Bob Fitzgerald (bf...@capital.net) wrote:
: I am drawn to them for this reason as well, plus I like overkill. If I NEED
: 200ma, I prefer to use a 300-400ma t-x, and the big Hammond plate supplies
: are the only way to get this reserve. Ahhhhh....., the sweet smell of
: engineered, cool running overkill power supply for my new monoblocks.... :~)
I think if you check Surplus Sales of Nebraska (www.surplussales.com) you
may find some nice Collins potted plate transformers in voltages that will
be of use to you. You'll need more trannies, unfortunately, for filaments
and bias supplies...
My main concern with using silicon rectifiers with a choke-input filter
is to insure that there is adequate series R to damp the network and prevent
"filter bounce". Using a large capacitor after the choke helps because the
Q goes down for a given resistance.
Henry Pasternack
fdeu...@blackrock.com wrote:
: (Of course, the first step in designing a linear voltage regulator is designing
: an unregulated supply...)
Yes. I was looking into tube-based regulation a while back and it turned
out to be a big headache getting the raw B+ to fall into a reasonable range
to insure proper margins across the pass tubes. This was for a Class AB amp
with strong variation in average and instantaneous load current.
If you look at common regulator circuits, they use a large plate resistor
for the driver tube, and as the raw B+ draws down, the voltage drop across
this resistor becomes small. There is potentially a slewing problem at the
grid of the pass tube in the event that the grid voltage has to change quickly
to accommodate a sudden rise in load current. The transient response of the
regulator, evidently, will depend on the output capacitor filling in while
the active circuitry catches up. It's painful to do the math to model this
behavior since the usual small-signal analysis doesn't apply. I don't do
this stuff for a living, and I don't have SPICE, so I kind of shelved the
investigation pending the purchase of some textbooks on the subject. But
there don't seem to be any decent books on linear supply design these days,
given that most power supply designers are into switch-mode.
None of the usual "Glass Audio" type articles, nor the classic undergrad
tube texts, or even the RDH, get into this level of detail. I ought to
build one of these up and measure it, but I honestly don't have the time.
I'd be keen on someone putting one of these circuits into SPICE so we can
explore the dynamic behavior in depth.
Anyway, my point is that I agree with you that designing a regulator,
especially for such a high B+ voltage, is an orders-of-magnitude bigger
headache than doing a passive supply right in the first place. At least
if you want to do it right.
Sheldon D. Stokes
> > I encourage people to try choke-input filters. And Reverse
> >Polish Notation calculators.
>
> Check - check.
Wait, isn't that:
check <enter> check
Sheldon
--
#include(dumbquote.h)
Remove SPAM_BE_GONE from my e-mail to reply directly to me
Brad Thompson
Henry Pasternack wrote:
>
> fdeu...@blackrock.com wrote:
> : (Of course, the first step in designing a linear voltage regulator is designing
> : an unregulated supply...)
>
> Yes. I was looking into tube-based regulation a while back and it turned
> out to be a big headache getting the raw B+ to fall into a reasonable range
> to insure proper margins across the pass tubes.
<snip>
Hello--
One technique which I dimly recall seeing in test instruments (e.g.,
RF/microwave signal generators built in the 1950s-1960s by Kay
Instruments) used a constant-voltage transformer (CVT) as a
preregulator. A CVT features complex magnetic-circuit design and an
external resonator capacitor to produce AC regulated to a few percent.
After rectification, the resultant DC drives a conventional regulator,
but preregulation eases some of the regulator's design constraints.
However, CVTs are sensitive to line frequency-- generally not a problem
in most locales; also, in my experience they're not readily available on
the surplus market. A CVT also runs thermally hot.
Used CVTs sometimes show up at photo swap meets, where 1:1 ratio (117V
in to 117V out) units for powering enlarger lamps are available for a
few dollars.
73,
Brad AA1IP
Bob Fitzgerald
Sheldon D. Stokes wrote in message ...
>
>> > I encourage people to try choke-input filters. And Reverse
>> >Polish Notation calculators.
>>
>> Check - check.
>
>
>Wait, isn't that:
>
>check <enter> check
>
>Sheldon
Yeah, check <enter> check <plus> I suppose...
Amazingly, after being totaly confused for years by constantly going between
RPN at work and normal $3.00 calculators at home, I found myself actually
*selecting* RPN on my HP 100 Palmtop, which has a software selection for
either mode... why ask why?
Bob
Bob Fitzgerald
André Jute
Eight days ago I sent this message:
André Jute <an...@indigo.ie> wrote:
> Can anyone tell me what the centretapped secondary voltage requires to
> be before a pair of Svetlana's 6D22S halfwave rectifiers if I want to
> draw 1084V at 281mA after the rectifier tubes and before the pi filter.
> Full wave, cap input, no info on tranny winding resistance so assume
> some normal value--that's the reason I don't just run a test, that the
> info is required to specify a mains tranny not yet in existence.
>
[Snip, sources of partial information about 6D22S voltage drop]
>
> Any information, help or hints will be appreciated.
>
> Andre
> --
Heartfelt thanks to those on the newsgroup who made useful suggestions
and those who were kind enough to write to my mailbox with entertaining
remarks. As for the rest... After a couple of days of the usual
schoolyard bullies jerking off over a piccie of me, and imparting
absolutely no useful information, I simply put the last of these people
whom I still read, Pasternack, in my killfile for good, and worked it
out myself in a brief, agreeable correspondence with friends who don't
come here because of the appallingly poor S/N ratio.
The result, together with a generally applicable method, is posted at
http://indigo.ie/~andre/RealMcCoy.html#Mainssec
Andre
If the posturing wankers on the net won't help you, Ohm's Law will.
Henry Pasternack
=?ISO-8859-1?Q?Andr=E9_Jute?= (an...@indigo.ie) wrote:
: Heartfelt thanks to those on the newsgroup who made useful suggestions
: and those who were kind enough to write to my mailbox with entertaining
: remarks. As for the rest... After a couple of days of the usual
: schoolyard bullies jerking off over a piccie of me, and imparting
: absolutely no useful information, I simply put the last of these people
: whom I still read, Pasternack, in my killfile for good, and worked it
: out myself in a brief, agreeable correspondence with friends who don't
: come here because of the appallingly poor S/N ratio.
I'm sorry you aren't reading my postings anymore, Andre. No doubt some
kind reader of r.a.t. will "forward" this message to you for your perusal.
Your disparaging remarks about the "appallingly poor S/N ratio" of the
discussion here are unfortunate. I thought the thread was actually quite
productive. You must have, too, since much of the information in the
middle part of your article comes straight from my postings.
I disagree with your assessment of the "art" of transformer design.
Transformer design is engineering, not art. Now, actually building a
transformer is a craft. That's a related, but distinct activity.
Designing a power supply is the job of the amp designer. That means
choosing a filter topology, reading the rectifier curves, running the
math and working with the catalogs or transformer builder to arrive at
a complete solution. You're right, it's not rocket science.
But, as usual, you have shuffled and mumbled your way through a bunch
of calculations without really understanding them, coming up with an
answer that seems to be about correct, if only by accident.
For instance, you state that the rectifier drop should be doubled
because there are two plates in the circuit. But you are calculating
for only one half of the circuit, so the proper drop should be that
of one plate, or about 20 Volts. It works out in the end because the
variations in the rest of the design are so broad. Not reading my
postings? Let's see how long it takes for this oversight to be corrected
on your Web page.
You use Ohm's law to calculate DC voltage drops based on the DC load
current and winding resistance. But the current waveform in this part
of the circuit consists of short spikes of pulsating DC. The actual
RMS power dissipated will be up to several times greater than what you
calculate. That's why there are curves giving peak to average current
ratios as a function of filter capacitance and series resistance in the
power supply section of the RDH.
Your concluding remarks about the "diplomaed quarterwits on the
newsgroups" are most unfortunate. You use the problem of varying mains
voltage to illustrate the absurdity of trying to specify the raw DC
voltage to a very tight tolerance. You advise builders to "leave some
headroom in the design of the amp to account for mains variations."
This is precisely the advice this and other "quarterwits" gave in
response to your original four significant digits specification for
B+ voltage.
A rigorous analysis won't do anything to improve raw supply regulation
under conditions of varying mains voltage. It will, though, give the
designer the precise limits of DC output voltage for a given change
in AC input.
Once again, you have done a disservice to your readers by presenting
watered-down, jumbled material as though it is authoritative. Power
supply design is complex and there are straightforward ways to deal
with it. For the beginner, the "rough shortcut", lifted from my postings,
will work very well. For the advanced hobbyist or neophyte engineer, the
correct approach is to start with a textbook such as the RDH and work
through the details.
The reason I criticize you, Andre, isn't simply that I don't like
you and hold a grudge. It's because you are misinforming the public,
creating confusion when you have, instead, an opportunity to clarify
complex topics. And then you have the gall to attack in the most
unpleasant way people who take the time to set your mistakes straight.
This is really fucked, and takes away significantly from my enjoyment
of this newsgroup.