McIntosh Output Transformers
Eric
price, would hobby tube amp builders buy them?
Gene E. McCluney
<tiger...@hotmail.com> wrote:
> I am curious, if these transformers were available at a reasonable
> price, would hobby tube amp builders buy them?
The key here is "reasonable price", which is doubtful, since they are
very specialised with bifilar and in some cases trifilar windings on
the plate sides.
Gene McCluney
Frank Pierce Harrison
Gary Gerhart
> I am curious, if these transformers were available at a reasonable
> price, would hobby tube amp builders buy them?
>
I am curious, what do have up your sleeve? Are you planning to
unleash a secret stash of Mac OPTs or are you planning to build some?
If they are real Macs, then the question is, "what is a *reasonable*
price?" If they are reproductions then quality, construction,
appearance, etc. creep into the equation.
So, what's the scoop?
--
Gary Gerhart
www.GerhartAmps.com
RonSonic
>I am curious, if these transformers were available at a reasonable
>price, would hobby tube amp builders buy them?
There was a company selling reproductions of some of these. Sterling
Audio? Something like that. Pretty decent prices if I recall. Check
back issues of Glass Audio for the ads.
The Macs used some fairly unusual circuitry - tertiary cathode
windings and other craziness that doesn't much appeal to the homebrew
crowd. With those trannies you would build something like a replica of
a classic McIntosh, or set out on a fairly significant R&D project.
Still, it would be well to see them available. Just don't know what
sort of market there is.
Ron
Shiva
"Gary Gerhart" <Ga...@Gerhartamps.com> wrote in message
news:3D40A443...@Gerhartamps.com...
-dim
Steve O'Neill
Eric wrote in message ...
>I am curious, if these transformers were available at a reasonable
>price, would hobby tube amp builders buy them?
What type of Mac OPT do you speak of???
As I recall, Mac used two basic types of transformers: the classic Mac unity
coupled opt and a somewhat more conventional type that had cathode feedback
winding that wasn't in the same ratio as primary so wasn't considered unity
coupled. Lore has it that the classic Mac unity coupled opt employed bi or
tri filar primary windings using double insulated magnet wire AND double "C"
core and was usually potted. This style opt was used in their MC series of
mono and stereo amps among others. The other OPT used more conventional E-I
core, wasn't potted and showed up in Mac's downscale offerings such as the
Mac 1500 receiver.
--
Steve
Jerry G.
expensive to manufacture, and distribute. You can contact Mcintosh
directly for their parts. They do not use aftermarket dealers. The
dealer would go to them for the same parts.
--
Greetings,
Jerry Greenberg GLG Technologies GLG
==============================================
WebPage http://www.zoom-one.com
Electronics http://www.zoom-one.com/electron.htm
Instruments http://www.zoom-one.com/glgtech.htm
==============================================
"Eric" <tiger...@hotmail.com> wrote in message
news:c87b98d8.02072...@posting.google.com...
Gary Gerhart
>>
> Me too... Your site - nothing but dead links... What's the scoop?
> -dim
Well, it's a long boring story -- don't worry, I'll spare you the drama.
Short side of it is, it's soon to be completed...
--
Gary Gerhart
www.GerhartAmps.com
2B
R
"Gene E. McCluney" <mccl...@earthlink.net> wrote in message
news:250720021825594484%mccl...@earthlink.net...
Philip Lawrence
There was a US. firm making the Transformer for the MC30
Phil
I made two amps with them using 809's
William Waters
If you are asking about the vintage ones for tube gear, if they were
available even at an unreasonable price, many would still buy them.
Sadly, the ability to build them as originally done has been lost. Bill
Anton Elron
while. They had hired at least one person who had worked at Mac (Mc is
in fact a "starter" or low end employer in that area: you started
there and went to Link or IBM when you had experience
traditionally)on the wind line and Mac alleged patent and trade secret
infringement and, the company just dropped the product rather than
fight. (The Southern Tier NY area is pretty tight-knit among
electronics mfr's.)
The patents have long since run out.
Other unity coupled transformers have and do exist. EAR, Nestorovic
and even Quicksilver have all marketed commercial amplifiers with this
output configuration. Plitron sells a toroid output transformer wired
for this type of operation although it is certainly not mechanical fit
in a Mac amp and while it might be electrically serviceable a toroid
is going to present problems meeting spec without circuit changes.
A one man wind shop in Wisconsin offers to wind these and claims to
have copies of Mc wind and test procedure sheets. He wants a lot of
money and I doubt he sells any except as replacements for Macs.
I seriously doubt whether any of the Mac outputs were as expensive to
make as a top quality conventional unit. There is no lost secret mojo
involved, in fact everything you need to know is sitting in college
libraries: A student engineer built a "McIntosh" of his own with a
trifilar primary in 1956. His performance numbers weren't in the Mac
class, but I suspect he didn't get his transformers (he used a bifilar
driver too as in the pre-MC30 Macs)very well set up corewise or may
have poorly selected his cores.
To answer Tigerbomb77's question though, the answer is probably no,
because most of the amateur tube constructors out there are not
really capable of comprehending the Mac circuit, much less the
Bereskin or De Paravicini derivatives.The longer I fool with old
electrojunk and the people who do, the more I am convinced that the
whole single ended open loop triode thing is popular only for its
intellectual challengelessness.
Tigerbomb? Must be a Debhead too.
Anton Elron
> coupled opt and a somewhat more conventional type that had cathode feedback
> winding that wasn't in the same ratio as primary so wasn't considered unity
> coupled. Lore has it that the classic Mac unity coupled opt employed bi or
> tri filar primary windings using double insulated magnet wire AND double "C"
> core and was usually potted. This style opt was used in their MC series of
> mono and stereo amps among others. The other OPT used more conventional E-I
> core, wasn't potted and showed up in Mac's downscale offerings such as the
> Mac 1500 receiver.
The MA230 used a "more conventional" transformer with cathode
windings which was the genesis for Bruce Rozenblitt's writing in Glass
Audio and his company. I think he found a case of them somewhere. They
were built by an outside winder and were not unity-coupled. The 1500
also used a non-Mac wound output transformer but I don't know if they
are different.
The "real" Mac amplifiers all used the unity-coupled output stage
which, contrary to the sage future advice of Don Lancaster, they
extensively patented. (Didn't seem to harm them any.) Since no U.S.
patent can last for more than 34 years, any patent granted after July
of 1968 would be now out of force. The key patents were all granted
before 1960 apparently, to judge from Roger Russell's site and a
search at www.uspto.gov.
The key to McIntosh's geat success was their high sale price, dealer
discipline, advertising, cosmetics, and their ability to meet tight
self-imposed specs while-an aspect of Mc products not often
discussed-relatively low build costs.
Bruce Morgen
Another aspect of their marketing
success was the "we'll fix it
free on the spot if we have the
parts with us" feature of their
traveling amp testing clinics.
They stopped doing those when
some of the better Japanese amps
started measuring up to theirs
(on their own test gear!) at a
fraction of their prices.
Tom Bavis
Technophool page. http://www.audiophool.cjb.net/Techno.htm
Not a great scan since it's from a copy. If my friend can turn up the
original, I'll rescan it.
Tom Bavis
Robert Baer
>
> I am curious, if these transformers were available at a reasonable
> price, would hobby tube amp builders buy them?
You would have to know exactly the method of winding, the screen tap
point, and the cathode winding turns as well, along with the core
lamination material and thickness grade(s) used for optimum
reproduction.
Anton Elron
> Eric:
> If you are asking about the vintage ones for tube gear, if they were
> available even at an unreasonable price, many would still buy them.
> Sadly, the ability to build them as originally done has been lost. Bill
Are you high?
Mac for one could still build them inasmuch as they still use the
original winding machine Frank McIntosh and Sid Corderman cobbled up
and many of the same low paid female assemblers are still there.
However, others can and have built these tranformers. What's lacking
is the business case for doing so.
RonSonic
wrote:
Bingo.
Last year needed parts to resurrect some severely beat 2300s - was
delighted to find that McIntosh could supply populated driver boards.
Was disgusted to find NTE and ECG junk on these poorly assembled
boards when they arrived. That's something new to me, rebuilding
McIntosh stuff to get it right. Yuck. Called them on their component
selection, they insisted that the parts used all met their very
demanding specifications which remain as high as ever. Yeah, right.
Ron
Rich Andrews
news:j8c5kuon9alt74a1d...@4ax.com:
I believe that the Clarion buyout has everything to do with that. If one
reads Roger Russell's site, you can see that the Clarion mismanagement
caused all manner of problems (and apparantly still does). What Roger
doesn't say in the following paragraph says volumes. Keep in mind this was
after Clarion had made wholesale changes to speaker designs.
"The LS310, LS330 and LS350 go into production. These systems are based on
the XL1, XL10 and XR250 that Carl and I had designed, but incorporate the
driver and cabinet changes. The cabinet assembly and finish is excellent."
The only thing positive he can say about the speakers is in regard to the
cabinets? Roger obviously did not approve of the design and probably not
the performance either.
To save bandwidth, it might be helpful to read the page
http://www.sundial.net/~rogerr/lsd2.htm starting with the year 1990.
I can understand why Roger left McIntosh. It makes no sense to stay where
one's sound engineering decisions are countermanded by the marketing dept.
or the art dept.
About the only new speaker system I would buy from Mac now would be the XR-
290 if it is still available.
r
--
"I will not be pushed, filed, stamped, indexed, briefed, de-briefed, or
numbered...My life is my own."
"I am not a number. I am a free man."
No. 6
Robert Baer
Also, Dynaco made some good copies.
Robert Baer
Piss around all you want; NTE and ECG transistors and ICs are made the
same way as Fairchild, Signetics, National, etc & etc make them.
RHersh
first that I have heard.
Ross Hershberger
(Listening to Ron Sexsmith on MC30s and Spica TC60s)
RonSonic
Sure, and Orion and National branded vacuum tubes are made the same
way as Mullard, RCA and GE.
Perhaps this is one of those those things where a statistically unfair
sample causes a bad impression, that lasts longer than is fair. The
other stuff doesn't cost that much more and I am handy with
substitution and spec books. The difference in parts cost is nothing
compared to the labor costs. I'll stick with the premium stuff.
I'll be the first to admit that I am not the most knowledgeable solid
state guy, but I work with some and trust their experience and my own.
Your mileage obviously varies - no problem.
BTW, those boards needed to be reworked, transistor bodies were
touching traces they shouldn't have been.
Ron
Anton Elron
> You would have to know exactly the method of winding, the screen tap
> point, and the cathode winding turns as well, along with the core
> lamination material and thickness grade(s) used for optimum
> reproduction.
You are obviously putting the wrong kind of poppy seeds on your
burger buns. Since we offed the Taliban , who despite their faults
nearly ended smack production, farmers all over Afghanistan are
looking really hard for those seeds now...
Anton Elron
> point, and the cathode winding turns as well, along with the core
> lamination material and thickness grade(s) used for optimum
> reproduction.
The McIntosh transformer does not use screen taps, there are plate
and cathode windings which are identical for AC but very different for
DC,and they are wound bifilarly. Almost all the old ones were wound on
C-cores. Exactly which cores were used, the number of sections used,
and the number of turns and the gauge and insulation quality are
unknowns but all these things will be determinable from a teardown.
Turns ratio and DC resistance are easily measurable as are primary and
leakage inductance and interwinding capacitance.
Please, take a moment to look at a schematic of any of the popular
Mac power amps.
Eric
have screen windings which are cofilar with the plate and cathode
windings. The Lockheart amplifier which precedes the MC75 in having a
trifilar primary uses one of the windings for the screen also.
Not to make you jealous, I had lunch with the baroness last week.
She's the girl every boy wants to be, isn't she?
Robert Baer
Perhaps ther are 2 versions; one with screen taps and one without.
The screen tap version appears to be superior.
DC resistance is not relevant; wiring method (ie: bifilar, "pi", etc)
is very important, and wire size is useful so that the number of turns
(useful knowledge) may be reproduced in the space given.
Turns ratio can be determined only by taking the transformer apart and
counting turns of one winding, OR adding a known number of turns and
making appropiate measurements.
Robert Baer
...Now tell me, *why* would a boy want to be a girl (or vice-versa)????
Anton Elron
> She's the girl every boy wants to be, isn't she?
>
At least I'd have _known_ what to do with her...
CompUser
"Robert Baer" wrote ...
> Turns ratio can be determined only by taking the transformer apart and
> counting turns of one winding, OR adding a known number of turns and
> making appropiate measurements.
I thought you could calculate the turns ratio by applying a known
low voltage (say, 6.3 volts) to one side and checking the other
side to see what the induced voltage was...no?
Steve
Anton Elron
<snip>
>
> Perhaps ther are 2 versions; one with screen taps and one without.
> The screen tap version appears to be superior.
> DC resistance is not relevant; wiring method (ie: bifilar, "pi", etc)
> is very important, and wire size is useful so that the number of turns
> (useful knowledge) may be reproduced in the space given.
> Turns ratio can be determined only by taking the transformer apart and
> counting turns of one winding, OR adding a known number of turns and
> making appropiate measurements.
You are a bolt. Where did you get your license to drive on Usenet?
There ARE NO screen "taps" on any McIntosh unity coupled transformer.
None. There can't be. Your only options are to have a separate screen
winding, cross couple the screens to the opposite side's plate, or
triode-connect them.
All transformers operate on the physical law that their voltage ratio
is directly a function of their turns ratio. Therefore if you put a
known AC voltage across any winding, the voltages across the others
will be equal to the turns ratio between the windings, as long as
suficient current is available to excite the transformer and minus the
voltage drop from currents dropped through the other windings (which
there is none if the transformer is "phased out" open circuit.)
A boy who wants to be a girl is called "gay". (That's an
oversimplification but it will do for now.)
Now go study for your learner's permit before the keyboard police
find you out.
Fred Nachbaur
Anton Elron wrote:
>>[...]
>
>
>
> You are a bolt.
Bolt? You mean, as in the complement to a nut? Or did you mean to say
"dolt"?
> Where did you get your license to drive on Usenet?
Comes with the computer. Also available in Cracker-jack.
Cheers,
Fred
--
+--------------------------------------------+
| Music: http://www3.telus.net/dogstarmusic/ |
| Projects, Vacuum Tubes & other stuff: |
| http://www.dogstar.dantimax.dk |
+--------------------------------------------+
Steve O'Neill
Anton Elron wrote in message ...
<snip>
>There ARE NO screen "taps" on any McIntosh unity coupled transformer.
>None. There can't be. Your only options are to have a separate screen
>winding, cross couple the screens to the opposite side's plate, or
>triode-connect them.
I'm sure the person you're responding to has confused the cross coupled
screen arrangement with screen taps of partial triode operation. Some of
Mac's schematics are drawn such that it's difficult to glance at the
representation of the opt and see what's really been drawn. Kind of like
some tube tester schematics with 100 parallel lines from sockets. Seems
that the unity coupled scheme is easy to misinterpret on schematics. I've
seen posts that have referred to the cathode windings as a cathode choke.
Patience and tolerance might be warranted in cases like these
<snip>
> A boy who wants to be a girl is called "gay". (That's an
>oversimplification but it will do for now.)
>
> Now go study for your learner's permit before the keyboard police
>find you out.
Well, this WAS Xposted to alt.guitar.amps ;-)
--
Steve
Robert Baer
So, doing it your way gets (say) 5:1 ... insufficent info.
How *many* turns primary, how *many* turns secondary, etc is what is
useful.
Robert Baer
1)
"cannot be screen taps" - really? Never heard of the Williamson
circuit?????
Or how about the Crowhurst circuit?
2)
Turns ratio is useless when trying to copy a transformer. one needs to
know how *many* turns were used (and winding method for critical uses).
3)
Do you know, on the PC, the differnece between a male mouse and a
female mouse?
Robert Baer
1)
"cannot be screen taps" - really? Never heard of the Williamson
circuit?????
Or how about the Crowhurst circuit?
2)
Turns ratio is useless when trying to copy a transformer. one needs to
know how *many* turns were used (and winding method for critical uses).
3)
Do you know, on the PC, the difference between a male mouse and a
female mouse?
Robert M. Braught
"Insufficient info" for *what*? Experimentation? No. Practical
application of surplus parts obviously designed for a particular service
(audio output, line driver, power transformers)? No. The 'black box'
dynamic test Steve mentioned is completely useful, albeit limited. FWIW
we could get out our sweep function generator, lab amp, programmable
load, etc and find out 'more' (but 'more' may have no real value in our
'one off' design using the odd xfmr! ;-)
"Insufficient info" for gnat's ass reverse engineering intended for
drawing up plans to 'copy' a transformer? Yes, undoubtedly.
-Robert
QTS
http://www.Braught.com
RonSonic
There's a reason the McIntosh circuit has its own name - No screen
taps.
>2)
> Turns ratio is useless when trying to copy a transformer. one needs to
>know how *many* turns were used (and winding method for critical uses).
Yes the turn count is important, but that isn't what you said. You
said turns ratio.
>3)
> Do you know, on the PC, the differnece between a male mouse and a
>female mouse?
Fire away, I'll cue the drummer for a rim shot.
Ron
Richard D Pierce
Turns ratio can be determined only by taking the transformer
apart and counting turns of one winding, OR adding a known
number of turns and making appropiate measurements.
Notice that Mr Baer SPECIFICALLY said "TURNS RATIO can only
be determined, ..."
Then CompUser wrote:
"I thought you could calculate the turns ratio by applying a
known low voltage (say, 6.3 volts) to one side and checking
the other side to see what the induced voltage was...no?
Now, just from the viewpoint of someone who HAS done transformer
design and testing, this statement is correct: the voltage ratio
between primary and secondary is the same as the turns ratio.
TO that, Mr Baer retorts:
"So, doing it your way gets (say) 5:1 ... insufficent info.
How *many* turns primary, how *many* turns secondary, etc is
what is useful.
But that's NOT what Mr. Baer said: Mr. Baer specifically stated
that trusn ratio could not be determined without disassembling
the transformer and counting turns. And that statement is false.
Further, I'd bet that I can wind a transformer core that, if Mr.
Baer were to disassemble it and count the turns, he'd get the
turns ratio wrong, where a measurement of the ratio of input
voltage and output voltage would get it right (excercise for the
student: name at least two ways how this could be done).
Further, it's QUITE possible to get MUCH more information that
Mr. Baer supposes about the design details of a transformer,
including number of turns (or a damned close approximation
thereof) without disassembling the transformer. We used to do it
all the time. The chief engineer I worked for could take a
competitor's transformer and inside a half hour, clone it
without opening it up.
For example, want to know how many turns ratio? Easy, measure
the ratio of the input to output voltage. Gets you to within
about 2% if you're even moderately careful. Want to know the
number of turns on the primary? Easy, expose just enough of the
primary winding, maybe 1/4" to measure the wire gauge. Now
measure the DC resistance. The two allow you to calculate the
length of the primary winding to within a couple percent. Now
measure the physical dimensions of the winding bundle and from
that calculate the number of turns. Multiply by the turns ratio,
you have the secondary. Measure it's gauge and DC resistance and
calculate backwards and see how close you get. Iterate towards
convergence. This guy, just doing simple stuff like that could
design a visual and performance clone to a transformer and the
original remained quite intact. And he'd do this all with the
simplest of test equipment (signal generator, good AC voltmeter,
a micrometer, a pencil, pad of paper and a slide rule.) If he
wanted to getr fancy, he'd do som interwinding capicitance
measurements and bandwidth measurements and come up with a bunch
of other stuff.
But, Mr. Baer, you stated that "turns ratio" required
disassembly of the transformer, and that's just simply false.
You further implied that no useful information could be
ghathered without disassembly, and that is just as false. Maybe
YOU can't get any useful information, but it is there waiting
for anyone with sufficient skill and instrumentation to pick up
at one's leisure.
--
| Dick Pierce |
| Professional Audio Development |
| 1-781/826-4953 Voice and FAX |
| DPi...@world.std.com |
Patrick Turner
Richard that it is possible to measure just about any tranny
to find out most of the characteristics.
Somebody real clever would measure all the leakage inductances
between the windings, and the capacitances to see where in relation
they might be
placed on the wind up.
But maybe it would take longer than a half hour.
One can then confirm the findings by sawing through
the top layers, peeling the layers back, and painstakingly
counting all the turns, and layers,
and how the thing was wound.
I once did this to a Leak OPT to findout more about
what the guys did 40 years ago.
The sample I cut open had shorted turns, which made
measurements impossible, but there were two channels,
and from the working channel, I got the picture.
( I found out the Leak fellas did as little as possible,
there was hardly any interleaving, the wire was far too thin,
and the iron core was too small.)
I recently examined an EAR509 OPT, and found out
all about what was inside the tranny, by just
applying 6.3 volts at 500 Hz to one of the primary windings,
with no load on the amp, (and with it turned off!).
I might add, the EAR tranny isn't simple, it has plate, cathode,
screen, and feedback windings, as well as quite a few layers of
secondaries, in various sections.
I wind all my own PTs, OPTs, and chokes.
Patrick Turner.
Anton Elron
> 1)
> "cannot be screen taps" - really? Never heard of the Williamson
> circuit?????
> Or how about the Crowhurst circuit?
> 2)
> Turns ratio is useless when trying to copy a transformer. one needs to
> know how *many* turns were used (and winding method for critical uses).
> 3)
> Do you know, on the PC, the difference between a male mouse and a
> female mouse?
This guy is threading himself in farther and farther.
First, the Williamson circuit,(by which we mean the D.T.N. Williamson
"Wireless World" circuit and not Reg Williamson's solid state efforts
later),WIRED AS TRIODES the output tubes. It was Hafler and Keroes
that popularized the "Ultra-Linear Williamson" with screen taps.
It also had the cathodes at, or approximately at, DC. The McIntosh
unity coupled circuit does not. It "splits the difference" between the
conventional plate loaded configuration and cathode follower
operation.
Tim Williams
> 3)
> Do you know, on the PC, the differnece between a male mouse and a
> female mouse?
A female mouse has a lot more buttons?
Tim
Serial :( 2-button mouse
--
"Yeah, and I'm not easily impressed. Whoa, a blue car!"
- Homer Simpson
Tony Hwang
My answer, female mouse is fatter and slower, specially the pregnant
one.
And they squeak a lot.
Tony
CompUser
"Robert Baer wrote ...
> > "Robert Baer" wrote ...
> >
> > > Turns ratio can be determined only by taking the transformer apart
and
> > > counting turns of one winding, OR adding a known number of turns and
> > > making appropiate measurements.
> >
> > I thought you could calculate the turns ratio by applying a known
> > low voltage (say, 6.3 volts) to one side and checking the other
> > side to see what the induced voltage was...no?
> >
> > Steve
>
> So, doing it your way gets (say) 5:1 ... insufficent info.
Which is exactly what you said you were looking for...
turns ratio! Sheesh.
Steve
Robert Baer
Never heard of mouse balls?
RonSonic
<rober...@earthlink.net> wrote:
Just never actually used a mouse that didn't have one and it didn't
occur to me. Besides, I was sure those were steers not females.
Ron
Robert Baer
Frenchie say zee optical meece no ball.
Greg Pierce
Chris M
> >> The McIntosh transformer does not use screen taps, there are plate
> >>and cathode windings which are identical for AC but very different for
> >>DC,and they are wound bifilarly. Almost all the old ones were wound on
> >>C-cores. Exactly which cores were used, the number of sections used,
> >>and the number of turns and the gauge and insulation quality are
> >>unknowns but all these things will be determinable from a teardown.
> >>Turns ratio and DC resistance are easily measurable as are primary and
> >>leakage inductance and interwinding capacitance.
> >
Hold the horses......
Many moons ago.....before I did a teardown on a MC-30 OPT...I was able to
make enough measurments to figure the turns and core area of the C-cores
and winding layer arangement....and my calculations were nearly on the
botton when they were confirmed durring tear-down...
The trick is finding peak inductance vs. flux density to nail down the core
material ...the rest just falls into place..
Regards
Chris
Anton Elron
some more detailed info on Mac OPTs. Where does he get the cores? I
think Westinghouse supplied a lot of the early ones. I also think they
had a small air gap which was calibrated with fishpaper (employees
thought it was called fish paper after engineer Larry Fish, was one
frequently repeated story.)
Greg Pierce <greg....@cox.net> wrote in message news:<3D660116...@cox.net>...
> OK, maybe I am a bit daft here - who is the 'Baroness'?
Debbie Harry. She picked this up while singing with Roy Nathanson's
Jazz Passengers.
Patrick Turner
Anton Elron wrote:
> Chris Merren has had some good stuff on here, maybe he will put up
> some more detailed info on Mac OPTs. Where does he get the cores? I
> think Westinghouse supplied a lot of the early ones. I also think they
> had a small air gap which was calibrated with fishpaper (employees
> thought it was called fish paper after engineer Larry Fish, was one
> frequently repeated story.)
The Audio Encyclopedia, by Tremain, 1963,
has some info about McIntosh OPTs.
I think the cores are double C-cores,
like arranged so the windings are like a normal E&I
tranny, and the c-cores make a 00 with the windings
around the join in the 0s.
I have a mosfet PP amp with double c-cores,
and yes, it has a fine layer of plastic film to gap the cores,
but there ain't nothin fishy about it, the mu of the iron
is high, about 5,000, and tended to saturate abruptly at very LF,
so the gapped core tends to smooth out this effect.
The gap does not want to be any larger than it should be,
or else the primary inductance falls too much.
and the bass suffers. There is an art to know what sized gap is required,
but roll your own cigarrette paper might be about right.
I have used GOSS E&I lams made here in Aust, and the mu max was 17,000,
but I found no need to gap the core, and maximum interleaving
of laminations gives the lowest bottom end, and the problem with
the LF saturation in a tube amp didn't seem to matter.
I used rock and roll, and with no input filter on the amp,
so bandwidth was down to about 2 Hz.
Then I wind up the wick, to clipping, with a dummy load,
and sure, there is some knocking / choking sounds coming from the
OPT, and same thing with a pink noise source.
When an input filter is applied, so input F pole is say 13 Hz,
the problem is much less.
At a 1/4 full power, no problem at all.
So whether or not you gap a PP amp core is
not absolutely important.
At very low F, and a high enough applied voltage,
and with a gap, the steel will still saturate.
To get the frequency of saturation lowest, use
plenty of primary turns, but then the HF suffers,
so one must strike a balance between HF extension, LF grunt,
and winding losses, for any given core.
The trouble with OPTs is there are many things to consider together.
But they do allow fabulous clarity, when well wound.
Patrick Turner.
Choky
--
Choky
Prodanovic Aleksandar
choky*remove*@eunet.yu
YU
-------------------------------
remove *remove* to reply!
-------------------------------
"Anton Elron" <anton...@hotmail.com> wrote in message
news:ea82a75b.0208...@posting.google.com...
Greg Pierce
Hmm.. Interesting; I never knew that.
Now I AM jealous :-p
Greg
Chris M
tradename was refered to as "HYPERSIL" ....
The cores were stamped A-28, which meant it was Audio grade, which meant it
was .013" thick lams that were spun-up and cut..
It was also known as C-97...
"Hypersil" was just Westinghouse's tradename for their grain-oriented
silicon steel...Just like today "Magnesil" it is refered to by some
makers..It really boils down to being M6 core material when you do the full
measurements on the core.....
The Mac OPT's were typically 5 primary (in series) and 4 secondaries (in
parallel) and used all the same wire gauge for all the windings..no
interleaving papers...just Quad-formvar wire spun-up... Fish-paper was used
for the bobbin which was just folded...string held the windings to the
bobbin so it could be varnished and baked...
The 8 ohm taps were all really 9 ohms do to the winding geometry..
regards
Chris
RHersh
Very interesting. That dispels some of the mystery and misinformation about
these OPTs. A couple of questions:
1: What is the ratio between the cathode primary winding and plate primary
winding? 1:1 would seem to be implied.
2: What is the design primary impedance?
Thanks
Ross Hershberger
Patrick Turner
RHersh wrote:
The primaries on a McIintosh OPT consist of two
equal turn windings, one for the cathodes, one for the
anodes, and both have a CT.
The cathode winding CT is grounded, and the anode winding CT
is connected to the B+, of say +470 volts.
The mode of operation is mainly class B operation, ie,
the bias current is low, so for much of one sine wave cycle,
only one of the output tubes is conducting, while the other is cut off
by the negative going grid voltage.
During the Class B part of the cycle, the load seen by the tube
doing work is 1/4 of the rated anode to anode load.
In the case of the MC75, the anode to anode load is 4 kOhms,
and so each tube sees 1 k.
The loads on the tubes are no different to what is going on
in a conventional plate loaded amp with the same a-a load,
and same tubes.
So if you have 75 watts into 4 k, then 547 vrms would exist
across the a-a load of 4k in a normal amp,
ie, that would be 547 vrms across the two ends of the primary.
This means 273 vrms occurs at the anode of each output tube.
But with the McI, the part of the primary normally devoted to one of the output
tube's
plate circuit is divided into two, and half used in the anode, half in the
cathode,
so the 273 vrms is further divided by 2 so that +136 vrms is at the anode,
while -136 vrms is at the cathode.
Meanwhile, at the other tube, -136 vrms is at the anode,
and +136 vrms is at the cathode.
From the tube's point of view, it thinks it has exactly the same loading
as the normal amp with just one big anode winding.
In the McI, the fact that you have two windings, with equal turns,
means that the signal voltages for anode and cathode are the same
for both tubes, but just McI connect the tubes so
that the phases of signals are + and - as required.
McI use bifilar wound plate cathode windings to
ensure tight magnetic coupling between the two tubes,
and this minimises crossover distortion effects
caused by lack of coupling found in normal amps.
The bandwidth of the McI output stage is very darn good,
and the guy who designed their OPTs so long ago was no fool,
and the mystique rests in the cleverness which is still regarded
as clever compared to what is done today.
The only reservation I have about the McI OPTs
is that the bifilar windings have such a high direct voltage difference,
held at bay only by the wire insulation, and not by
a layer of insulation between cathode and anode windings.
Layer wound transformers made to the McI recipe
can be done without bifilar windings, and with alternate layer windings,
and can employ polypropylene caps to bypass ends of windings which
would otherwise be only tightly coupled magnetically.
The caps might even do a better job of shunting what low leakage
inductance that may be present.
The fact that the anode and cathode windings are equal means that
the grid input voltage must be greater than the magnitude of the
anode or cathode voltages, so that if say 27 vrms between
cathode and grid are required to produce 273 vrms across the tube to
put the power into the load of 1 k, then 136v + 27v = 163 vrms is required
at the grid.
This equates to an amount of local NFB from a gain reduction
from about 10 to 1.67, = to a gain reduction = of about 6 times,
approximately = 15 dB.
The EAR509 from England uses a similar arrangement
of "unity gain" or equal plate and cathode windings.
In the McIntosh OPT arrangement
the maximum magnetic flux density, Bm, in the OPT is the same
as in a normal amp.
The Bm is generated by a voltage of 273 vrms across N turns,
but the N consists of parallel cathode and anode windings each of N turns,
compared to 547 vrms developed across 2N turns in a normal amp.
So the core size and total number of turns used and turns / volt are
identical to the normal way of building an OPT, to ensure a low
Bm at some low F, to prevent stauration effects.
So, to get back to your original query about impedances,
The tubes see the same 4 k anode to anode load as in any other normal amp.
The issue is confusing, since there are two parallel windings.
If you consider the two windings as being just N turns instead of
2 N turns, then the load for N turns is 1 k.
Have I thrown any light on the issue,
or does it still appear confusing as ever?
The tubes in a normal amp might make 5% or more THD,
but this is reduced by 15 dB, or six times in the McI
output circuit, but then some additional thd is generated
by having to make up to about 140 vrms of drive voltage
to each output tube grid.
The McI uses global NFB to reduce the THD to equal that of other
75 watt amps that you or I might build.
But the McI will have a far lower output impedance
than what we might achieve with only global NFB,
since the total amount of NFB is much higher.
The EAR509 using 2 x PL509 tubes achieves 100 watts,
at similar thd levels as the McI, but since its secondaries
on its OPT are not included in the global loop, the Rout
is much higher than the McI.
In practice, just for myself, I find the idea
of using equal anode and cathode windings in an OPT
are quite unecessary.
I have an amp with 4 x 6550 which has 12.5 % of the total primary
turns in its cathode windings, and it uses only 16 dB of
balanced screen NFB to a pair of EL84 drivers,
and I get 0.07% thd at 80 watts, using a far larger amount of class
A action than the McI.
The drive amp only consists of the one balanced stage using two EL84, not
like EAR509, or McIntosh, with 3 stage drive amps.
The amp is integrated, and uses a line stage longtail pair without any
feedback loops to drive the power amp.
Sould you build such things, whether one sounds better than the other
is a moot point that has been argued by the poets for at least 50 years
so far.
Probably, the McI would be a better amp if it had a lower B+,
higher idle tube current and hence a lot more class A % in its power output.
Maybe discerning listeners might hear the difference.
Patrick Turner.
Chris M
I would have to say that Patrick did a wonderfull job of being detailed in his
explanation..
One thing to note... You mentioned the impedance...
Well the impedance is 1/4 of what it would normally be...
The MC30 is 900 ohms.... which would be like a 3.6K plate load if it had no
cathode winding..
As Patrick stated...a Mac OPT has same number of turns on the primary..only that
it's distrubuted 50% to the anodes and 50% to the cathodes.., thus 100% feedback
and UNITY coupling.....
Regards
Chris
Chris M
Mac's with 6L6 valves (MC-30, MC-40) had 900ohm OPT's....that would be
equivalent to a 3.6K plate load with no cathode loading..
Mac's with 6550 valves (MC-60, MC-75) had 575 ohm OPT's..which would be
equivalent to a 2.3K plate load with no cathode loading..
The early Mac stuf like the 20w-2 had 6V6 valves and used a 1250 ohm
OPT...equivalent to using 5K plate load if no cathode loading ...
The 50W amp had a 1000 ohm OPT....which would be like a 4K plate loaded OPT
with no cathode winding...
Regards
Chris
Patrick Turner
Chris M wrote:
Thanks Chris, I took 4 k as the value of load actually seen
by the two output tubes.
Obviously the McI amps have a fair old range of load
rated impedances for the tubes to work into.
In the real world, rated impedances are a bit meaningless,
since speaker loads vary widely for a given nominal impedance.
And so does the value of load seen by the tubes.
Some particular speaker load might cause the load seen by the tubes to be
4 k a-a at one frequency, but away from this F the load seen by the tubes
might vary from 2 k to 50 k.
And with a lot of class B content in the amp's working,
there is a high tendency for a pretty awfull amount of IMD to be produced
by the amp.
However, McI's use of a fairly huge amount of NFB is supposed
to reduce all this back down to inaudible levels.
The process is catching the horse after it has bolted.
And we see the horse walking, when in fact it is having to gallop.
I can't help saying that amps which are linear even without
any NFB do attract my attention better.
Then hardly any NFB needs to be applied,
and the amp is simpler in its concept, even if more tubes
need to be used.
Let's face it, the company accountants know loops of NFB
are cheaper than tubes!
Patrick Turner.