[John Byrns]

Some years back there was considerable discussion in this group of a series of
articles, which IIRC were said to have been originally published in the British
magazine Wireless World, which touted the advantages of British amplifier design
over US amplifier design. Specifically, IIRC, British design was represented by
the "Acoustical" output circuit, while US design was represented by the
"Ultralinear" output circuit. These articles seemed to focus mainly on the
shortcomings of the "Ultralinear" circuit and the advantages of the "Acoustical"
circuit.

As Patrick has pointed out, I have gone senile, as a result I can't remember the
Authors or Titles of these articles, can anyone refresh my memory?

The above is just idle curiosity, what I am really interested in is an article
that I believe was connected with the above mentioned series of articles, if not
one of them, which described the requirements on the design of Output
Transformers for the "Ultralinear" circuit. The focus was on the requirements
that were necessary to prevent the output stage from becoming a giant high power
oscillator as a result of leakage inductance between the plate and screen
sections of the OPT causing phase shifts changing the plate to screen grid
feedback from negative to positive.

I wish to find and read this article again, is anyone familiar the article I am
describing? I think I may have a copy of the article on my computer, however I
haven't a clue how to locate it among the hundreds of thousands of files that I
have squirreled away over the years.

Any help identifying and locating this article would be greatly appreciated?

--
Regards,

John Byrns

Surf my web pages at, http://fmamradios.com/

[Andre Jute[_2_]]

Mmm. I think the article you want may be this, or in reply to it:

"Amplifiers and Superlatives " by D.T.N. Williamson and P.J. Walker,
Wireless World, September 1952

I can let you have the article by Hafler and Keroes which started the
kerfuffle, Audio Engineering, November 1951: "An Ultra-linear
Amplifier". Write to me at andrejute at coolmainpress with the
commercial extension so I can get an address to send it to.

Andre Jute
Visit Jute on Bicycles at
http://coolmainpress.com/BICYCLING.html


On Jul 10, 6:12*pm, John Byrns wrote:
Some years back there was considerable discussion in this group of a series of
articles, which IIRC were said to have been originally published in the British
magazine Wireless World, which touted the advantages of British amplifier design
over US amplifier design. *Specifically, IIRC, British design was represented by
the "Acoustical" output circuit, while US design was represented by the
"Ultralinear" output circuit. *These articles seemed to focus mainly on the
shortcomings of the "Ultralinear" circuit and the advantages of the "Acoustical"
circuit.

As Patrick has pointed out, I have gone senile, as a result I can't remember the
Authors or Titles of these articles, can anyone refresh my memory?

The above is just idle curiosity, what I am really interested in is an article
that I believe was connected with the above mentioned series of articles, if not
one of them, which described the requirements on the design of Output
Transformers for the "Ultralinear" circuit. *The focus was on the requirements
that were necessary to prevent the output stage from becoming a giant high power
oscillator as a result of leakage inductance between the plate and screen
sections of the OPT causing phase shifts changing the plate to screen grid
feedback from negative to positive.

I wish to find and read this article again, is anyone familiar the article I am
describing? *I think I may have a copy of the article on my computer, however I
haven't a clue how to locate it among the hundreds of thousands of files that I
have squirreled away over the years.

Any help identifying and locating this article would be greatly appreciated?

--
Regards,

John Byrns

Surf my web pages at, *http://fmamradios.com/

[John Byrns]

In article ,
Andre Jute wrote:

Thanks Andre, yes those are the articles that started the controversy. Both
articles are available on the web, and I think I have the November 1951 issue of
Audio Engineering in my library.

The article I am trying to remember, and find, specifically dealt with the
technical problems of building an output transformer for the Ultralinear
circuit. As I remember it the basic thrust was that it wasn't as simple as some
people apparently thought. I assume this could have been a response to people
that may have written in saying that the Ultralinear circuit could use a simpler
output transformer than either the Acoustical or Williamson circuits. The
article had the flavor of debunking a notion like that, as it went into
considerable detail on the complexity required in the transformer to prevent the
output stage from oscillating due to the plate to screen grid feedback becoming
positive.

Mmm. I think the article you want may be this, or in reply to it:

"Amplifiers and Superlatives " by D.T.N. Williamson and P.J. Walker,
Wireless World, September 1952

I can let you have the article by Hafler and Keroes which started the
kerfuffle, Audio Engineering, November 1951: "An Ultra-linear
Amplifier". Write to me at andrejute at coolmainpress with the
commercial extension so I can get an address to send it to.

Andre Jute
Visit Jute on Bicycles at
http://coolmainpress.com/BICYCLING.html


On Jul 10, 6:12*pm, John Byrns wrote:
Some years back there was considerable discussion in this group of a series
of
articles, which IIRC were said to have been originally published in the
British
magazine Wireless World, which touted the advantages of British amplifier
design
over US amplifier design. *Specifically, IIRC, British design was
represented by
the "Acoustical" output circuit, while US design was represented by the
"Ultralinear" output circuit. *These articles seemed to focus mainly on the
shortcomings of the "Ultralinear" circuit and the advantages of the
"Acoustical"
circuit.

As Patrick has pointed out, I have gone senile, as a result I can't
remember the
Authors or Titles of these articles, can anyone refresh my memory?

The above is just idle curiosity, what I am really interested in is an
article
that I believe was connected with the above mentioned series of articles,
if not
one of them, which described the requirements on the design of Output
Transformers for the "Ultralinear" circuit. *The focus was on the
requirements
that were necessary to prevent the output stage from becoming a giant high
power
oscillator as a result of leakage inductance between the plate and screen
sections of the OPT causing phase shifts changing the plate to screen grid
feedback from negative to positive.

I wish to find and read this article again, is anyone familiar the article
I am
describing? *I think I may have a copy of the article on my computer,
however I
haven't a clue how to locate it among the hundreds of thousands of files
that I
have squirreled away over the years.

Any help identifying and locating this article would be greatly
appreciated?

--
Regards,

John Byrns

Surf my web pages at, http://fmamradios.com/

[John L Stewart]

Regards,

John Byrns

Surf my web pages at, http://fmamradios.com/[/quote]

Crowhurst had similar concerns & mentions that in some of his writings but doesn't get into the nuts & bolts of it. His concerns are at the ultrasonic frequencies. However I've not seen the article you have referenced.

I've never had those problems at all. For example the 6LU8 running SEUL thru a Hammond 125E has got to be a very low cost & no complications at all.

Cheers, John

[Patrick Turner]

On Jul 11, 9:54*am, John Byrns wrote:
In article ,
*Andre Jute wrote:

Thanks Andre, yes those are the articles that started the controversy. *Both
articles are available on the web, and I think I have the November 1951 issue of
Audio Engineering in my library.

The article I am trying to remember, and find, specifically dealt with the
technical problems of building an output transformer for the Ultralinear
circuit. *As I remember it the basic thrust was that it wasn't as simple as some
people apparently thought. *I assume this could have been a response to people
that may have written in saying that the Ultralinear circuit could use a simpler
output transformer than either the Acoustical or Williamson circuits. *The
article had the flavor of debunking a notion like that, as it went into
considerable detail on the complexity required in the transformer to prevent the
output stage from oscillating due to the plate to screen grid feedback becoming
positive.





Mmm. I think the article you want may be this, or in reply to it:

"Amplifiers and Superlatives " by D.T.N. Williamson and P.J. Walker,
Wireless World, September 1952

I can let you have the article by Hafler and Keroes which started the
kerfuffle, Audio Engineering, November 1951: "An Ultra-linear
Amplifier". Write to me at andrejute at coolmainpress with the
commercial extension so I can get an address to send it to.

Andre Jute
Visit Jute on Bicycles at
*http://coolmainpress.com/BICYCLING.html

On Jul 10, 6:12*pm, John Byrns wrote:
Some years back there was considerable discussion in this group of a series
of
articles, which IIRC were said to have been originally published in the
British
magazine Wireless World, which touted the advantages of British amplifier
design
over US amplifier design. *Specifically, IIRC, British design was
represented by
the "Acoustical" output circuit, while US design was represented by the
"Ultralinear" output circuit. *These articles seemed to focus mainly on the
shortcomings of the "Ultralinear" circuit and the advantages of the
"Acoustical"
circuit.

As Patrick has pointed out, I have gone senile, as a result I can't
remember the
Authors or Titles of these articles, can anyone refresh my memory?

The above is just idle curiosity, what I am really interested in is an
article
that I believe was connected with the above mentioned series of articles,
if not
one of them, which described the requirements on the design of Output
Transformers for the "Ultralinear" circuit. *The focus was on the
requirements
that were necessary to prevent the output stage from becoming a giant high
power
oscillator as a result of leakage inductance between the plate and screen
sections of the OPT causing phase shifts changing the plate to screen grid
feedback from negative to positive.

I wish to find and read this article again, is anyone familiar the article
I am
describing? *I think I may have a copy of the article on my computer,
however I
haven't a clue how to locate it among the hundreds of thousands of files
that I
have squirreled away over the years.

Any help identifying and locating this article would be greatly
appreciated?

--
Regards,

John Byrns

Basically, nearly everyone involved in the amplifier industry was
always on the lookout for some way of getting right away from
Williamson's ideas about interleaving P&S windings in OPTs. The OPT
was a difficult item to have wound properly, needing well practised
and skilled tradesmen and trades-women who of course cost money and
were well unionised.

So there was this tendency to believe the screen taps were all you
needed to use with an OPT with a P-S-P simple interleaving pattern,
and suddenly the phase errors and shortcomings of such a ****ing awful
winding config would somehow dissapear. Anyway, the FEW who didn't
charge along with the uneducated mob of wannabe amp makers realised
the UL connection offered no free lunch and so for 2 x KT66, best
results could only be had if Willy's ideas were fully adhered to. But
after WW2, demand mushroomed and quality nose dived and the rest is a
history of tricks and compromises played on unsuspecting customers.

Quad tried to keep true to Willy, but the Quad-II OPT is nowhere near
as good as it could have been, had it been designed with more iron,
slightly less turns but of thicker wire, etc, etc, as my prvious
recent posts on that issue reveal. Quad did have good quality control
though, as a separate issue to the actual quality, which was mediocre
at best. That's for sure, and Quad suited lotsa ppl. But I have found
a UL amp with KT66 can work just as well as Quad-II, you just need to
apply more global NFB, and install better critical damping than Quad
ever wanted to, but didn't, because such things cost a penny extra.
Quad stuck to their idea because it was a feature which was
marketorrially exploitable, and it didn't matter if most ppl didn't
understand the operating principles or 1951 magazine arguments. Jus'
think, Walker in tweed suit, puffing a pipefull of smoke, wording on
with elegant incoherence about amps et all......anyone listening, like
the BBC dudes were bowled over real easy.

Patrick Turner.

[Estee Eff]

On Jul 11, 2:12*am, John Byrns wrote:
Some years back there was considerable discussion in this group of a series of
articles, which IIRC were said to have been originally published in the British
magazine Wireless World, which touted the advantages of British amplifier design
over US amplifier design. *Specifically, IIRC, British design was represented by
the "Acoustical" output circuit, while US design was represented by the
"Ultralinear" output circuit. *These articles seemed to focus mainly on the
shortcomings of the "Ultralinear" circuit and the advantages of the "Acoustical"
circuit.

As Patrick has pointed out, I have gone senile, as a result I can't remember the
Authors or Titles of these articles, can anyone refresh my memory?

The above is just idle curiosity, what I am really interested in is an article
that I believe was connected with the above mentioned series of articles, if not
one of them, which described the requirements on the design of Output
Transformers for the "Ultralinear" circuit. *The focus was on the requirements
that were necessary to prevent the output stage from becoming a giant high power
oscillator as a result of leakage inductance between the plate and screen
sections of the OPT causing phase shifts changing the plate to screen grid
feedback from negative to positive.

I wish to find and read this article again, is anyone familiar the article I am
describing? *I think I may have a copy of the article on my computer, however I
haven't a clue how to locate it among the hundreds of thousands of files that I
have squirreled away over the years.

Any help identifying and locating this article would be greatly appreciated?

--
Regards,

John Byrns

Surf my web pages at, *http://fmamradios.com/

Hi John,
I believe that the article that you are thinking of is by D M Leakey &
R M Gilson. If you Google for "gilson wireless world" follow the link
to Douglas Self's site.
Regards,
Ian.

[John Byrns]

In article ,
Estee Eff wrote:

On Jul 11, 2:12*am, John Byrns wrote:
Some years back there was considerable discussion in this group of a series
of
articles, which IIRC were said to have been originally published in the
British
magazine Wireless World, which touted the advantages of British amplifier
design
over US amplifier design. *Specifically, IIRC, British design was
represented by
the "Acoustical" output circuit, while US design was represented by the
"Ultralinear" output circuit. *These articles seemed to focus mainly on the
shortcomings of the "Ultralinear" circuit and the advantages of the
"Acoustical"
circuit.

As Patrick has pointed out, I have gone senile, as a result I can't
remember the
Authors or Titles of these articles, can anyone refresh my memory?

The above is just idle curiosity, what I am really interested in is an
article
that I believe was connected with the above mentioned series of articles,
if not
one of them, which described the requirements on the design of Output
Transformers for the "Ultralinear" circuit. *The focus was on the
requirements
that were necessary to prevent the output stage from becoming a giant high
power
oscillator as a result of leakage inductance between the plate and screen
sections of the OPT causing phase shifts changing the plate to screen grid
feedback from negative to positive.

I wish to find and read this article again, is anyone familiar the article
I am
describing? *I think I may have a copy of the article on my computer,
however I
haven't a clue how to locate it among the hundreds of thousands of files
that I
have squirreled away over the years.

Any help identifying and locating this article would be greatly
appreciated?


Hi John,
I believe that the article that you are thinking of is by D M Leakey &
R M Gilson. If you Google for "gilson wireless world" follow the link
to Douglas Self's site.
Regards,
Ian.

Thanks Ian,

That is exactly the article I was looking for! Now to read it through and try
to understand it.

--
Regards,

John Byrns

Surf my web pages at, http://fmamradios.com/

[Patrick Turner]

On Jul 13, 11:47*pm, John Byrns wrote:
In article ,
*Estee Eff wrote:





On Jul 11, 2:12*am, John Byrns wrote:
Some years back there was considerable discussion in this group of a series
of
articles, which IIRC were said to have been originally published in the
British
magazine Wireless World, which touted the advantages of British amplifier
design
over US amplifier design. *Specifically, IIRC, British design was
represented by
the "Acoustical" output circuit, while US design was represented by the
"Ultralinear" output circuit. *These articles seemed to focus mainly on the
shortcomings of the "Ultralinear" circuit and the advantages of the
"Acoustical"
circuit.

As Patrick has pointed out, I have gone senile, as a result I can't
remember the
Authors or Titles of these articles, can anyone refresh my memory?

The above is just idle curiosity, what I am really interested in is an
article
that I believe was connected with the above mentioned series of articles,
if not
one of them, which described the requirements on the design of Output
Transformers for the "Ultralinear" circuit. *The focus was on the
requirements
that were necessary to prevent the output stage from becoming a giant high
power
oscillator as a result of leakage inductance between the plate and screen
sections of the OPT causing phase shifts changing the plate to screen grid
feedback from negative to positive.

I wish to find and read this article again, is anyone familiar the article
I am
describing? *I think I may have a copy of the article on my computer,
however I
haven't a clue how to locate it among the hundreds of thousands of files
that I
have squirreled away over the years.

Any help identifying and locating this article would be greatly
appreciated?

Hi John,
I believe that the article that you are thinking of is by D M Leakey &
R M Gilson. If you Google for "gilson wireless world" follow the link
to Douglas Self's site.
Regards,
Ian.

Thanks Ian,

That is exactly the article I was looking for! *Now to read it through and try
to understand it.

--
Regards,

John Byrns

Surf my web pages at, *http://fmamradios.com/- Hide quoted text -

- Show quoted text -

I recall reading such an article in abour 1994 when I was making
prototype OPTs for future amps I wanted to build. Unfortunately
Leakey&Gilson have very little to offer. Williamson had beaten them
many years before to get to good design principles. Notice the
ponderous way L&G proceed to propose reverse wound coils and so forth
but ending up with the obvious simplest method with interleaving of
1/4S - 1/2P -1/2S -1/2P -1/4S.
This was used by countless makers of low to medium quality affordable
amps, and as everyone srious about home built audio should know, never
copy what a major manufacturer does because half the design is ruined
by accountants, or an engineer worried about costs.

In 1956, it was very common to have 15 ohm and 3.75 ohm load matchses
with secs in series or parallel. There was rarely any attempt for
better load matching possiblities. Cost too much. Unecessary, etc. Oh
yeah?
Was BS, still is. But the authors don't tell us there is a match for
8.44 ohms if the outer secs are paralleled and then series with the
centre sec winding. Current density is unequal, and there may be more
instability, but it is doable, and better than using a 3.75 winding to
drive an 8 ohm speaker in an amp capable of only maybe 10W max.

The UL tranny specs have Np = 3,880 turns for RLa-a = 6.97kohms.
Ns can be :-
1. Two parallel windings of 90t for RL Sec = 3.75 ohms,
2. Series winding with all sec turns in series of 18t for 15 ohms,
3. Centre 90t winding in series with outer 45t secs in parallel for
135t for 8.44 ohms.

Now becayse the OPT is such a miniature type of concoction so typical
or parsimonious designers of 1956 who hated to see anyone getting
served too much iron and copper, the LL was reasonably low, and one
might stablise the amp OK.

Infinitely better is to use a bigger core and limit Np to 2,800 turns
which will reduce all inductances including LL by a factor of about
0.5, but raising AFe will boost Lp back to good enough.

Primary is 14 layers of wire at 200t per layer. There are 3 P sections
with a centre section of 6 layers with CT for B+. The outer 2 P
sections have 4 layers each.

Th sec is 4 windings of 66 turns with the last on sec divided into 3
windings of 22 turns.
This gives :-

4 // ( 66t x 4 ) = 66t for 7k0 : 3.89 ohms,
3 // ( [66t + 22t] x 3 ) + 66t ) = 88t for 7k0 : 6.91 ohms,
2 // ( 66t + 66t ) = 132 t for 7k0 : 15.6 ohms.

The G2 taps will be at 3/7 of the 1/2 primaries, ie located to the
wires coming out each side of innner secs so taps along primary layers
are not needed.

LL varies with the square of the number of interleavings, and anyone
will find my proposed design is far better than anything by L&G.

Now amp makers can source what are called long window wasteless E&I
core material where the window L x H dimension is the same as the
total I, so that the core dimensions may be :-
Overall plan area = 96mm x 72mm. Window = 72mm x 12mm. Tongue = 24mm,
and Stack may be 36mm. The long window increases traverse width of
each layer therefore reducing LL.

P is 0.28mm Cu dia wire, Sec = 0.85mm Cu dia wire and P-S insulation =
0.35mm, and p-p insulation = 0.05mm, so winding height total = 11.1mm
allowing for wire enamel on grade 2 winding wire.

Using E&I with window = 16mm x 96mm gives thicker wire and insulations
and 60W capability without difficulty. But as soon as you get right
away from Olde British Engineering, the sun shines on your efforts and
you end up with a decent amp. Ordinary wasteless E&I lams are fine for
OPTs without trying to source rare long window type E&I.

Patrick Turner.

[John Byrns]

In article ,
Patrick Turner wrote:

On Jul 13, 11:47*pm, John Byrns wrote:
In article
,
*Estee Eff wrote:

Hi John,
I believe that the article that you are thinking of is by D M Leakey &
R M Gilson. If you Google for "gilson wireless world" follow the link
to Douglas Self's site.
Regards,
Ian.

Thanks Ian,

That is exactly the article I was looking for! *Now to read it through
and try to understand it.

I recall reading such an article in abour 1994 when I was making
prototype OPTs for future amps I wanted to build. Unfortunately
Leakey&Gilson have very little to offer. Williamson had beaten them
many years before to get to good design principles. Notice the
ponderous way L&G proceed to propose reverse wound coils and so forth
but ending up with the obvious simplest method with interleaving of
1/4S - 1/2P -1/2S -1/2P -1/4S.

Where did "L&G" mention reverse wound coils, I guess I missed that part?

What do you feel was ponderous about "L&G's" presentation??

This was used by countless makers of low to medium quality affordable
amps, and as everyone srious about home built audio should know, never
copy what a major manufacturer does because half the design is ruined
by accountants, or an engineer worried about costs.

You are obsessed with transformers for people with a lot of money to spend on
their high end audio pursuits. Not everyone can afford the transformers you
propose, or if they can afford them they may not feel the need to have
transformers of the specifications you propose. The audio market place has/had
need for more than the very high end products you build, many people are/were
completely satisfied with less. This "L&G" transformer design is a perfectly
valid design, just because it doesn't suit your requirements doesn't mean that
it is somehow invalid.

In 1956, it was very common to have 15 ohm and 3.75 ohm load matchses
with secs in series or parallel. There was rarely any attempt for
better load matching possiblities. Cost too much. Unecessary, etc. Oh
yeah?
Was BS, still is. But the authors don't tell us there is a match for
8.44 ohms if the outer secs are paralleled and then series with the
centre sec winding. Current density is unequal, and there may be more
instability, but it is doable, and better than using a 3.75 winding to
drive an 8 ohm speaker in an amp capable of only maybe 10W max.

The UL tranny specs have Np = 3,880 turns for RLa-a = 6.97kohms.
Ns can be :-
1. Two parallel windings of 90t for RL Sec = 3.75 ohms,
2. Series winding with all sec turns in series of 18t for 15 ohms,
3. Centre 90t winding in series with outer 45t secs in parallel for
135t for 8.44 ohms.

This is not really an issue with the transformer, it is more of an issue of what
complexity the end user can deal with when connecting his speakers. When
presented with connecting three secondary sections as you propose, there is
going to be a fair chance that an error may be made. Bringing out only two
sections simplifies things somewhat, reducing the chance of error. A single
section would be ideal. There is probably a reason why QUAD only brought out
the connections for two secondary sections, when they could have easily brought
out more.

Now becayse the OPT is such a miniature type of concoction so typical
or parsimonious designers of 1956 who hated to see anyone getting
served too much iron and copper, the LL was reasonably low, and one
might stablise the amp OK.

That was the whole point, to meet the power and response specifications and
yield a stable output stage, while not throwing money away on over engineering
and more copper and iron than is necessary to do the required job.

Infinitely better is to use a bigger core and limit Np to 2,800 turns
which will reduce all inductances including LL by a factor of about
0.5, but raising AFe will boost Lp back to good enough.

That is only "infinitely better" if you are pursuing a cost is no object design.
Not everyone can pay the price for that, and others that can pay may not want or
need it.

How do you propose raising AFe?

Primary is 14 layers of wire at 200t per layer. There are 3 P sections
with a centre section of 6 layers with CT for B+. The outer 2 P
sections have 4 layers each.

Th sec is 4 windings of 66 turns with the last on sec divided into 3
windings of 22 turns.
This gives :-

4 // ( 66t x 4 ) = 66t for 7k0 : 3.89 ohms,
3 // ( [66t + 22t] x 3 ) + 66t ) = 88t for 7k0 : 6.91 ohms,
2 // ( 66t + 66t ) = 132 t for 7k0 : 15.6 ohms.

The G2 taps will be at 3/7 of the 1/2 primaries, ie located to the
wires coming out each side of innner secs so taps along primary layers
are not needed.

G2 taps at 3/7 is a 43% tapping, that doesn't fit the 20% specification "L&G"
called for. To meet the "L&G" 20% tapping spec. a tap would be required in the
middle of a primary layer.

LL varies with the square of the number of interleavings, and anyone
will find my proposed design is far better than anything by L&G.

Only if you need lower LL than "L&G" proposed.

Now amp makers can source what are called long window wasteless E&I
core material where the window L x H dimension is the same as the
total I, so that the core dimensions may be :-
Overall plan area = 96mm x 72mm. Window = 72mm x 12mm. Tongue = 24mm,
and Stack may be 36mm. The long window increases traverse width of
each layer therefore reducing LL.

What was the size of the lams in the "L&G" transformer design? They mention the
stack height but not the dimensions of the lams.

P is 0.28mm Cu dia wire, Sec = 0.85mm Cu dia wire and P-S insulation =
0.35mm, and p-p insulation = 0.05mm, so winding height total = 11.1mm
allowing for wire enamel on grade 2 winding wire.

Using E&I with window = 16mm x 96mm gives thicker wire and insulations
and 60W capability without difficulty. But as soon as you get right
away from Olde British Engineering, the sun shines on your efforts and
you end up with a decent amp. Ordinary wasteless E&I lams are fine for
OPTs without trying to source rare long window type E&I.

The same issue of Wireless World that has the "L&G" article also has an
interesting article titled "Tetrodes With Screen Feedback", a good read
available on the same website.

The KT55 datasheet also mentions the plate to screen capacitors mentioned in the
"L&G" article.

--
Regards,

John Byrns

Surf my web pages at, http://fmamradios.com/

[Patrick Turner]

On Jul 15, 7:55*am, John Byrns wrote:
In article ,
*Patrick Turner wrote:





On Jul 13, 11:47 pm, John Byrns wrote:
In article
,
Estee Eff wrote:

Hi John,
I believe that the article that you are thinking of is by D M Leakey &
R M Gilson. If you Google for "gilson wireless world" follow the link
to Douglas Self's site.
Regards,
Ian.

Thanks Ian,

That is exactly the article I was looking for! Now to read it through
and try to understand it.

I recall reading such an article in abour 1994 when I was making
prototype OPTs for future amps I wanted to build. Unfortunately
Leakey&Gilson have very little to offer. Williamson had beaten them
many years before to get to good design principles. Notice the
ponderous way L&G proceed to propose reverse wound coils and so forth
but ending up with the obvious simplest method with interleaving of
1/4S - 1/2P -1/2S -1/2P -1/4S.

Where did "L&G" mention reverse wound coils, I guess I missed that part?

See the centre top winding layout schematic on Page 31 of the 4 page
article.

What do you feel was ponderous about "L&G's" presentation??

Because Williamson had already got to the guts of things. L&G were
doing unecessary research into something already known about, tryna
finda way to simplicity that's suit accountants, when Blind Freddy
could have done it better. If you'd designed and wound a few trannies,
you'd know this. Its very difficult to see the obvious if you don't
practise the craft and just wanna sit around having acedemic
discussions.

This was used by countless makers of low to medium quality affordable
amps, and as everyone srious about home built audio should know, never
copy what a major manufacturer does because half the design is ruined
by accountants, or an engineer worried about costs.

You are obsessed with transformers for people with a lot of money to spend on
their high end audio pursuits. *

Indeed Iam only interested in building good amps for people who want
un-compromised engineering.

I refuse to reproduce the awful crap foisted upon unsuspecting members
of the public at supposed bargain prices.

Not everyone can afford the transformers you
propose, or if they can afford them they may not feel the need to have
transformers of the specifications you propose. *

Let those who won't save enough for good eat cake if they want. I
don't care about ppl with no money, and they sure don't care about
me.

The audio market place has/had
need for more than the very high end products you build, many people are/were
completely satisfied with less. *

Of course people are satisfied - for awhile. The Chinese used to have
1 billion bicycles and cars were a rare sight on the few Chinese roads
capable to taking a car, in 1985. But now look where they are. To be
rich is glorious, and nobody is satisfied for longer than 5 minutes.

People in the West routinely become ****ed off with good stuff they
bought and put it on the tip to buy better stuff.

I don't try to copy hi-end which tries desperately to Style the Amp to
make it look right. This turns out to be more important for the
creators than the circuit integrity, which is often just re-cycled
designs from 1955 but often with a whole pile of little add ons which
are easy because of printed circuit boards. Don't get me started on Hi-
****in-End. So often when I have to rebuild amps made by hi-end makers
I am reminded of their technical incompetence.

This "L&G" transformer design is a perfectly
valid design, just because it doesn't suit your requirements doesn't mean that
it is somehow invalid.

Let ppl buy as much valid invalidity as they wish. Just what do 2 x
N709 do with music and modern insensitive speakers? Nothing much. Much
junk designed in 1955 is no longer appropriate in 2011 because it
fails to meet modern expectations. L&G filled pages of WW rather
nicely, and WW existed to make a profit, and most ppl were poor and
miserable and high expectations were for the rich, who mostly didn't
give a **** about the techno talk in WW, and they just bought the best
at a shop as usual, if it was available, and often that was a vanity
because only a few of the rich have time to appreciate music at home
because they are too concerned with counting their pennies.

For a decent amp in 1956, at least 2 x KT66 or 6L6 were needed. l

In 1956, it was very common to have 15 ohm and 3.75 ohm load matchses
with secs in series or parallel. There was rarely any attempt for
better load matching possiblities. Cost too much. Unecessary, etc. Oh
yeah?
Was BS, still is. But the authors don't tell us there is a match for
8.44 ohms if the outer secs are paralleled and then series with the
centre sec winding. Current density is unequal, and there may be more
instability, but it is doable, and better than using a 3.75 winding to
drive an 8 ohm speaker in an amp capable of only maybe 10W max.

The UL tranny specs have Np = 3,880 turns for RLa-a = 6.97kohms.
Ns can be :-
1. Two parallel windings of 90t for RL Sec = 3.75 ohms,
2. Series winding with all sec turns in series of 18t for 15 ohms,
3. Centre 90t winding in series with outer 45t secs in parallel for
135t for 8.44 ohms.

This is not really an issue with the transformer, it is more of an issue of what
complexity the end user can deal with when connecting his speakers. *

So do you think ppl were mainly STUPID in 1956? Many were, of course;
they drank and smoked too much and had poor education, and levels of
stupidity were high. But the parallel or series thing could be avoided
with a tapped sec. What makers failed to do was to realise that if
tapped secs are to be used then you still have to use the same number
sec sections, and each none is identical and has the same taps so they
can all be paralled to give Com, 4, 8 16. This means many more Sec
turns have to be used and its a lot more work to do this right. but
makers hated the ****ing costs of anything more than the minimum lousy
****ing thing they could get away with.

I say let all those so concerned with costs go take a running jump off
a cliff somewhere.

When
presented with connecting three secondary sections as you propose, there is
going to be a fair chance that an error may be made. *

Indeed. The Home Operator has smoked vast number of cigarettes,
amplifiers, heaters, toasters, firewood, dinners, and cars, and
sometimes a wife. He routinely smoked all sorts of stuff. He's into
combustion. It fascinates him. He'll find a way to plug a 2A3 into an
amp the wrong way if the socket allows, and sit and wonder about the
silence and the hot power trannies. Then he wonders why they went
cool. A customer I has cooked both the potted PTs in a pair of Sun
amps this way. It was an expensive repair, but I installed active
protection in case he did it again. Owners drink at night with hi-fi
going. This is how TTs get wrecked. There is is always a fair chance
**** Will Happen.

Quad-II amps are regularly used with 4 ohm speakers while the OPT is
set for 16 ohms. This means the RLa-a is about 1k0 instead of 4k0. Who
understands that? Not a living soul!

Owners routinely plug 4 ohm speakers into the Com-16 outlet.

So it matters not what any maker does about load matching, the Home
Operator will find a way to **** things up if there is a way.

I'm here for the educated ppl with humility who always enquire about
the pitfalls of dealing with all things in life.
They seem not to cause too much smoke.

The chance of error can be VASTLY reduced in makers were to place
suitable warnings and signage on their gear on the rear panels and
using large black lettering on white backgrounds so ppl will read the
message. But much is missing from rear panels of amps.

Bringing out only two
sections simplifies things somewhat, reducing the chance of error. *A single
section would be ideal. *There is probably a reason why QUAD only brought out
the connections for two secondary sections, when they could have easily brought
out more.

With Quad, it requires a service tech to change load matching. The
bottom cover is removed and soldered links are changed by someone with
a soldering iron, and who knew what to do. There was NO information on
the bottom cover and ppl were expected to read the owner manual, which
ppl don't always do. So Quads could cause muddles.

Leak had a plug you pulled out and turned around to a different
position to change Z match. If you lost the little plug, then No
MUSIC. The amp was saved. Leaks awful OPTs.

But more could have been done but wasn't. After 1960 Quad used a
method for changing preamp load matching for cartriges which involved
a square printed circuit board card which could be inserted 4
different ways into a slot with a multi contact socket with decently
made connectors that would last being plugged and unplugged 50,000
times. Such a scheme is not hard to make work. But not in 1955. Too
dificult. Too expensive. But the card is unlikely to get lost. A
similar arrangement can be achieved with an octal tube socket and tou
use differently configured plugs. But ppl lose plugs, then wire them
wrong. One cannot win. When SS was finally invented then these trouble
s dissapeared because SS amps could be used with any load above 2 ohms
- there was 60dB of NFB to keep things linear. But even so, the Home
Operator managed to smoke a mountain sized pile of SS amps. The
maunfacturers ensured this would occur to keep themselves employed.
Not one manufacturer has fitted a protection circuit capable of
detecting a load value below 2.5 ohms and then turning off the amp
immediately. There are countless Home Operators who have managed to
short circuit amps with signal present, or make appalling efforts
preparing speaker cabling. I've had to often repair the results of
device failures caused by speaker cable shorts or speaker driver
damage.


Now becayse the OPT is such a miniature type of concoction so typical
or parsimonious designers of 1956 who hated to see anyone getting
served too much iron and copper, the LL was reasonably low, and one
might stablise the amp OK.

That was the whole point, to meet the power and response specifications and
yield a stable output stage, while not throwing money away on over engineering
and more copper and iron than is necessary to do the required job.

This policy translated to manufacture of mediocre garbage. The Morris
motor car in 1956 was a Mobile Horror.

It suited the masses with no money, but they had enough to buy smokes
and booze and trashy food and clothes et all. And the baby boom showed
they had little restraint in the bedroom. Inconsistency ruled. But the
rich bought Jaguars and Rolls Royces. Good stuff could be made, just
not cheaply enough. Everyone poor wanted better of course, NOBODY was
really satisfied. Don't you realise the more ppl have, the more they
want? This is as valid as saying the less ppl have the more thay want.
This is why the world will never ever run out of the one vast
commodity which exists - DEMAND.



Infinitely better is to use a bigger core and limit Np to 2,800 turns
which will reduce all inductances including LL by a factor of about
0.5, but raising AFe will boost Lp back to good enough.

That is only "infinitely better" if you are pursuing a cost is no object design. *
Not everyone can pay the price for that, and others that can pay may not want or
need it.



How do you propose raising AFe?

The same winding layout and turn count and wire size may be used for
the same E&I lam size and the height of the stack size can be used to
varied to suit the wanted PO to a large extent.

Doing things my way in 1956 would not have cost much more.

But no doubt I'd have been sacked 5 minutes after being hired in 1956.
In reponse I would have started my own company to make stuff they way
I wanted, and not how accountants wanted. I might/might not have
succeeded. I don't care what might have been.

Primary is 14 layers of wire at 200t per layer. There are 3 P sections
with a centre section of 6 layers with CT for B+. The outer 2 P
sections have 4 layers each.

Th sec is 4 windings of 66 turns with the last on sec divided into 3
windings of 22 turns.
This gives :-

4 // ( 66t x 4 ) = 66t for 7k0 : 3.89 ohms,
3 // ( [66t + 22t] x 3 ) + 66t ) = 88t for 7k0 : 6.91 ohms,
2 // ( 66t + 66t ) *= 132 t for 7k0 : 15.6 ohms.

The G2 taps will be at 3/7 of the 1/2 primaries, ie located to the
wires coming out each side of innner secs so taps along primary layers
are not needed.

G2 taps at 3/7 is a 43% tapping, that doesn't fit the 20% specification "L&G"
called for. *To meet the "L&G" 20% tapping spec. a tap would be required in the
middle of a primary layer.

Taps placed somewhere along a layer in in any OPT are poor practice
when there is an available place to bring ut a tap at the end of each
layer. Forget being married to 20%. taps can be placed at ends of all
P layers, 2 layers is 2/7 = 28.6% which is near enough for where a low
% UL tap is needed. Using taps less than this is next ro useless.
Mullard suggested 43% taps work best for UL. Leak used 50%. I've used
up to 66%, which works best for pure class A.

LL varies with the square of the number of interleavings, and anyone
will find my proposed design is far better than anything by L&G.

Only if you need lower LL than "L&G" proposed.

Of if you want the best. L&G offer the Morris solution. Nothing
brilliant.


Now amp makers can source what are called long window wasteless E&I
core material where the window L x H dimension is the same as the
total I, so that the core dimensions may be :-
Overall plan area = 96mm x 72mm. Window = 72mm x 12mm. Tongue = 24mm,
and Stack may be 36mm. The long window increases traverse width of
each layer therefore reducing LL.

What was the size of the lams in the "L&G" transformer design? *They mention the
stack height but not the dimensions of the lams.

Info is deliberately omitted. Don't let ppl know too much. But assume
the Afe of core is a square section with S = T.

P is 0.28mm Cu dia wire, Sec = 0.85mm Cu dia wire and P-S insulation =
0.35mm, and p-p insulation = 0.05mm, so winding height total = 11.1mm
allowing for wire enamel on grade 2 winding wire.

Using E&I with window = 16mm x 96mm gives thicker wire and insulations
and 60W capability without difficulty. But as soon as you get right
away from Olde British Engineering, the sun shines on your efforts and
you end up with a decent amp. Ordinary wasteless E&I lams are fine for
OPTs without trying to source rare long window type E&I.

The same issue of Wireless World that has the "L&G" article also has an
interesting article titled "Tetrodes With Screen Feedback", a good read
available on the same website.

UL connection is a method of screen FB.

The KT55 datasheet also mentions the plate to screen capacitors mentioned in the
"L&G" article.

I don't like anode to G2 caps in UL amps because it always loads a
winding with pure C that causes resonances, and is Wrong. But Zobels
from anode to screen sometimes work as well as Zobels across each 1/2
primary to damp HF instability.

Read my website and forget all those old boring grey British Wannabes.
The best info from WW and other sources has been distilled and stored
for all at my website.

Patrick Turner.

[John Byrns]

In article ,
Patrick Turner wrote:

Because Williamson had already got to the guts of things. L&G were
doing unecessary research into something already known about, tryna
finda way to simplicity that's suit accountants, when Blind Freddy
could have done it better. If you'd designed and wound a few trannies,
you'd know this. Its very difficult to see the obvious if you don't
practise the craft and just wanna sit around having acedemic
discussions.

Why then didn't Williamson provide more information and discussion on how to
design and build output transformers to meet less rigorous specifications than
he demanded for his "Williamson Amplifier"? The bottom line is that Williamson
didn't explain how to make these compromises, he just drove straight ahead
toward the specifications he wanted and left it at that.

It is relatively easy to design and build a cost is no object transformer such
as the ones you produce. It takes a little more skill to come up with a good
compromise when cost enters into the equation. I personally would be very
interested in the methods employed by "Blind Freddy" to design a better
transformer than "L&G" at the same price point. While "Blind Freddy" may have
been able to do it, it is far from obvious that you have the experience and
skill necessary to do it.


Let ppl buy as much valid invalidity as they wish. Just what do 2 x
N709 do with music and modern insensitive speakers? Nothing much.


The problem isn't the 2 x N709, the problem is with modern insensitive speakers,
another design copout resulting from the easy availability boat loads of solid
state power. Large high power tube amplifiers are not very Green, better to use
a small tube amp with more efficient speakers.

I had never heard of N709s before, they appear to be a knockoff of the EL84, a
great little audio power tube.

--
Regards,

John Byrns

Surf my web pages at, http://fmamradios.com/

[Patrick Turner]

On Jul 16, 9:28*am, John Byrns wrote:
In article ,
*Patrick Turner wrote:

Because Williamson had already got to the guts of things. L&G were
doing unecessary research into something already known about, tryna
finda way to simplicity that's suit accountants, when Blind Freddy
could have done it better. If you'd designed and wound a few trannies,
you'd know this. Its very difficult to see the obvious if you don't
practise the craft and just wanna sit around having acedemic
discussions.

Why then didn't Williamson provide more information and discussion on how to
design and build output transformers to meet less rigorous specifications than
he demanded for his "Williamson Amplifier"? *The bottom line is that Williamson
didn't explain how to make these compromises, he just drove straight ahead
toward the specifications he wanted and left it at that.

And a great man he was. He wasn't into dumbing down ideas.



It is relatively easy to design and build a cost is no object transformer such
as the ones you produce. *It takes a little more skill to come up with a good
compromise when cost enters into the equation. *I personally would be very
interested in the methods employed by "Blind Freddy" to design a better
transformer than "L&G" at the same price point. *While "Blind Freddy" may have
been able to do it, it is far from obvious that you have the experience and
skill necessary to do it.

Dynaco ST70 OPTs = Blind Freddy Quality. Just one of thousands of
examples of crap.

Let ppl buy as much valid invalidity as they wish. Just what do 2 x
N709 do with music and modern insensitive speakers? Nothing much.

The problem isn't the 2 x N709, the problem is with modern insensitive speakers,
another design copout resulting from the easy availability boat loads of solid
state power. *Large high power tube amplifiers are not very Green, better to use
a small tube amp with more efficient speakers.

And where does one find adequate sensitive speakers now? They are just
not made at any volume anywhere, or everyone would have them and none
of my customers have any, except for a few who have horn loaded
things, Lowthers, etc, and they cost a pile.

95% of ppl I work for have speakers rated for 87dB SPL per watt at 1M,
96dB/W/M has long gone, so 32W is needed where once 8W was fine. Class
AB SS amps with low bias currents are green enough. But ppl WANT MORE
AND MORE AND MORE and so they have huge flat TV screens and 5+1
surround amps and all this stuff which all adds up, despite the amps
going to PWM and being 96% efficient. CO2 emissions from western
nation households are increasing despite some people switching to
greener methds of doing things.

15" Dual Concentric Tannoys were excellent, and thrive on 9 watts, but
needed a huge box to get real bass and this cost heaps and upset the
missus.


I had never heard of N709s before, they appear to be a knockoff of the EL84, a
great little audio power tube.

I like EL84/6BQ5 as driver triodes in power amps. Yes, they are good
OP tubes for peanut power, but EL34 kill them.

Patrick Turner.