45 Driving 845?
Ross Hershberger
I've been working on a silly idea to use my current fave triode, the 45 to
directly drive an 845. The design involves about 800V B+ for the output and a
5K pri OPT. The 45 would have a 10K resistive load, have cathode bias and
would run from supplies of about -440V and +250V. Driver current would be 35mA
and A-K voltage about 270V. The 45 would have to swing about 200V P-P for full
output from the 845. I like the idea of 'opposing' DHT curves to cancel some
distortion products. The input tube would be a 76 or 56 in the current plan.
The only major obstacles are the number of power supplies for 3 stages and a B-
supply and the fact that the input tube's anode is at +250 while the 45's grid
is at about -430, requiring a coupling cap of over 1KV rating. I have all of
the parts on hand to do this except for a few common chokes.
Has anyone trried this? Let me know before I start cutting metal for the 4
chassis that this silliness would require. Thanks!
Correlation is not causation.
Ross Hershberger
Royal Oak MI
RHE...@AOL.COM
GRANT G 10
direct coupled to the 845. This works great!! Best way to do direct-coupling.
I have also tried choke loading the DHT driver but preferred Mills resistor.
Other good choice for the driver tube would be a globe #10 (more voltage swing
since mu=7 or 8). I think you'll have a very good sounding amp on your hands.
For the high voltage cap required between input and driver tubes, I recommend
the 1000V Copper Foil Jensen Oils. I also like to use high transconductance
input tubes for this kind of topology (WE 417A/5842, 5687, WE 437A, etc.)
Also, using adjustable tube regulated power supplies for the +250V and -440V is
a good idea for dialing in and providing a stable bias for the 845. Good Luck!
-Grant
Grover Gardner
>
> Grant;
> Yes, the design that I pencilled up direct couples the plate of the 45 to the
> grid og the 845. I planned for an extra-big cathode resistor on the 45
> bypassed with a pot to allow the bias to be adjusted (thereby also adjusting
> the 845 operating point).
Whoa! How does this work? You're going to put the 845 grid at positive
potential, no? If the 845 supply sits on top of the 45 supply, I guess it's
possible, but I'm having trouble picturing this.
> You mention the Jensen 1KV coupling caps. Are they available from Angela?
You don't need this rating. The only voltage you need be concerned about is
the plate voltage at the forward end of the cap, say 250vdc, so a 400vdc cap
would suffice. Unless I'm mistaken, the cap doesn't care where the
terminating grid sits.
I don't want to sound discouraging, but this is a fantastically complicated
design! For starters, I can't think of how you actually turn this all on
without some sort of problem. I think you can achieve something very close
with a much simpler topology.
First off, the 76 has a gain of, what, 13? With a resistive load, let's say
10 at best. A resistively-loaded 45 will maybe get you a gain of 3, probably
less. You're going to need a hell of a preamp to drive this 30 volts to 250+.
I suppose something with an output of 8-10 volts would do it. The Ankoru
uses a 7044 in SRPP, right? That's a gain of 20 or so, plus an IT-loaded 300B
for a gain of 3.8--that's less than 80, but with one of their MONDO output
preamps you can get the swing you need.
If you want direct-coupling, may I suggest that you use your 800vdc B+ and
directly-couple the first two stages, then cap-couple to the 845. Use
separate supplies if you like. I wouldn't drive an 845 with a 45, seems like
too much objectivity to me, but I could be wrong--it wouldn't be the first
time :-).
Sorry, I know this is boring but I'm having trouble getting a handle of some
of the ideas you've raised.
--
Grover Gardner
gro...@postoffice.att.net
Ross Hershberger
Yes, the design that I pencilled up direct couples the plate of the 45 to the
grid og the 845. I planned for an extra-big cathode resistor on the 45
bypassed with a pot to allow the bias to be adjusted (thereby also adjusting
pair of those laying around too. That would allow me to eventually go to
higher B+ on the 845, as the 10Y will swing more volts as a driver than the 45
will. I'll sketch up a version with 10Y driver and see how that looks on
paper. The OPTs that I have are E-P KL10KBs and are designed as 10K pri with
120 mA max. I would use them as 5K at B+ of 800V or lower. I also have on
hand additional power iron to make a 1200V 250ma B+ supply with LCLC filtering.
This is a huge set of iron that would take serious engineering to assemble
into a working supply. I'm currently planning around more modest B+ of 800V
from a Thordarson 1800VCT PT and a 38H-16H swinging choke. My plan would be to
build the audio chassis as monos, with a separate chassis holding the
input/driver supplies and theit fil supplies. A fourth chassis would hold the
output B+ supply and 845 fil supplies. This all has to be broken down into
separate packages for physical reasons. I recently reinjured a bad shoulder
and I'm seriously limited in what I can lift.
The input tube question is up in the air. My present SE 45 amp uses a 6SL7,
but I think that tube gives away too much bass. I want my next SE to be truly
full range if possible. I initially drew in the 76 because it's a single
triode with acceptable specs. I was thinking of a 7044, which I believe is
similar to the 5687 that you cite. The A-N Ankoru goes
7044/cap/300b/IT/parallel 845. There's a pair of those A-N ITs laying around
here somewhere, but I don't think they're needed for thjis project. Choke
loading the driver crossed my mind, but I don't like the (on paper) rolloff in
the bass. With two cathode biased stages there will be too much of that
already. I have never worked with the 417/5842. People seem divided about it.
The specs certainly are impressive. Some say it's not linear enough. Others
say 'screw that, it sounds awesome'. On paper it's easy to pick devices out of
books according to specs but experience tells me that that doesn't always give
the best audible results (sigh).
Whether I build this or not will largely depend on the mass once I get the bits
all in one heap. There's no point in building something that's going to shred
my ligament again when I try to flip it over.
You mention the Jensen 1KV coupling caps. Are they available from Angela? My
current printout doesn't list them. If there's another source, please let me
know.
I have to admit I really get a kick out of sitting down with a calculator, a
pile of books and designing stuff on paper. Mostly sketching unlikely things
(cathode drive 300Bs, cathode loaded PP, etc). It's entertaining to take an
idea that seems reasonable on the surface, analyze it and find the drawbacks
and flaws. This is a long way from actually making good sound, but learning
what looks bad on paper is one step toward learinid what works well in reality.
Thanks again for the help and encouragement.
Cheers
Ross H
In article <19990701195305...@ng-fp1.aol.com>, gran...@aol.com
(GRANT G 10) writes:
GRANT G 10
think. I definitely would go with the direct-coupling instead of the AN IT's
you have. The 7044 is similar to the 5687, but it's a little less delicate and
more powerful-sounding. I still like the WE417A and 437A best. Remember that
you'll need quiet DC filament supplies for the driver tubes as any residual
hum is amplified by the 845. Have fun, Grant
GRANT G 10
350V drop from 220V potential, so grid of 845 is -130V. The 1000V cap is
needed from the plate of the input tube (~200V) to the grid of the 45 (-440V_
difference of 660V without the AC component.
This design may be complicated but very worth it!!! My design uses WE 437A cap
coupled to WE VT-2 direct-coupled to Telefunken RV-218.
-Grant
Grover Gardner
catch on :-). I guess it's not that much more complicated than mine, with a
direct-coupled CF driver... It never occurred to me you had to rate the cap
for the difference in dc potential between the plate and following grid. You
learn something new (and safe) every day!
--
Grover Gardner
gro...@postoffice.att.net
Ross Hershberger
tortured circuit description and see the sense of what I'm trying to do.
Trying to describe three stages and four supplies in text makes me want to get
the scanner working and just throw a picture up on my AOL page space, saving
the equivalent thousand words for something more fun.
If I can pester y'all for a little more data, please recommend operating
parameters for the 417 and 437 driving a 45 or 10Y. I have no curves for WE
tubes and very few schems in hand that use them. IIRC the 417 is an RF single
triode with high transconductance, Rp about 1800 Ohms, 4.5W PD, likes to run at
about 18ma and needs stoppers on all of its grid pins to stop oscillation.
What load and Ep would it like to see? Same questions for the 437, for which
I have never seen data or applications.
As for the complexity, I think I'm up to it. This is actually a bit simpler
than the PP 300B amps that I built, which surprised me by operating just as
intended on the first try. It's all stuff that I've done before and looks
pretty straightforward. The HT safety issue concerns me, but I managed to
overhaul and mod a pair of Altec 1570Bs and live to tell, so maybe I've got a
little luck left.
Thanks again for the help. This advice is precious and it can't be bought from
a catalog at any price.
Cheers
Ross H
In article <19990702001614...@ng-fr1.aol.com>, gran...@aol.com
(GRANT G 10) writes:
>
>Grover: The 45 will draw about 35ma thru the 10k plate resistor so there'll
>be
>350V drop from 220V potential, so grid of 845 is -130V. The 1000V cap is
>needed from the plate of the input tube (~200V) to the grid of the 45 (-440V_
>difference of 660V without the AC component.
>
>This design may be complicated but very worth it!!! My design uses WE 437A
>cap
>coupled to WE VT-2 direct-coupled to Telefunken RV-218.
>
>-Grant
>
bobcx
Ross Hershberger <rhe...@aol.comremove> wrote in message
news:19990702072402...@ngol02.aol.com...
> If I can pester y'all for a little more data, please recommend operating
> parameters for the 417 and 437 driving a 45 or 10Y. I have no curves for
WE
> tubes and very few schems in hand that use them.
Ross,
Check out http://www.vt52.com/ for the data on both.
A place to start for an operating point would be about 150V, 10ma, and a
10K plate resistor. You might also take a look a the 2A3 as a driver.
Arthur Loesch uses a SP 2A3 in his 845 design. I have not personally heard
it, but have heard good things about it. When I get around to building an
845 amp, I will probably go with a SP 2A3.
Bob
GRANT G 10
cathode resistor, 5k plate resistor. 437 runs at 180V, 100 ohms cathode
resistor, 3.5k plate resistor. On that web site there's a schematic under DIY
for a single tube power amp using 437A (1 watt). Black Gate non-polar 220uF
6.3V is a great cathode bypass for either.
-Grant
CERREM
>potential, no? If the 845 supply sits on top of the 45 supply, I guess it's
>possible, but I'm having trouble picturing this.
Here's another thing to try...
I have built direct coupled amps with some of the stages using only negative
supply..
This way I have the plate at 0 volts...
I have one amp where 211 OPT is connected between ground and plate..
CHEERS
CM
Grover Gardner
that I understand it better. I like Cerrem's idea of stages with totally
negative supplies, too.
Here are some *helpful* thoughts :-). The 10 typically wants 450 plate volts,
I think, and also demands a higher load--this may complicate matters or not,
but it has nice gain. Pricey, though, if you don't already own a pair. Might
I suggest the 46 or 47? More gain, slightly lower current and an extremely
lively and 3-D presentation, plus they are cheap. But you have the 45s
already, so... I do like the idea of a higher-gain, higher trans tube up
front. To my ears, you got to have some gain and punch somewhere in the amp,
and the input tube is a good place, IMHO.
--
Grover Gardner
gro...@postoffice.att.net
Ross Hershberger
Thanks for the info and references on the 5842 and 437. This morning I
remembered an old SP article on a preamp featuring the 5842 so I scanned that
at lunch and worked up a driver schem with 5842/10Y. I calculated a gain of
11dB, which is OK for a SE power amp. I'm concerned about the 10Y's high Rp of
5K Ohms as a driver. I'll have to work out the input capacitance of the 845
and see how much the driver will sag at 20kHz. If it's in the ballpark (<1dB),
I'm leaning toward the 10Y as a driver due to its lower current requirement and
higher Ep capability relative to the 45.
Thanx a million for the advices and references. I really hope this works
out. If I screw up with the HT supply, be nice to my widow when she flogs my
stuff on Ebay...
Cheers
Ross H
GRANT G 10
higher Rp of the 10Y will make a very small difference on the output impedance
of the driver section (keeping the same 10k resistor on the 10Y). I'd go with
the 10. BTW, the Globe version is my favorite.
-Grant
Andre Jute
> Choke
> loading the driver crossed my mind, but I don't like the (on paper) rolloff in
> the bass.
Do try it before you just dismiss it from a paper study. There was an
amusing exchange on the Joenet a couple of years ago when I let slip I
chokeloaded a 6SL7 driver to 300B, and I was warned that I would have a
bass *boost*. This was one of the most successful amps I ever designed.
You can tune it with the choke, but mostly with the cap, to have
virtually any sound you like. Most of those we built did in fact have a
bass boost, for opera lovers, who are hedonists and don't care much for
a theoretically flat response. I love the choke loaded driver among
other reasons because to me, after that experience, it seems like a
shortcut to a tunable amp: you just swap one coupling cap per channel
and you have a different amp. It is also great for using with horns less
than full size to boost the bass a little, sort of built-in EQ.
Andre
--
Andre Jute an...@indigo.ie COMMUNICATION JUTE
--come with Andre's Jute's Classical Jukebox live to the
West Cork Chamber Music Festival
http://indigo.ie/~andre/WestCorkChamberMusicFest.html
bobcx
Andre Jute <an...@indigo.ie> wrote in message
news:1duceir.1pu...@ts01-001.bantry.indigo.ie...
> There was an
> amusing exchange on the Joenet a couple of years ago when I let slip I
> chokeloaded a 6SL7 driver to 300B, and I was warned that I would have a
> bass *boost*.
>This was one of the most successful amps I ever designed.
You mean that "YOU NEVER DESIGNED".
Don't you realize that everyone knows you are a phony. Please see the
bullshit below. for more JUTE JOKES come to
http://www.communicationjute.com
********************************************************************
T30 Choke coupled 300B SE. An extremely
felicitous sound and a very successful model with
27 pairs built. Availability limited by essential NOS
components. When available from USD6800
depending on tubes and finish.
The special NOS chokes required for T30 are now finished. I'm working on
another version of this choke coupled amp, Type 81, with British and
Swedish tranny winders. Tell me what sort of music you like and I will tell
you if this amp is suitable--it is designed to be an opera-lover's amp,
warm and enveloping, good on voices and orchestral alike. Type 81 is likely
to start under USD8000 (depends on how much further development costs me
and the tranny winders). Deliveries probably early next year. Five pairs of
T81 already on order.
T30, 1pr Real McCoy Type 30, choke-coupled 300B
SE, bought by us from the estate of a deceased
client. Rebuilt as required including new Western
Electric 300B and repolished burr walnut trim. As
new. USD10000 firm.
Sold on that day. If I find another pair, I shall keep it for myself. Any
further pairs for sale will most likely be in excess of $30,000, which is
what the chap who bought the one above from me got less than a month later
from someone who already has one of my 845.
Manfred Huber
>I'm concerned about the 10Y's high Rp of 5K Ohms as a driver.
I would not be concerned about this. The total input capacitance of a
845 is about 100pF (including miller effect and stray capacitance).
With a driver hving with an output impedance of about 5k you get -3dB
at over 300kHz.
I did not calculate the slew rate limitations of
this configuration but with 18mA of plate current in the driver
I´m sure that this is also no problem.
I say go for it!
Regards
Manfred
Ross Hershberger
Thanks for the comments. I did a rough approximation in my head and came up
with the same. No worries about HF rolloff, apparently. The 10Y looks like a
good candidate. I should be able to get gain of 5 or 6 out of the stage, way
better than a 45 or 2A3. On advice received here I'm leaning toward the 5842
as an input stage as well.
More details as I solidify the design.
Cheers
Ross H
In article <7llven$nv9$1...@news07.btx.dtag.de>, MHu...@t-online.de (Manfred
Huber) writes:
Andre Jute
> I did not calculate the slew rate limitations of
> this configuration but with 18mA of plate current in the driver
> I´m sure that this is also no problem.
Assuming 150V signal to the 845, 18mA will get Ross circa 42kHz if his
OPT can handle it. Compare Tim de Pavaricini's Yoshino 845, which was
reported barely to make 20kHz flat with dP's own handwound trannies, and
you will see how difficult it is.
****
As an aside, I'm always amazed at claims of stratospheric HF extensions
for commercial 845 products.
****
I know that for EE's it makes a terrific difference to be able to say
their own 845 goes a notch or two higher than the next guy's, and I used
to worry about it too, but now I think that if a big-tube amp is
designed in every stage within broadly conservative outlines, you can
grab a casual look at a 20k-25k curve on a fixed scope (my favourite
personal scope is an old Tek 2215 because it is quick and easy to set up
and I don't have to RTFM to discover what the knobs do) and if the curve
is clean, concentrate on tuning the quality of the sound rather than the
HF extension; setting up a better scope and downloading the curves to a
computer with a big screen for zooming in on sections of the curve is a
waste of time with a tube as linear as the 845. It is one of those
common cases where trying to find a single parameter of the "goodness"
of the amp (in this case HF) will have large deleterious effects on its
overall livability quotient. All you need is enough beef on the driver,
for an HV 845 the nearer 20mA the better, and after that even the best
OPT trx will set the limit.
What is in fact far more important to the ease an 845 will bring to your
ear than the HF extension is being careful not to give it a grotesque
bass boost. In fact, a perceived, as distinct from measured, "flat"
response on 845 (for my taste, of course) rolls them off not well below
your intended speakers but only reasonably below. If you make them too
warm, they soon start sounding woolly.
More to the point, and this applies to most DHT in SE with ZNFB, for
every tube or topology in which HF will be a technical problem, a much
bigger problem in human psychology raises its head. This is the property
of hearing in which sound must be perceived as balanced at the frequency
extremes if it is not to cause psychological discomfort. Thus an amp
which makes only 10-12kHz will sound very odd if it goes down to 20Hz;
it will sound much better if it rolls off at 50Hz or possibly higher.
Equally, if it goes up to 100kHz, you'd better not roll it off above a
couple of Hz at most at the low end, or it will immediately sound odd to
an experienced listener. In the particular case of the 845, since for
technical reasons you will in any event achieve a less than spectacular
(by the lotta-NFB norm) HF extension, too low a bass rolloff makes it
sound woolly *at the ear*, regardless of what it measures like.
Of course, this is also a good argument for wringing the max HF out of
it, so that you can get bass extension with street cred... <G> But I am
now mature enough not to bother about credibility with the inhabitants
of street corners, and I don't listen to my amp with a scope. Besides,
the sort of guy who can afford to buy or build a really good SE 845 amp
is of an age where the least of his concerns is HF extension, because he
can't hear above 15-16kHz any more.
****
On Wednesday I attended a concert where I thrice heard a cello
octet--Villa-Lobos with Patricia Rozario singing, a Jane O'Leary
premiere, a John Tavener premiere with a ninth cello playing the solo
part. An 845 is just the tube to play that lot on when it comes out on
CD! The 845 is made for cello music.
And the double bass, for you jazz fans. It is no accident that those
smoky jazz stations, which still broadcast on AM when I toured the
States on one of those 99 buck go anywhere once Greyhound tickets in the
1960s, used 845 as their tube of preference. God, that music lives in
the memory.
Kurt Strain
<19990704061339...@ngol03.aol.com>...
>Manfred;
> Thanks for the comments. I did a rough approximation in my head and came
up
>with the same. No worries about HF rolloff, apparently. The 10Y looks
like a
>good candidate. I should be able to get gain of 5 or 6 out of the stage,
way
>better than a 45 or 2A3. On advice received here I'm leaning toward the
5842
>as an input stage as well.
I like this thread. I think it's pretty interesting with all the ideas.
I was thinking about how I would do it, and of course I'm different.
I think a good quality more affordable solution would be a 6SN7
input to a 27 driver to the 845. At least that's my first pass guess
at this. The 6SN7 is a tube that has most gain, and you can
choose one that tailor's the sound since there's so many to choose
from and they sound quite different. The 27 is quite affordable and
very linear over a wide swing with simple RC coupling, and will
provide a gain of about 8. With the 6SN7 providing a gain of about
15 (RC coupled), you have a gain of 120 to drive the 845, plenty.
The 6SN7 could be biased in a number of ways and remain fine,
and the 27 I would bias at Vp = 180V, Ip = 5 mA, B+ = 400V,
Rp = 44K, Rk = 2.7K for Vgk = -13.5V. This will swing 210Vpp
before the soft clip point (Vgk = 0). It will operate class A1 and not
pull the 845 grid positive much, but I would probably not care about
the lost dB or two.
I haven't used the 27 as a driver, but it looks well built for the task
and it's great as a line stage amp. The 27 sounds better than the
56 or 76 to me. The 27 sounds very good in combination with my
6SN7 used in my OTL driver stage. I think people overkill the
driver stages current capacity or output impedance. This 27
driver would have an output impedance of 7.5K. Ipk would be
about 7.5 mA. Into a 100 pF load, slew rate limiting would be
75 V/us, which at 200 Vpp would only be slew rate limited at
119 KHz. It would still be transformer limited. I think using power
triodes would be fine, but technically overkill for the most part
unless you feel they sound much better. I'm not sure of that.
I choose the 6SN7 for the first stage although it too would be
plenty to drive the 845 for the most part, but since I found the 27
to be just that much cleaner and sweeter than the 6SN7 I would want
to try it with the larger signal.
Note I have no experience with 45's or 10's. They are expensive
and require some beefy supplies to get them to do what it seems
is intended - high current drive. It makes for a difficult, large and
tough to maintain amp, I think. Of course this is probably part of
the fun and challenge of it.
For simple and sweet, plenty of gain and no-hassle indirect
heated cathode stages (you could use AC filament power no
problem for these), this would be my first attempt at this amp -
6SN7 and 27 are not an arm and a leg. I might actually build this
some day.
Kurt
Grover Gardner
follower directly-coupled to the 845 grid. I've tried choke-loaded drivers,
and while very nice, I find the rolloff a bit too much on top. Ross was
worried about LF losses, but the choke-load gives me more concerns about HF
extension. This is fine if it's not a problem for one's ears, but I like the
ease and bandwidth of RC stages best. I have tried it without the CF driver,
and while it's very nice and open, I like the control and added musical
complexity afforded by the CF. The 6900 is hard to get--a 6BL7 would work
just as well, a bit warmer and fatter, probably.
Some people don't like the distortion characteristics of the 12AT7, but I
prefer the solid, 3-D quality this tube imparts when used unbypassed. I
should add that all stages are unbypassed, and this too is a matter of
preference. Again, I like way unbypassing carves the instruments and voices
out of space, as opposed to a "portrait" of the sound, so to speak. Losing
the cathode bypass cap is, to me, far preferable to any direct-coupling which
may involve bypassing the cathode resistor. But this is not the way a lot of
people feel. It's a less tubey, more streamlined sound, not always what
people expect from a tube amp. You also have to choose resistors, coupling
caps and other components carefully to avoid the "obscure" or veiled sound
many people associate with unbypassed (and paralled) stages.
--
Grover Gardner
gro...@postoffice.att.net
Kurt Strain
of other people's designs and evaluations and preferences the more
it shows how there's room for every taste when it comes to tube
design. Don't you find that tubes are more "workable" than
transistors? I mean we can make a tube amp sound how we like
it much more easily than with transistors, correct? I think so, at
least.
I like it when others write about how they did something and why
it worked for them. It always leaves me thinking "now how can I
grab a little more of what he's getting and not lose what I like?"
That's the game I try to play, at least. It's so apparent that we cannot
grab everything and get it perfect, and preference is so much a part
of it I cannot understand where someone else is coming from without
these detailed descriptions. So thanks all for sharing your thoughts.
It's truly insightful.
I think it would benefit all us DIY designers if there were more people
willing to tell us what they've done and what they get out of it like
Grover has here. It really gets us calibrated and helps us go down
the right design path with fewer wrong turns for our individual tastes,
I believe. So all you guys building these amps, tell us what tricks
work out for you and why. Grover's post here is a good example.
We just don't have the time to try everything. Except maybe for
TubeGarden, SE experimenter maniac that he is. :-)
Push-pull amps seem less popular to design. I would like to know
the tricks of push-pull a little better. There's a few of you out there
who prefer push-pull. I know what a good push-pull amp can do,
they can do quite spectacular things.
By the way, does anyone know what I should call the "Fully Differential
Balanced Design" OTL amp I have, the Atma-Sphere circlotron?
I still think it's push-pull even though many call it a "double Single-
ended". BAT makes a transformer-coupled "double Single-ended"
circlotron-style amp. Putting in inverted SE amps together in one
package still makes me say it's push-pull, although it has been
categorized "Single-ended" in one magazine listing of amps by
topology. I think push-pull has been unfairly knocked out of
favor by a fad-like thinking mode. It should be no shame in labelling
this amp push-pull I think, but they want the SE label to get more
attention, for marketing reasons I believe. A rose by any other name
would sound the same. What do you think?
Kurt
Grover Gardner wrote in message <37801B67...@postoffice.att.net>...
Acrosound
but in a direction hinted at in your second post of the day to this thread, and
I quote you:
::Push-pull amps seem less popular to design. I ::would like to know
::the tricks of push-pull a little better. There's a ::few of you out there
::who prefer push-pull. I know what a good push-::pull amp can do,
::they can do quite spectacular things.
::By the way, does anyone know what I should call ::the "Fully Differential
::Balanced Design"
Why couldn't someone take your 6SN7-27-845 topology and double it for a push
pull amp assuming that you either have a balanced (differential) source input
to the amplifier or use a phase splitting trans right at the front door. Or am
I missing something very basic?
For the longest time I have thought of building a 845 push-pull amp somewhere
along the lines of the Altec 287F amp or the Musician's Amplifier Senior (which
required a slave amp).
you got me cookin'
MSL
Ross Hershberger
(Acrosound) writes:
>
>Why couldn't someone take your 6SN7-27-845 topology and double it for a push
>pull amp assuming that you either have a balanced (differential) source input
>to the amplifier or use a phase splitting trans right at the front door. Or
>am
>I missing something very basic?
>
>For the longest time I have thought of building a 845 push-pull amp somewhere
>along the lines of the Altec 287F amp or the Musician's Amplifier Senior
>(which
>required a slave amp).
>
>you got me cookin'
>
>MSL
>
Mike;
By coincidence, I woke up this morning thinking about this same issue and
looked up some things in RDH4 to confirm my thoughts. Two SE stages can be fed
identical out of phase signals, and the OPT secondaries tied together out of
phase to form a mono PP amp. To the extent that the 'channels' have identical
characteristics, the even order harmonics would cancel. If the OPT secondaries
are connected with the commons tied together and an 8 Ohm load connected
between the former 8 ohm taps (bridge style), the output will rise by about 30%
over the sum of the two channels' individual outputs in SE as each tube sees a
lower load, as long as the resulting load is not lower than about Rp to each
tube. An advantage: You could build a cheap, sloppy common power supply
(voltage doubler?) and it would act 'stiff' because the currents cancel as long
as neither tube is driven to cutoff. No change in current in the PS means no
voltage drop due to PS internal resistance. Drawback: unbalanced current in
OPTs. Since you have separate secondary windings for each phase, you could
even go as far as summing them acoustically by driving identical drivers in the
same speaker enclosure, inverting the phase at one driver's leads, of course.
I don't know if that would have advantages over summing electrically at the
OPTs secs, but it's an interesting idea and it should still cancel even
harmonics.
A SE stereo amp could be built with the capability of being run as pseudo push
pull by using transformer coupling at some stage with the capability of
inverting one channel's IT winding connection to provide phase inversion. One
possible drawback: the IT's windings probably don't have equal capacitance to
each end, due to winding pattern (MSL comments??). This would cause phase
shift between the phases at HF.
The power supply advantage (currents cancel when identical signals out of
phase) also applies for stereo (left and right) SEs with a common supply. Feed
one channel in out of phase, invert it back at the OPT sec to maintain phase
relationship between the channels and, to the extent that there is information
'in phase' between the left and right channels, the currents will cancel in the
PS, making it act 'stiffer'. Think about PP class A1 if this isn't clear. The
'difference' signals between the two channels would have an interesting effect.
The PS impedance would tend to subtract a little of each channel's unique
information from the other. Sort of crosstalk in reverse. This might actually
increase separation instead of degrading it as a common PS can do. The worst
case, no common information between left and right, would be no worse than a
standard stereo SE with common supply and a mono input signal to both channels.
I'm betting that there is substantial identical information between left and
right more often than not.
Well, that's enough confusion for one day. I caught my wife on her computer
last night reading all of my RAT posts through Dejanews. Now she's all fired
up about helping me build a SE 845 amp. She wants to teach me Autocad for the
chassis layout and we discussed handing the patterns over to her brother's
shop, an automotive fixture build outfit to be built and finished. If you see
posts here from TRIL...@AOL.COM that's my CAD designer wife looking over my
shoulder.
Cheers
Ross H
news.ntplx.net
yesterday I sat down for 2 monoblocks or a stereo, either, when my
son (almost 15) went in. He talked to me about soft-f/x hes using for
little 3d sceneries.
We ended up to draw a little stell chassis, 4 cylinders at the corners,
OT, PT, 2 power tubes ...
He made the tubes and the trannies very nicely (hammonds with
label on top) and he did also the animation.
Now you can see the amp turning arround from each side as AVI-image.
Not yet know where to publish the 0.5 megs...I'll post that separately.
After that I called some local sheet metal companies for 1 or 2 pieces
including all breakouts and even the chrome finish. He stated to me, that
his price merely is based on preparation work, that is, preparing a dxf-file
from my scetches. And it semms to be affordable.
Seems to me, that I'm now going to do that myself with AutoCAD during
next week. So I'll get my own Kits once design is completed ?
I love the idea of making an amp also mechanically in perfect shape.
So, keep me up to date about your wifes influence on your projects.
Regards, Michael
Ross Hershberger schrieb in Nachricht
<19990705083042...@ngol05.aol.com>...
<snip>
Acrosound
::By coincidence, I woke up this morning thinking about this same issue
<snip>
::Two SE stages can be fed
::identical out of phase signals, and the OPT secondaries ::tied together out
of
::phase to form a mono PP amp.
What I had in mind was simpler... one output transformer and using a balanced
source (say the bal out of a DAC or preamp) and then using Kurt's topology as a
fully differential circuit. This way ya don't have to double your money on
outputs....
<bigger snip>
::Since you have separate secondary windings for each ::phase, you could
::even go as far as summing them acoustically by driving ::identical drivers in
the
::same speaker enclosure, inverting the phase at one ::driver's leads, of
course.
::I don't know if that would have advantages over ::summing electrically at the
::OPTs secs, but it's an interesting idea and it should still ::cancel even
::harmonics.
but this seems to be the long hard road to the same end as using a PP output
trans suitably spec'ed for the task at hand. Or am I missing a point or
feature that your trying to bring out in your discussion???
::One
::possible drawback: the IT's windings probably don't have ::equal capacitance
to
::each end, due to winding pattern (MSL comments??).
Not too terribly difficult to acheive equal capacitances to ground (as well as
equal leakage to ground) in-and-of-itself. One sure fire way to achieve such
is to wind on a two section bobbin.... with each half being a mirror image of
the other half. Then you will have a high degree of electrical symmetry. That
said symmetry is very important but only one of the very important things that
you want to achieve.
::This would cause phase
::shift between the phases at HF.
Even in a symmetrically wound (say with bal capacities and leakages to ground)
IT... your still going to have some phase shifts both at the low end and the
high end. What your after is that the phase shifts be small in magnitude as
well as symmetrical. But your not going to eliminate phase shifts in the
response all together... hopefully control it well and push it out far enough
(in the frequency spectrum) that it does not come back and haunt you. But it
will definitely be there....
<smaller snip than previous>
:Feed
:one channel in out of phase, invert it back at the OPT sec :to maintain phase
:relationship .... <snip>
but.. why don't we just hook up each phase of the differential amp to one end
(and opposite) of the output trans on the primary side? This seems so much
easier... again, apologies if I am missing something here.
::Think about PP class A1 if this isn't clear. The
'::difference' signals between the two channels would ::have an interesting
effect.
::The PS impedance would tend to subtract a little of ::each channel's unique
::information from the other. Sort of crosstalk in ::reverse. This might
actually
::increase separation instead of degrading it as a common ::PS can do.
Grover and I were talking about the power supplies this morning and I asked him
about the same point you are raising. If this is a "discernible" real world
"feel-it-in-the-seat-of-your-pants" benefit (niether of us were sure it was or
wasn't) then it must also be weighed against the added complexity and cost of
(minimaly) four power supplies (phase/anti-phase--- two channel stereo)....
and Grove, if I got him right, was concerned about getting all four supplies
and topologies to all hold hands and behave electrically like each other. As I
have understood it from others.. some of the difficulty of doing fully
differential circuits (especially, perhaps with tubes) is to get dead balls the
same gain structures and etc...
::Well, that's enough confusion for one day. I caught my ::wife on her
computer
::last night reading all of my RAT posts through ::Dejanews. Now she's all
fired
::up about helping me build a SE 845 amp.
ALL RIGHT Ross and Gina.... making music together. And they are relative
newlyweds to boot IIRC.
:If you see
::posts here from TRIL...@AOL.COM
If I may ask... what is her handle "trilln451" come from... or was it random
assignment from AOL???
gotta get,
MSL
Grover Gardner
> I think it would benefit all us DIY designers if there were more people
> willing to tell us what they've done and what they get out of it like
> Grover has here. It really gets us calibrated and helps us go down
> the right design path with fewer wrong turns for our individual tastes,
> I believe. So all you guys building these amps, tell us what tricks
> work out for you and why. Grover's post here is a good example.
> We just don't have the time to try everything. Except maybe for
> TubeGarden, SE experimenter maniac that he is. :-)
I hasten to add that I had a lot of "help" with this amp from a more
experienced designer than myself, having heard what he has been able to
schieve with the 845, but I also found, as you mentioned, that some of his
ideas didn't give me the sound I wanted or weren't appropriate to my
components, so I've either altered them or abandoned them in favor of
something similar. I didn't think I'd resort to the 6SN7, but the more I use
in this instance, the more I like what it does for me.
>
> Push-pull amps seem less popular to design. I would like to know
> the tricks of push-pull a little better. There's a few of you out there
> who prefer push-pull. I know what a good push-pull amp can do,
> they can do quite spectacular things.
I think PP amps are in many ways harder to do. Getting the phase balance
right is hard, but see below. Again, this friend of mine does PP circuits
which beat SE all around for clairty and immediacy, but I'll be damned if I've
been able to duplicate his sound with my limited knowledge of PP circuitry.
>
> By the way, does anyone know what I should call the "Fully Differential
> Balanced Design" OTL amp I have, the Atma-Sphere circlotron?
> I still think it's push-pull even though many call it a "double Single-
> ended". BAT makes a transformer-coupled "double Single-ended"
> circlotron-style amp. Putting in inverted SE amps together in one
> package still makes me say it's push-pull, although it has been
> categorized "Single-ended" in one magazine listing of amps by
> topology. I think push-pull has been unfairly knocked out of
> favor by a fad-like thinking mode. It should be no shame in labelling
> this amp push-pull I think, but they want the SE label to get more
> attention, for marketing reasons I believe. A rose by any other name
> would sound the same. What do you think?
If you haven't checked out the Tube Cad Journal at www.glass-ware.com, do so.
It's one of the great freebees on the net. They have an extensive article on
a balanced line stage preamp using a cathodyne phase splitter run into
cathode-followers. They suggest using it to drive a PP amp sans splitter,
like the Urei 845 PP amp, with which I am not familiar. I'm am curious about
this and would like to try it, since it can be used for PP or SE drive.
--
Grover Gardner
gro...@postoffice.att.net
Andre Jute
> Why couldn't someone take your 6SN7-27-845 topology and double it for a
> push pull amp assuming that you either have a balanced (differential)
> source input to the amplifier or use a phase splitting trans right at the
> front door. Or am I missing something very basic?
Kondo did in fact build a fully differential 211 PP amp, tranny coupled
at the input and at all stages. My understanding was that it did not
sell and was overshadowed by the Ongaku, possibly precisely because of
the fashion for SE. His topology was of course not 6SN7-27-845 because
he has his own paradigm.
In fact, many people who start out on SE 845 do in fact end up with PP
845. I have heard one, and it was wonderful, but I heard it only for a
couple of hours, so cannot say if in the long term I would conclude it
is "superior" in any way to SE 845. The thing is that the HVHC 845 I
prefer (to LVLC 845 in the Japanese manner) already has enough oomph to
make dynamics, transients etc, the usual things one finds wrong with
low-power amps except on horns or very high sensitivity speakers,
irrelevant even on panels (I test with a capella voices at modest
listening levels--if you're a loud heavy metal freak, YMMV). Thus the
extra power of a PP 845 is, in the largest sense, not what it is about.
PP 845 do have an ease though that give a sound I noted as "creamy".
Check into Stephan Puechmore's site (I seem to remember finding it via
Claudio Bonavolta's site or just search for "Puechmore") for a
reasonably economical PP 845 and--this is the interesting part--its
entire genesis as SE 211 and then SE 845 iterations as this energetic
and knowledgeable fellow searched for his ideal sound.
****
In Ross's next post, it sounds to me like he is talking about what the
Japanese call single-ended push-pull, SEPP, but that for a mono amp he
wants to add a second OPT, which might make the effect different.
John Byrns
rhe...@aol.comremove (Ross Hershberger) wrote:
> A SE stereo amp could be built with the capability of being run as pseudo push
> pull by using transformer coupling at some stage with the capability of
> inverting one channel's IT winding connection to provide phase inversion. One
> possible drawback: the IT's windings probably don't have equal capacitance to
> each end, due to winding pattern (MSL comments??). This would cause phase
> shift between the phases at HF.
> The power supply advantage (currents cancel when identical signals out of
> phase) also applies for stereo (left and right) SEs with a common
supply. Feed
> one channel in out of phase, invert it back at the OPT sec to maintain phase
> relationship between the channels and, to the extent that there is information
> 'in phase' between the left and right channels, the currents will cancel
in the
> PS, making it act 'stiffer'. Think about PP class A1 if this isn't clear.
Hi Ross,
So far, so good.
> The 'difference' signals between the two channels would have an
interesting effect.
> The PS impedance would tend to subtract a little of each channel's unique
> information from the other. Sort of crosstalk in reverse. This might
actually
> increase separation instead of degrading it as a common PS can do.
I'm not sure that is true, although it is confusing to think about, so I
could be getting it wrong. In the normal situation, where the channels
are driven in the same phase, I believe the cross talk at the output, due
to common power supply impedance, would be in phase for resistance coupled
driver stages, as you said, while the cross talk due to a conventional
transformer coupled triode output stage would be out of phase. If the
output stage is assumed to be the dominant source of cross talk in this
discussion, then when the two channels are connected with one out of
phase, the resultant cross talk due to the output stage would be in phase
at the speaker. This would tend to decrease separation, rather than
increase it as you suggest. SE "para feed" output stages would probably
work like the RC coupled driver stage, and would result in increased
separation as you speculated.
Regards,
John Byrns
Surf my web pages at, http://www.enteract.com/~jbyrns/index.html
Alan Douglas
>:If you see
>::posts here from TRIL...@AOL.COM
>If I may ask... what is her handle "trilln451" come from... or was it random
>assignment from AOL???
Tsk, tsk, not a Hitchhikers Guide to the Galaxy fan? I have the
original BBC radio series on tape, the TV series (not nearly as good),
and the book. Trillian was the astrophysicist with the white mice.
Actually Gina did explain it once here: the remainder is from
Fahrenheit 451.
Cheers, Alan
Ross Hershberger
In article <37813591...@news.newsguy.com>, adou...@gis.net (Alan
Good memory! That was over two years ago and you are correct. She's the
computer maven of the household, and as an automotive fixture designer, knows a
lot more about drafting and layout than I. She's gradually learning the basics
of tube amplification, power supply circuits and component assembly. We don't
know yet what her role will be in this next project, but I hope it will end up
being 'our' amplifier rather than mine this time. Our musical involvement so
far has been mostly sharing CDs and going to concerts. Tonight we're going
downtown (Detroit) to see Crystal Method, Orbital and The Low-Fi All Stars, a
Techno show. Two weeks ago it was Richard Thompson, the English
singer/songwriter from Fairport Convention. Next week, Chemical Brothers, more
Techno. After that, Lilith Fair festival, with Sheryl Crow, Liz Phair, Sarah
Mclaughlin, Lauryn Hill and others. Music is one of the most compelling things
that we share and I'm delighted to be able to involve her in the hardware end
of the hobby. Putting together an audio system that handles acoustic folk as
well as hard Techno has been a challenge, but we're getting there.
Ross Hershberger
Thanks for your comments about IT leakages. It's good to know that they need
not be a limiting factor in themselves.
It certainly is simpler to build a PP amp with a single OPT. What interested
me about the thought question of opposing SE amps driving the same load was; at
what point does one lose the single-ended-ness of the sound? The topology of
using SE OPTs was an extreme case to show that 'SE' components can give some of
the same advantages of PP, and when do they stop sounding 'SE'? I confess I
can't see that there would be a difference between parallel SE and Class A1 PP
as described (grabbing Nomex suit and hood). I won't build something like that
as a practical amp as it doesn't suit my needs. That idea will be filed for
later experimentation when I get around to building a simple SE stereo 2A3 amp
with a pair of DS-025s that I have stashed.
Thanks again for your contribution.
Cheers
Ross H
In article <19990705120748...@ng-fn1.aol.com>, acro...@aol.com
(Acrosound) writes:
>::One
>::possible drawback: the IT's windings probably don't have ::equal
>capacitance
>to
>::each end, due to winding pattern (MSL comments??).
>
>Not too terribly difficult to acheive equal capacitances to ground (as well
>as
>equal leakage to ground) in-and-of-itself. One sure fire way to achieve such
>is to wind on a two section bobbin.... with each half being a mirror image of
>the other half. Then you will have a high degree of electrical symmetry.
>That
>said symmetry is very important but only one of the very important things
>that
>you want to achieve.
>
>::This would cause phase
>::shift between the phases at HF.
>
>Even in a symmetrically wound (say with bal capacities and leakages to
>ground)
>IT... your still going to have some phase shifts both at the low end and the
>high end. What your after is that the phase shifts be small in magnitude as
>well as symmetrical. But your not going to eliminate phase shifts in the
>response all together... hopefully control it well and push it out far enough
>(in the frequency spectrum) that it does not come back and haunt you. But it
>will definitely be there....
>
>
><smaller snip than previous>
>
>:Feed
>:one channel in out of phase, invert it back at the OPT sec :to maintain
>phase
>:relationship .... <snip>
>
>but.. why don't we just hook up each phase of the differential amp to one end
>(and opposite) of the output trans on the primary side? This seems so much
>easier... again, apologies if I am missing something here.
>
>
>::Think about PP class A1 if this isn't clear. The
>'::difference' signals between the two channels would ::have an interesting
>effect.
> ::The PS impedance would tend to subtract a little of ::each channel's
>unique
>::information from the other. Sort of crosstalk in ::reverse. This might
>actually
>::increase separation instead of degrading it as a common ::PS can do.
>
>Grover and I were talking about the power supplies this morning and I asked
>him
>about the same point you are raising. If this is a "discernible" real world
>"feel-it-in-the-seat-of-your-pants" benefit (niether of us were sure it was
>or
>wasn't) then it must also be weighed against the added complexity and cost of
>(minimaly) four power supplies (phase/anti-phase--- two channel stereo)....
>
>and Grove, if I got him right, was concerned about getting all four supplies
>and topologies to all hold hands and behave electrically like each other. As
>I
>have understood it from others.. some of the difficulty of doing fully
>differential circuits (especially, perhaps with tubes) is to get dead balls
>the
>same gain structures and etc...
>
>
>::Well, that's enough confusion for one day. I caught my ::wife on her
>computer
>::last night reading all of my RAT posts through ::Dejanews. Now she's all
>fired
>::up about helping me build a SE 845 amp.
>
>ALL RIGHT Ross and Gina.... making music together. And they are relative
>newlyweds to boot IIRC.
>
>:If you see
>::posts here from TRIL...@AOL.COM
>
>If I may ask... what is her handle "trilln451" come from... or was it random
>assignment from AOL???
>
>
>gotta get,
>
>MSL
Ross Hershberger
common PS by running one channel in reverse phase at the input and reversing it
again at the output.
In article <jbyrns-0507...@207-229-148-71.d.enteract.com>,
jby...@enteract.com (John Byrns) writes:
>
>> The 'difference' signals between the two channels would have an
>interesting effect.
>> The PS impedance would tend to subtract a little of each channel's unique
>> information from the other. Sort of crosstalk in reverse. This might
>actually
>> increase separation instead of degrading it as a common PS can do.
>
>I'm not sure that is true, although it is confusing to think about, so I
>could be getting it wrong. In the normal situation, where the channels
>are driven in the same phase, I believe the cross talk at the output, due
>to common power supply impedance, would be in phase for resistance coupled
>driver stages, as you said, while the cross talk due to a conventional
>transformer coupled triode output stage would be out of phase. If the
>output stage is assumed to be the dominant source of cross talk in this
>discussion, then when the two channels are connected with one out of
>phase, the resultant cross talk due to the output stage would be in phase
>at the speaker. This would tend to decrease separation, rather than
>increase it as you suggest. SE "para feed" output stages would probably
>work like the RC coupled driver stage, and would result in increased
>separation as you speculated.
>
>
>Regards,
>
>John Byrns
>
John;
I'm not following you very well here. I was intending to address only
transformer-output power stages, leaving aside the effect of PS crosstalk on
lower-level stages. It may be useful to consider the crosstalk as common mode
noise and I'll think about it some more that way and see if it gets clearer.
It certainly seems that with a shared PS, having one channel run in opposite
phase should reduce crosstalk rather than increase it, with the best case
results occurring when the program material is true mono. I'm not really sure
how to go about exploiting that or even investigating it, but it's interesting
to think about. I have to admit that this is over my head and I could be
overlooking something that would be obvious to others.
Thanks for your comments.
Cheers
Ross H
John Byrns
rhe...@aol.comremove (Ross Hershberger) wrote:
> John;
> I'm not following you very well here. I was intending to address only
> transformer-output power stages, leaving aside the effect of PS crosstalk on
> lower-level stages. It may be useful to consider the crosstalk as common mode
> noise and I'll think about it some more that way and see if it gets clearer.
> It certainly seems that with a shared PS, having one channel run in opposite
> phase should reduce crosstalk rather than increase it, with the best case
> results occurring when the program material is true mono. I'm not really sure
> how to go about exploiting that or even investigating it, but it's interesting
> to think about. I have to admit that this is over my head and I could be
> overlooking something that would be obvious to others.
Hi Ross,
Yes, I too was trying to address only the output stage, and specifically
an ordinary transformer coupled SE triode output stage. The situation is
different in the case of a "para feed" output stage.
Let's forget about the effect of the cross talk on "separation" for a
moment, and just look at the nature of the crosstalk. Consider a stereo
SE triode amplifier with a common power supply of finite impedance. Let's
start by applying a signal to only one channel, and observe the phase of
the output on the second undriven channel, with respect to the output of
the driven channel. The simplest, non mathematical, way to look at the
situation is to observe that when the current through the output tube in
the driven channel increases, the Voltage drop across the common power
supply impedance will cause the B+ feeding the undriven channel to
decrease. It is generally accepted that when you decrease the plate
Voltage of a triode tube, the current through the tube will also
decrease. This decrease in current through the undriven channels output
tube will cause a cross talk signal at the speaker terminals which is out
of phase with the signal at the speaker terminals of the driven stage,
where the current is increasing.
An outline of a more formal analysis is as follows. When the current flow
through the output transformer of the driven channel, from the B+ supply
to the plate of the tube, increases, the increased current must be
supplied from the B+ supply node. In addition to the output transformer
of the driven channel, there are two other things tied to the supply
node. First is the B+ supply itself, and second is the output transformer
of the undriven channel. The increased current flowing into the driven
channel comes from this node, and a current divider is formed by the B+
supply and the undriven channel to supply the increased current to the
driven channel. The fraction of the current supplied by the undriven
channel depends on the relative impedance of the B+ supply and of the
undriven channel. The fraction of the current supplied by the undriven
channel will flow through the second channel's output transformer from the
plate to the supply node, or in the opposite the direction of the current
change in the driven channel. Again this shows that the crosstalk will be
out of phase at the speaker terminals, and the exact amount of cross talk
can be calculated if all the impedance's are known.
Now that we know that the crosstalk is out of phase, we can go back to
your original statement that "This (inverting the phase of one channel)
might actually increase separation instead of degrading it as a common PS
can do." My take on this is that the out of phase crosstalk increases the
"separation" with the normal connection, where the channels are connected
in the same phase. Inverting the phase of one channel would put the
crosstalk in phase, and reduce the separation.
Kurt Strain
: Kurt: Your earlier post was dynomite. really got me psyched up....
: but in a direction hinted at in your second post of the day to this thread, and
: I quote you:
: ::Push-pull amps seem less popular to design. I ::would like to know
: ::the tricks of push-pull a little better. There's a ::few of you out there
: ::who prefer push-pull. I know what a good push-::pull amp can do,
: ::they can do quite spectacular things.
: ::By the way, does anyone know what I should call ::the "Fully Differential
: ::Balanced Design"
: Why couldn't someone take your 6SN7-27-845 topology and double it for a push
: pull amp assuming that you either have a balanced (differential) source input
: to the amplifier or use a phase splitting trans right at the front door. Or am
: I missing something very basic?
: For the longest time I have thought of building a 845 push-pull amp somewhere
: along the lines of the Altec 287F amp or the Musician's Amplifier Senior (which
: required a slave amp).
: you got me cookin'
: MSL
I've got a design for the 845 push-pull amp laid out now. It's got
a 6SN7 diff amp input stage with tubed constant current source. That's
the phase splitter, and it feed two 27 amp stages, capacitor coupled.
These stages are both run by +-300V rails. The 27's are common cathode
amps with bypass caps that are then DC coupled from the anodes right
into the 845 push-pull stage. This stage also has trimpots for balance
and output bias in the anode circuit, bypassed to keep the pots out of
the load effectively.
The driver power supply uses a 275-0-275 VAC / 5 VAC / 6.3 VAC which
can power the whole thing, including all filaments (27's 2.5VAC filaments
wired in series).
I have 10VCT 4A filament transformers for the 845's. The 845's will be
fixed biased through the 27 anode at -100VDC (adjustable) setting the
27 plate voltage at 200VDC and the 845's grids at -100VDC. I have the
845's operating at 800 VDC with a voltage doubler with 1KV oil cap bypass.
Do Chinese 845's remain stable with fixed bias, or will I need some
cathode biasing to keep it stable? Note I have to power up the drivers
before the output stage or it will run full bore on power up.
Now the big question for Mike. What OPT to use for this? I think I need
one with an Raa between 8K and 10K, and I want 4, 8, and 16 ohm output
options. The bias current will be about 180 mA for class A operation.
Power output looks to be about 50W/ch. I need at least 40W/ch for the
the speakers I want to use these with. Impedance is high, efficiency
is not. So I need a decent push-pull OPT with enough power capacity in
it to do the job here. Any suggestions?
Kurt
Ross Hershberger
Excellent analysis of crosstalk! (full text below). The way that you explain
it also clarifies other effects of supply impedance that had been puzzling me.
This is much more useful than the deeply technical discussion in RDH4 that I
had been trying to work from. Thanks for taking the time to put this in clear
terms that will be useful for seat of the pants evaluations.
Regards
Ross H
In article <jbyrns-0607...@207-229-173-212.d.enteract.com>,
jby...@enteract.com (John Byrns) writes: