How Important Is the Sound Of the Phase Inverter?

[Armand]

I know that the sections should be well matched, but how much does the PI
affect the overall sound of an amp, gain notwithstanding? Thanks.

[Fred Nachbaur]

Armand wrote:

> I know that the sections should be well matched, but how much does the PI
> affect the overall sound of an amp, gain notwithstanding? Thanks.
>

Any stage in an amplifier is like a link in a chain. If you have a weak
link, the chain will break no matter how strong the rest of the links are.

So if your phase inverter introduces a lot of distortion, you'll have a
lot of distortion! There are a great many ways of accomplishing phase
inversion; they all have certain characteristics which *can* affect the
overall result -- just like any other amplifier stage.

Cheers,
Fred
--
+--------------------------------------------+
| Music: http://www3.telus.net/dogstarmusic/ |
| Projects: http://dogstar.dantimax.dk |
+--------------------------------------------+

[Nothing40]

(Armand) wrote in message >...
> I know that the sections should be well matched, but how much does the PI
> affect the overall sound of an amp, gain notwithstanding? Thanks.

I'm not really sure...All of the amp I have,and have made (except SE)
use the simple "cathodyne" PI. Balance between tubes,load
resistors,coupling caps,etc is quite important,as the PI may output a
non-symmetrical signal,or have one output with a lower
amplitude,etc..All of which equals distortion!

I'm not sure many PI circuits have a whole lot of gain,I know the
cathodyne is unity gain (well,not really,slightly less.)
Idealy,the PI would not alter the signal in any way,except for
"splitting" it. :)

[Henry 007]

"Armand" > wrote in message
...
> I know that the sections should be well matched, but how much does the PI
> affect the overall sound of an amp, gain notwithstanding? Thanks.
>

I like to use a dual triode for each channel, with an AC balance adjustment
on the anode resistors, which allows you to adjust the balance of drive to
the output tubes before NFB is applied. Basically a knob which allows you to
balance the output of the drivers.

[Patrick Turner]

Henry 007 wrote:

> "Armand" > wrote in message
> ...
> > I know that the sections should be well matched, but how much does the PI
> > affect the overall sound of an amp, gain notwithstanding? Thanks.
> >
>
> I like to use a dual triode for each channel, with an AC balance adjustment
> on the anode resistors, which allows you to adjust the balance of drive to
> the output tubes before NFB is applied. Basically a knob which allows you to
> balance the output of the drivers.

The PI with the lowest THD and widest BW and highest voltage swing,
and which is easy to drive, and with adequate gain, and with excellent
balanced drive is the long tail pair, for all the large octal power tubes.
A typical circuit is shown at
http://www.turneraudio.com.au/htmlwebpgs02/schema5050ulabinteg.htm

But for circuits using EL84, 6V6,
the circuit at http://www.turneraudio.com.au/htmlwebpgs02/schem1010ulabinteg.htm

should sound very nice, because the Concertina PI isn't being driven hard,
and nor is the tube ahead of it.
The CPI tube itself has a gain of only 2, counting cathode and anode output
voltages combined, compared to the grid input.
so when a 12AX7 is used for this stage, there is local FB
which reduces the open loop gain about 30 times, and the thd at
2 x 5 volts output is less than 0.05%.
BUT, the input tube on its own has to make just over 5 volts
will have much more thd, which will total perhaps 0.3%,
and all 2H.
To this we could add the the thd of the output stage, which might be 3%,
and mainly all 3H.
So at the output, expect to measure about 3.3%, and this is without FB,
for a class AB UL amp.
When FB is added, ALL this thd gets reduced perhaps by 5 times,
to 0.6%. The mix of distortion products stays about constant for the first
few watts, with or without FB, and its just the quantity that is reduced by FB.
Some extra higher harmonics are created by the FB, but usually,
these "second order" artifacts are not audible in well made tube amps.

And the sound? That's for you to find out for yourself.

I think one should be mentally flexible about drive amp circuits;
I like to stick to medium U triodes, perhaps with a 12AX/AT/AY7
at the input.

Patrick Turner.

[Armand]

Thanks. That answers my question for the most part.

But I was also thinking about the tube used. Should one spend the money for
say an RCA BP 6SL7, the inverter in this case, or settle for something less
expensive?




In article >,
says...
>
>
>"Armand" > wrote in message
...
>> I know that the sections should be well matched, but how much does the PI
>> affect the overall sound of an amp, gain notwithstanding? Thanks.
>>
>
>Hi, Armand,
>
>Short answer. It doesn't.!!!!!!
>I have often rebuilt the same PP design with different inverters. The
>distortion contributed by the inverter stage is nothing compared to that
>produced by the following stages, and you certainly won't hear it, provided
>the two antiphase outputs are closely balanced.
>So why are there different inverter configurations and which one is best ??
>Only two inverters seem to be regularly used now. The cathodyne and the
LTP.
>Which you choose is not dictated by which sounds best, it is decided by the
>AC output required, the impedance of the output load you are driving, and
>the available B+ voltage.
>kind regards
>jim
>
>

[TubeGarden]

Hi RATs!

Yes. Buy as many tubes as you can find. Use the one you like the best.

Nobody but you knows what you will like, and you won't, neither, unless you
try.

Do not assume price is the way to perfection ;)

Happy Ears!
Al


Alan J. Marcy
Phoenix, AZ

PWC/mystic/Earhead

[Patrick Turner]

Ronald wrote:

> What do you guys think of this splitter :
> http://www.angelfire.com/electronic2/audio_tubes/6L6schema.html
> I think it works great in this / my amp .....
>
> Ronald.

Your PI is what's known as the paraphase splitter.
In your case the drive for the bottom gain driver tube comes from the the
low output Z of the top follower tube.
Since the 6SL7 is a very linear triode when it sees a high Z load value
as you have there, this phase inverter should work fine, up to a high F.
The Ro from the mu-follower as you have done it might be as low as 5kohms,
so driving the outputs is no problem, especially since
they are in tetrode mode, and have low miller input C.

Probably, the balance of such a paraphase amp could vary.
so some form of trimmer R could be used temporarily when setting up,
and then find the right value for balance, then replace with a fixed value R,
and the balance shouldn't change for years.
Even 5% unbalance in signal drive won't be heard.

BTW, in your case, NFB can be applied to the cathode of the input valve
from the amp output if needed.
Trimming the R values may still be needed, but I see you have the variety
of paraphase which is self balancing, with the 3 x 180k arranged just so.
I'd prefer to have even lower bias R than 180k for the output tubes,
probably OK for 6L6, but they'd need to be lower value for 6550.

Just watch out for LF instability with FB, because there is CR
coupling between the gain tubes and followers on top,
and then the CR to output grids, then you have bypassed Rk
at the outputs, and one of the input tubes, and then there is
the inductance of the OPT across the load, and the gain of the output tubes is
high,
especially with no load, so beware LF oscillations.

Patrick Turner.


>
>
> "Patrick Turner" > schreef in bericht
> ...
> >
> >
> > Henry 007 wrote:
> >
> > > "Armand" > wrote in message
> > > ...
> > > > I know that the sections should be well matched, but how much does the
> PI
> > > > affect the overall sound of an amp, gain notwithstanding? Thanks.
> > > >
> > >
> > > I like to use a dual triode for each channel, with an AC balance
> adjustment
> > > on the anode resistors, which allows you to adjust the balance of drive
> to
> > > the output tubes before NFB is applied. Basically a knob which allows
> you to
> > > balance the output of the drivers.
> >
> > The PI with the lowest THD and widest BW and highest voltage swing,
> > and which is easy to drive, and with adequate gain, and with excellent
> > balanced drive is the long tail pair, for all the large octal power tubes.
> > A typical circuit is shown at
> > http://www.turneraudio.com.au/htmlwebpgs02/schema5050ulabinteg.htm
> >
> > But for circuits using EL84, 6V6,
> > the circuit at
> http://www.turneraudio.com.au/htmlwebpgs02/schem1010ulabinteg.htm
> >
> > should sound very nice, because the Concertina PI isn't being driven hard,
> > and nor is the tube ahead of it.
> > The CPI tube itself has a gain of only 2, counting cathode and anode
> output
> > voltages combined, compared to the grid input.
> > so when a 12AX7 is used for this stage, there is local FB
> > which reduces the open loop gain about 30 times, and the thd at
> > 2 x 5 volts output is less than 0.05%.
> > BUT, the input tube on its own has to make just over 5 volts
> > will have much more thd, which will total perhaps 0.3%,
> > and all 2H.
> > To this we could add the the thd of the output stage, which might be 3%,
> > and mainly all 3H.
> > So at the output, expect to measure about 3.3%, and this is without FB,
> > for a class AB UL amp.
> > When FB is added, ALL this thd gets reduced perhaps by 5 times,
> > to 0.6%. The mix of distortion products stays about constant for the first
> > few watts, with or without FB, and its just the quantity that is reduced
> by FB.
> > Some extra higher harmonics are created by the FB, but usually,
> > these "second order" artifacts are not audible in well made tube amps.
> >
> > And the sound? That's for you to find out for yourself.
> >
> > I think one should be mentally flexible about drive amp circuits;
> > I like to stick to medium U triodes, perhaps with a 12AX/AT/AY7
> > at the input.
> >
> > Patrick Turner.
> >
> >

[Patrick Turner]

jim wrote:

> "Patrick Turner" > wrote in message
> ...
> >
> >
> > Henry 007 wrote:
> >
> > > "Armand" > wrote in message
> > > ...
> > > > I know that the sections should be well matched, but how much does the
> PI
> > > > affect the overall sound of an amp, gain notwithstanding? Thanks.
> > > >
> > >
> > > I like to use a dual triode for each channel, with an AC balance
> adjustment
> > > on the anode resistors, which allows you to adjust the balance of drive
> to
> > > the output tubes before NFB is applied. Basically a knob which allows
> you to
> > > balance the output of the drivers.
> >
> > The PI with the lowest THD and widest BW and highest voltage swing,
> > and which is easy to drive, and with adequate gain, and with excellent
> > balanced drive is the long tail pair, for all the large octal power tubes.
> > A typical circuit is shown at
> > http://www.turneraudio.com.au/htmlwebpgs02/schema5050ulabinteg.htm
> >
> > But for circuits using EL84, 6V6,
> > the circuit at
> http://www.turneraudio.com.au/htmlwebpgs02/schem1010ulabinteg.htm
> >
>
> Well ??????
> The man asked whether you could 'HEAR' the difference between different
> phase inverters. No, done properly, you can't . .... Phase inversion is a
> pain in the arse. Bust it in two halves and stitch it back together with
> two valves and a split arse transformer is just asking for trouble... which
> sort of invertor sounds better ????
> You tell me !!! The LTP ?? Well, yeah ...... it does the job, but to get
> it to balance you either need a CCS, or a fiddle with the anode loads, and
> what about a difference in the Gm of the two halves of the double triode ??

Well old mate, you could have a 12AU7 on one side of the LTP, and a
6DJ8 on the other, and as long as there is a CCS and equal RLs to each side,
voila, equal output voltages.
All that would change would be the slight amount of uncancelled 2H.

>
> Not today or tomorrow but next month as things age. Or if I decide to put
> another make of 12AX7 in the hole ?? Concertina ?? Easy, cheap and
> sensible but no gain. ???
> My main point, however, is that you are blatantly advertizing Turnip Audio
> again !!!

My circuits have been used, in essence, by Mullard, Leak, VTL, to name a few.
All I did was drag their circuits kicking and screaming into the modern tube age

by means of using a transistor as a pure slave to the whims of the cathodes
and their currents.

At least the free box of turnips that go with every amp I sell
is a step in the right direction for customer goodwill.

>
> I have reported this to Andre, Arny and Stewart
> They will be contacting you shortly
> regards
> jim

Last time I had to complain about Andre whacking me around
with a damp lettuce leaf until I turned green, and it wasn't even
St Paddy's day.

Patrick Turner.

[Patrick Turner]

jim wrote:

> "Armand" > wrote in message
> ...
> > Thanks. That answers my question for the most part.
> >
> > But I was also thinking about the tube used. Should one spend the money
> for
> > say an RCA BP 6SL7, the inverter in this case, or settle for something
> less
> > expensive?
> >
> >
>
> No, no, no........
>
> If you want to spend some money, spend it on the OP transformer. The larger
> the better. A concertina inverter works with 100% NFB so whether the valve
> was made in Libya or Birmingham, the result will be the same.
> 6SL7 ?? Looks pretty but is inferior to the ****ty little 9 pin 12A??
> series things in terms of gain and interelectrode capacitances. In LTP
> configuration a 6SL7 or a 12AX7 are fine, but neither are good as a
> concertina due to their high anode resistance....... in concertina mode,
> you go for a 6SN7 or a 12AU/AT7

But horses for courses Jim.
In my ealier post one of the amps I linked to was a 10 watt thinge with 6GW8.
The triode half of one tube is the CPI, and triode in other tube is
the input tube. The circuit with 17dB of NFB is sensitive to about a volt
of input.
Thousands of such circuits were used, and both bean counters and listeners
were winners.
The drive voltages are about 7 vrms max to each outptu pentode.

The output impedance of the CPI is LOW, providing the load
the anode and cathode see is the same, and it is, in all circuits,
until at HF, some phase shift occurs, and the anode output voltage rolls off
more than the cathode voltage.
Measured separately, anode and cathode output R is very different,
but so what, the CPI stays well balanced to well over 50kHz,
even with 12AX7/6SL7.
The Williamson used 6SN7, the best triode of all time,
and the CPI acts as a buffer between input tube and the balanced driver stage,
which results in a terrific bandwidth for the whole four triode driver stage.

>
> Not that you will 'HEAR' whatever you put in as an inverter.
> What make ???................. My $3 Chinese 6SN7s sound great, NOS Brimars
> sound pretty much the same.... you choose.....
> regards
> jim

I don't hear much difference. A PI is just another stage you must have for
a PP amp, unless the amp has a balanced input,
then no PI is required.

Patrick Turner.

[Patrick Turner]

jim wrote:
snip,

> > Well old mate, you could have a 12AU7 on one side of the LTP, and a
> > 6DJ8 on the other, and as long as there is a CCS and equal RLs to each
> side,
> > voila, equal output voltages.
> > All that would change would be the slight amount of uncancelled 2H.
> >
>
> I don't think I'm quite ready to grasp that I can have a 6SL7 in one hole
> and an ECC83 in the other and have perfect balance from 1 to 20 kHz , but
> the answer to the man's question is --- no, you will not hear the
> inverter..... I have never heard an inverter, .. inverting ..

I must admit, I have never heard anyone claim that thay could hear
their music partially or wholly inverted,
or that they thought their music was upside down sound.

But then reversing the speaker leads for phase turns everything upside down,
but there is no need to stand on one's head, or face away from the speakers..

>
> A conventional R tailed LTP is a cathode follower driving the second half of
> the triode hence the 10% difference in the two outputs. Theory says that if
> the tail load is infinite, the anode outputs must be equal.... irrespective
> of the active device. I cannot argue with this...
> It all just seems to use so many bits ..... neg supplies, CCS pentode or MJE
> ?? etc, etc..

Well, I reckon the LTP, used by so many ppl has less parts than a williamson.
But the 10-10 litle amp has even fewer.

Now its possible to build a sublime 50 watt amp using 6 or 8 x EL84,
and have only two input triodes, and a paralleled 12AX7 for the input,
and a paralleled 12AT7 for the CPI would be a fine way to build the driver.

>
> I,m going to LTP/CCS an ECC83 and a 6L6GC with G2 connected to the anode,
> as an LTP splitter, and see what happens...........
>
> > > Not today or tomorrow but next month as things age. Or if I decide to
> put
> > > another make of 12AX7 in the hole ?? Concertina ?? Easy, cheap and
> > > sensible but no gain. ???
> > > My main point, however, is that you are blatantly advertizing Turnip
> Audio
> > > again !!!
> >
> > My circuits have been used, in essence, by Mullard, Leak, VTL, to name a
> few.
> > All I did was drag their circuits kicking and screaming into the modern
> tube age
>
> So, nothing is new. It's all been done. We have tighter spec passives and
> silicon things which we can hang on our valves........

It matters not that newness be daily invented,
but that out time be spent on making other people happy from
our labours, and great it is that much tubecraft has survived the passing
of the ages.

Patrick Turner.

>
>
> > by means of using a transistor as a pure slave to the whims of the
> cathodes
> > and their currents.
> >
> > At least the free box of turnips that go with every amp I sell
> > is a step in the right direction for customer goodwill.
> >
> > >
> > > I have reported this to Andre, Arny and Stewart
> > > They will be contacting you shortly
> > > regards
> > > jim
> >
> > Last time I had to complain about Andre whacking me around
> > with a damp lettuce leaf until I turned green, and it wasn't even
> > St Paddy's day.
> >
> > Patrick Turner.
> >
> >

[Ronald]

Hi Patrick (and all others) ,

Thanks for your answer .
I knew already it is a paraphase splitter , but a schematic tells more than
1000 words .....
I also did some tweaking on the circuit already .
I had a nasty 60kHz oscillation .
I skipped the bypass-C on the power tubes cathode and the R-network
for the splitter is somewhat different now .
It works great now .....
I only posted the link because the paraphase looked to be forgotten ......
And I like the principle of the paraphase very much .
Also there's no need for matcing tubes . It even works with a 6SN7 on top
and a 6SL7 for the bottom tube .
The only thing is you won't get the same Z-out for both halfs .....
The mu-follower variant will be able to drive the output tubes into AB2 I
think .
For the same reason it will drive (almost) any power tube .....

Ronald .

"Patrick Turner" > schreef in bericht
...
>
>
> Ronald wrote:
>
> > What do you guys think of this splitter :
> > http://www.angelfire.com/electronic2/audio_tubes/6L6schema.html
> > I think it works great in this / my amp .....
> >
> > Ronald.
>
> Your PI is what's known as the paraphase splitter.
> In your case the drive for the bottom gain driver tube comes from the the
> low output Z of the top follower tube.
> Since the 6SL7 is a very linear triode when it sees a high Z load value
> as you have there, this phase inverter should work fine, up to a high F.
> The Ro from the mu-follower as you have done it might be as low as 5kohms,
> so driving the outputs is no problem, especially since
> they are in tetrode mode, and have low miller input C.
>
> Probably, the balance of such a paraphase amp could vary.
> so some form of trimmer R could be used temporarily when setting up,
> and then find the right value for balance, then replace with a fixed value
R,
> and the balance shouldn't change for years.
> Even 5% unbalance in signal drive won't be heard.
>
> BTW, in your case, NFB can be applied to the cathode of the input valve
> from the amp output if needed.
> Trimming the R values may still be needed, but I see you have the variety
> of paraphase which is self balancing, with the 3 x 180k arranged just so.
> I'd prefer to have even lower bias R than 180k for the output tubes,
> probably OK for 6L6, but they'd need to be lower value for 6550.
>
> Just watch out for LF instability with FB, because there is CR
> coupling between the gain tubes and followers on top,
> and then the CR to output grids, then you have bypassed Rk
> at the outputs, and one of the input tubes, and then there is
> the inductance of the OPT across the load, and the gain of the output
tubes is
> high,
> especially with no load, so beware LF oscillations.
>
> Patrick Turner.
>
>
> >
> >
> > "Patrick Turner" > schreef in bericht
> > ...
> > >
> > >
> > > Henry 007 wrote:
> > >
> > > > "Armand" > wrote in message
> > > > ...
> > > > > I know that the sections should be well matched, but how much does
the
> > PI
> > > > > affect the overall sound of an amp, gain notwithstanding? Thanks.
> > > > >
> > > >
> > > > I like to use a dual triode for each channel, with an AC balance
> > adjustment
> > > > on the anode resistors, which allows you to adjust the balance of
drive
> > to
> > > > the output tubes before NFB is applied. Basically a knob which
allows
> > you to
> > > > balance the output of the drivers.
> > >
> > > The PI with the lowest THD and widest BW and highest voltage swing,
> > > and which is easy to drive, and with adequate gain, and with excellent
> > > balanced drive is the long tail pair, for all the large octal power
tubes.
> > > A typical circuit is shown at
> > > http://www.turneraudio.com.au/htmlwebpgs02/schema5050ulabinteg.htm
> > >
> > > But for circuits using EL84, 6V6,
> > > the circuit at
> > http://www.turneraudio.com.au/htmlwebpgs02/schem1010ulabinteg.htm
> > >
> > > should sound very nice, because the Concertina PI isn't being driven
hard,
> > > and nor is the tube ahead of it.
> > > The CPI tube itself has a gain of only 2, counting cathode and anode
> > output
> > > voltages combined, compared to the grid input.
> > > so when a 12AX7 is used for this stage, there is local FB
> > > which reduces the open loop gain about 30 times, and the thd at
> > > 2 x 5 volts output is less than 0.05%.
> > > BUT, the input tube on its own has to make just over 5 volts
> > > will have much more thd, which will total perhaps 0.3%,
> > > and all 2H.
> > > To this we could add the the thd of the output stage, which might be
3%,
> > > and mainly all 3H.
> > > So at the output, expect to measure about 3.3%, and this is without
FB,
> > > for a class AB UL amp.
> > > When FB is added, ALL this thd gets reduced perhaps by 5 times,
> > > to 0.6%. The mix of distortion products stays about constant for the
first
> > > few watts, with or without FB, and its just the quantity that is
reduced
> > by FB.
> > > Some extra higher harmonics are created by the FB, but usually,
> > > these "second order" artifacts are not audible in well made tube amps.
> > >
> > > And the sound? That's for you to find out for yourself.
> > >
> > > I think one should be mentally flexible about drive amp circuits;
> > > I like to stick to medium U triodes, perhaps with a 12AX/AT/AY7
> > > at the input.
> > >
> > > Patrick Turner.
> > >
> > >
>

[jim]

"Armand" > wrote in message
...
> I know that the sections should be well matched, but how much does the PI
> affect the overall sound of an amp, gain notwithstanding? Thanks.
>

Hi, Armand,

Short answer. It doesn't.!!!!!!
I have often rebuilt the same PP design with different inverters. The
distortion contributed by the inverter stage is nothing compared to that
produced by the following stages, and you certainly won't hear it, provided
the two antiphase outputs are closely balanced.
So why are there different inverter configurations and which one is best ??
Only two inverters seem to be regularly used now. The cathodyne and the LTP.
Which you choose is not dictated by which sounds best, it is decided by the
AC output required, the impedance of the output load you are driving, and
the available B+ voltage.
kind regards
jim

[Ian Iveson]

Also, most importantly in my case, it depends on the linearity of
the driven impedance, and hence on how the output stage is biased.
Not just on how big it is. One of Van der Veen's designs, for
example, uses a cathodyne ECC81 for 6*EL34.

But the slightest bit of grid current causes the cathodyne anode
output to collapse resulting in a flurry of odd harmonics at the
output of the amp.

I think there is an issue of slew rate limiting too, and/or
imbalance at HF. Current-derived cathode feedback is good for
turning a problem into a crisis.

So the cathodyne should be restricted to AB1 output stages, and it
should idle with plenty of current. An ECC83 is unlikely to be a
good candidate. An '82 or an '81 is better, as might be an '88.
You have to get enough voltage swing and enough current capability
for the task.

A LTP makes a good differential amplifier and you can call it a
splitter or an inverter if you want, but it's a very inelegant
approach, IMO.

What we need to do is to create an inverted copy of the original
signal. This is more than inverting, or splitting (whatever that
means). The copying part naturally raises the issue of distortion.

The paraphase might seem awkward for those looking for symmetry.
But you why should copying and inverting be a symmetrical process?
The paraphase uses a valve to produce an inverted copy. Perfect.
Tricky to get the details right so it stays in balance of course,
but so is the LTP.

All in all, the cathodyne used as a driver ensures balance and low
distortion only within a comparatively narrow corridor of operation,
outside of which it goes badly awry. A paraphase is always slightly
awry but maintains this over a much wider proportion of available HT
and load conditions. A LTP falls somewhere between the two.

Really you have to consider the amp as a whole. For example, how
much HT you have available depends partly on how much hum you can
tolerate at the driver stage, which also depends on how good the hum
rejection is of the driver.

cheers, Ian


"jim" > wrote in message
...
>
> "Armand" > wrote in message
> ...
> > I know that the sections should be well matched, but how much
does the PI
> > affect the overall sound of an amp, gain notwithstanding?
Thanks.
> >
>
> Hi, Armand,
>
> Short answer. It doesn't.!!!!!!
> I have often rebuilt the same PP design with different inverters.
The
> distortion contributed by the inverter stage is nothing compared
to that
> produced by the following stages, and you certainly won't hear it,
provided
> the two antiphase outputs are closely balanced.
> So why are there different inverter configurations and which one
is best ??
> Only two inverters seem to be regularly used now. The cathodyne
and the LTP.
> Which you choose is not dictated by which sounds best, it is
decided by the
> AC output required, the impedance of the output load you are
driving, and
> the available B+ voltage.
> kind regards
> jim
>
>

[Ian Iveson]

But it's big in Europe, I understand, or at least in Germany. Dunno
what the French use. Actually we don't hear much from them at all.
Maybe they don't have amplifiers in France. That'll be because of
the cafes I suppose. Must be worse in the antipodes, where music
just means banging sticks and blowing raspberries.

Especially if you're in class A and especially if you're trying to
avoid odd harmonics.

The quad 2 uses a floating paraphase for everything and there aren't
too many bits. It is quite complex in operation though and is
intended for fairly unusual output demands.

If other splitters have a sound it is a bad sound. The paraphase
can sound OK I am told.

cheers, Ian

"jim" > wrote
>
> Paraphase ?? If you like it, that's fine by me. Don,t change it.
What it
> sounds like is more important than how it measures to me. If I
was bothered
> about 0.0001% THD and 1% phase shift at 2 Mhz, I would have joined
up to
> uk.rec.audio
> Seems to use a lot of bits though.
> regards
> jim
>
>

[jim]

"Patrick Turner" > wrote in message
...
>
>
> Henry 007 wrote:
>
> > "Armand" > wrote in message
> > ...
> > > I know that the sections should be well matched, but how much does the
PI
> > > affect the overall sound of an amp, gain notwithstanding? Thanks.
> > >
> >
> > I like to use a dual triode for each channel, with an AC balance
adjustment
> > on the anode resistors, which allows you to adjust the balance of drive
to
> > the output tubes before NFB is applied. Basically a knob which allows
you to
> > balance the output of the drivers.
>
> The PI with the lowest THD and widest BW and highest voltage swing,
> and which is easy to drive, and with adequate gain, and with excellent
> balanced drive is the long tail pair, for all the large octal power tubes.
> A typical circuit is shown at
> http://www.turneraudio.com.au/htmlwebpgs02/schema5050ulabinteg.htm
>
> But for circuits using EL84, 6V6,
> the circuit at
http://www.turneraudio.com.au/htmlwebpgs02/schem1010ulabinteg.htm
>


Well ??????
The man asked whether you could 'HEAR' the difference between different
phase inverters. No, done properly, you can't . .... Phase inversion is a
pain in the arse. Bust it in two halves and stitch it back together with
two valves and a split arse transformer is just asking for trouble... which
sort of invertor sounds better ????
You tell me !!! The LTP ?? Well, yeah ...... it does the job, but to get
it to balance you either need a CCS, or a fiddle with the anode loads, and
what about a difference in the Gm of the two halves of the double triode ??
Not today or tomorrow but next month as things age. Or if I decide to put
another make of 12AX7 in the hole ?? Concertina ?? Easy, cheap and
sensible but no gain. ???
My main point, however, is that you are blatantly advertizing Turnip Audio
again !!!
I have reported this to Andre, Arny and Stewart
They will be contacting you shortly
regards
jim

[jim]

"Armand" > wrote in message
...
> Thanks. That answers my question for the most part.
>
> But I was also thinking about the tube used. Should one spend the money
for
> say an RCA BP 6SL7, the inverter in this case, or settle for something
less
> expensive?
>
>


No, no, no........

If you want to spend some money, spend it on the OP transformer. The larger
the better. A concertina inverter works with 100% NFB so whether the valve
was made in Libya or Birmingham, the result will be the same.
6SL7 ?? Looks pretty but is inferior to the ****ty little 9 pin 12A??
series things in terms of gain and interelectrode capacitances. In LTP
configuration a 6SL7 or a 12AX7 are fine, but neither are good as a
concertina due to their high anode resistance....... in concertina mode,
you go for a 6SN7 or a 12AU/AT7
Not that you will 'HEAR' whatever you put in as an inverter.
What make ???................. My $3 Chinese 6SN7s sound great, NOS Brimars
sound pretty much the same.... you choose.....
regards
jim

[Choky]

concertina,LTP,mumbo jumbo........
use xformer for PI
that's what real tubies does
hehe

--
Choky
Prodanovic Aleksandar
YU


"jim" > wrote in message
...
>
> > >
> > > No, no, no........
> > >
> > > If you want to spend some money, spend it on the OP transformer. The
> larger
> > > the better. A concertina inverter works with 100% NFB so whether the
> valve
> > > was made in Libya or Birmingham, the result will be the same.
> > > 6SL7 ?? Looks pretty but is inferior to the ****ty little 9 pin 12A??
> > > series things in terms of gain and interelectrode capacitances. In LTP
> > > configuration a 6SL7 or a 12AX7 are fine, but neither are good as a
> > > concertina due to their high anode resistance....... in concertina
> mode,
> > > you go for a 6SN7 or a 12AU/AT7
> >
> > But horses for courses Jim.
> > In my ealier post one of the amps I linked to was a 10 watt thinge with
> 6GW8.
>
> 6GW8 ?? .. Is this one of those little TP things ??
>
> > The triode half of one tube is the CPI, and triode in other tube is
> > the input tube. The circuit with 17dB of NFB is sensitive to about a
volt
> > of input.
> > Thousands of such circuits were used, and both bean counters and
listeners
> > were winners.
> > The drive voltages are about 7 vrms max to each outptu pentode.
>
>
> Now, you're taking the ****, here...... I had a Dansette record player
with
> an ECL82 in it in 1972.. I thought we were trying to move on.... IV input
??
> Oh, wow !!!
> For what out ?? 10W ?? In terms of efficiency, my electric fire does a
> better job .......
>
> > The output impedance of the CPI is LOW,
>
> No it isn't. It's variable. The OP impedances of the two outputs go in
> opposite directions as the frequency changes. They are however low/equal
> enough to be only of interest to my dog !!!!!!
>
> providing the load
> > the anode and cathode see is the same, and it is, in all circuits,
> > until at HF, some phase shift occurs, and the anode output voltage rolls
> off
> > more than the cathode voltage.
> > Measured separately, anode and cathode output R is very different,
> > but so what, the CPI stays well balanced to well over 50kHz,
> > even with 12AX7/6SL7.
> > The Williamson used 6SN7, the best triode of all time,
>
> At least we agree on something......
>
> > and the CPI acts as a buffer between input tube and the balanced driver
> stage,
> > which results in a terrific bandwidth for the whole four triode driver
> stage.
> >
>
> So why are we not all building Williamson lookalikes ??
>
>
> > >
> > > Not that you will 'HEAR' whatever you put in as an inverter.
> > > What make ???................. My $3 Chinese 6SN7s sound great, NOS
> Brimars
> > > sound pretty much the same.... you choose.....
> > > regards
> > > jim
> >
> > I don't hear much difference. A PI is just another stage you must have
for
> > a PP amp, unless the amp has a balanced input,
> > then no PI is required.
>
> Well. yes.. You want to even try thinking about the earth/earth loop
> problems with seperate turntable, CD player, tape, radio, etc......
>
> > Patrick Turner.
> >
>
>

[jim]

> > > > ...
> > > > > I know that the sections should be well matched, but how much does
the
> > PI
> > > > > affect the overall sound of an amp, gain notwithstanding? Thanks.
> > > > >
> > > >
> > > > I like to use a dual triode for each channel, with an AC balance
> > adjustment
> > > > on the anode resistors, which allows you to adjust the balance of
drive
> > to
> > > > the output tubes before NFB is applied. Basically a knob which
allows
> > you to
> > > > balance the output of the drivers.
> > >
> > > The PI with the lowest THD and widest BW and highest voltage swing,
> > > and which is easy to drive, and with adequate gain, and with excellent
> > > balanced drive is the long tail pair, for all the large octal power
tubes.
> > > A typical circuit is shown at
> > > http://www.turneraudio.com.au/htmlwebpgs02/schema5050ulabinteg.htm
> > >
> > > But for circuits using EL84, 6V6,
> > > the circuit at
> > http://www.turneraudio.com.au/htmlwebpgs02/schem1010ulabinteg.htm
> > >
> >
> > Well ??????
> > The man asked whether you could 'HEAR' the difference between
different
> > phase inverters. No, done properly, you can't . .... Phase inversion is
a
> > pain in the arse. Bust it in two halves and stitch it back together
with
> > two valves and a split arse transformer is just asking for trouble...
which
> > sort of invertor sounds better ????
> > You tell me !!! The LTP ?? Well, yeah ...... it does the job, but to
get
> > it to balance you either need a CCS, or a fiddle with the anode loads,
and
> > what about a difference in the Gm of the two halves of the double triode
??
>
> Well old mate, you could have a 12AU7 on one side of the LTP, and a
> 6DJ8 on the other, and as long as there is a CCS and equal RLs to each
side,
> voila, equal output voltages.
> All that would change would be the slight amount of uncancelled 2H.
>

I don't think I'm quite ready to grasp that I can have a 6SL7 in one hole
and an ECC83 in the other and have perfect balance from 1 to 20 kHz , but
the answer to the man's question is --- no, you will not hear the
inverter..... I have never heard an inverter, .. inverting ..
A conventional R tailed LTP is a cathode follower driving the second half of
the triode hence the 10% difference in the two outputs. Theory says that if
the tail load is infinite, the anode outputs must be equal.... irrespective
of the active device. I cannot argue with this...
It all just seems to use so many bits ..... neg supplies, CCS pentode or MJE
?? etc, etc..
I,m going to LTP/CCS an ECC83 and a 6L6GC with G2 connected to the anode,
as an LTP splitter, and see what happens...........


> > Not today or tomorrow but next month as things age. Or if I decide to
put
> > another make of 12AX7 in the hole ?? Concertina ?? Easy, cheap and
> > sensible but no gain. ???
> > My main point, however, is that you are blatantly advertizing Turnip
Audio
> > again !!!
>
> My circuits have been used, in essence, by Mullard, Leak, VTL, to name a
few.
> All I did was drag their circuits kicking and screaming into the modern
tube age

So, nothing is new. It's all been done. We have tighter spec passives and
silicon things which we can hang on our valves........

> by means of using a transistor as a pure slave to the whims of the
cathodes
> and their currents.
>
> At least the free box of turnips that go with every amp I sell
> is a step in the right direction for customer goodwill.
>
> >
> > I have reported this to Andre, Arny and Stewart
> > They will be contacting you shortly
> > regards
> > jim
>
> Last time I had to complain about Andre whacking me around
> with a damp lettuce leaf until I turned green, and it wasn't even
> St Paddy's day.
>
> Patrick Turner.
>
>

[Patrick Turner]

Ronald wrote:

> Hi Patrick (and all others) ,
>
> Thanks for your answer .
> I knew already it is a paraphase splitter , but a schematic tells more than
> 1000 words .....
> I also did some tweaking on the circuit already .
> I had a nasty 60kHz oscillation .
> I skipped the bypass-C on the power tubes cathode and the R-network
> for the splitter is somewhat different now .
> It works great now .....
> I only posted the link because the paraphase looked to be forgotten ......
> And I like the principle of the paraphase very much .
> Also there's no need for matcing tubes . It even works with a 6SN7 on top
> and a 6SL7 for the bottom tube .
> The only thing is you won't get the same Z-out for both halfs .....
> The mu-follower variant will be able to drive the output tubes into AB2 I
> think .
> For the same reason it will drive (almost) any power tube .....

I mentioned paraphase for the benefit of others in the group
who may not be familiar with such things.

The mu follower won't be a very good driver for class AB2,
because the grid current will charge up the coupling caps,
and you have cathode bias, which will shift as the Ck charge up,
during sustained levels where grid flows.

For AB2, you need fixed bias, and full cathode follower drive, with the cathodes

direct coupled to the output tube grids,
and cathode R taken to about -150v,
and with the biasing done by applying the adjustable fixed bias to the
cathode follower grid, which is then cap coupled to the anode circuit of a gain
tube.
There are other methods, using transformers and chokes,
but a really low drive impedance is needed for grid current operation,
since the input impedance to the grid when trying to
force the grid to go to +10v is very low.

Patrick Turner.

>
>
> Ronald .
>
> "Patrick Turner" > schreef in bericht
> ...
> >
> >
> > Ronald wrote:
> >
> > > What do you guys think of this splitter :
> > > http://www.angelfire.com/electronic2/audio_tubes/6L6schema.html
> > > I think it works great in this / my amp .....
> > >
> > > Ronald.
> >
> > Your PI is what's known as the paraphase splitter.
> > In your case the drive for the bottom gain driver tube comes from the the
> > low output Z of the top follower tube.
> > Since the 6SL7 is a very linear triode when it sees a high Z load value
> > as you have there, this phase inverter should work fine, up to a high F.
> > The Ro from the mu-follower as you have done it might be as low as 5kohms,
> > so driving the outputs is no problem, especially since
> > they are in tetrode mode, and have low miller input C.
> >
> > Probably, the balance of such a paraphase amp could vary.
> > so some form of trimmer R could be used temporarily when setting up,
> > and then find the right value for balance, then replace with a fixed value
> R,
> > and the balance shouldn't change for years.
> > Even 5% unbalance in signal drive won't be heard.
> >
> > BTW, in your case, NFB can be applied to the cathode of the input valve
> > from the amp output if needed.
> > Trimming the R values may still be needed, but I see you have the variety
> > of paraphase which is self balancing, with the 3 x 180k arranged just so.
> > I'd prefer to have even lower bias R than 180k for the output tubes,
> > probably OK for 6L6, but they'd need to be lower value for 6550.
> >
> > Just watch out for LF instability with FB, because there is CR
> > coupling between the gain tubes and followers on top,
> > and then the CR to output grids, then you have bypassed Rk
> > at the outputs, and one of the input tubes, and then there is
> > the inductance of the OPT across the load, and the gain of the output
> tubes is
> > high,
> > especially with no load, so beware LF oscillations.
> >
> > Patrick Turner.
> >
> >
> > >
> > >
> > > "Patrick Turner" > schreef in bericht
> > > ...
> > > >
> > > >
> > > > Henry 007 wrote:
> > > >
> > > > > "Armand" > wrote in message
> > > > > ...
> > > > > > I know that the sections should be well matched, but how much does
> the
> > > PI
> > > > > > affect the overall sound of an amp, gain notwithstanding? Thanks.
> > > > > >
> > > > >
> > > > > I like to use a dual triode for each channel, with an AC balance
> > > adjustment
> > > > > on the anode resistors, which allows you to adjust the balance of
> drive
> > > to
> > > > > the output tubes before NFB is applied. Basically a knob which
> allows
> > > you to
> > > > > balance the output of the drivers.
> > > >
> > > > The PI with the lowest THD and widest BW and highest voltage swing,
> > > > and which is easy to drive, and with adequate gain, and with excellent
> > > > balanced drive is the long tail pair, for all the large octal power
> tubes.
> > > > A typical circuit is shown at
> > > > http://www.turneraudio.com.au/htmlwebpgs02/schema5050ulabinteg.htm
> > > >
> > > > But for circuits using EL84, 6V6,
> > > > the circuit at
> > > http://www.turneraudio.com.au/htmlwebpgs02/schem1010ulabinteg.htm
> > > >
> > > > should sound very nice, because the Concertina PI isn't being driven
> hard,
> > > > and nor is the tube ahead of it.
> > > > The CPI tube itself has a gain of only 2, counting cathode and anode
> > > output
> > > > voltages combined, compared to the grid input.
> > > > so when a 12AX7 is used for this stage, there is local FB
> > > > which reduces the open loop gain about 30 times, and the thd at
> > > > 2 x 5 volts output is less than 0.05%.
> > > > BUT, the input tube on its own has to make just over 5 volts
> > > > will have much more thd, which will total perhaps 0.3%,
> > > > and all 2H.
> > > > To this we could add the the thd of the output stage, which might be
> 3%,
> > > > and mainly all 3H.
> > > > So at the output, expect to measure about 3.3%, and this is without
> FB,
> > > > for a class AB UL amp.
> > > > When FB is added, ALL this thd gets reduced perhaps by 5 times,
> > > > to 0.6%. The mix of distortion products stays about constant for the
> first
> > > > few watts, with or without FB, and its just the quantity that is
> reduced
> > > by FB.
> > > > Some extra higher harmonics are created by the FB, but usually,
> > > > these "second order" artifacts are not audible in well made tube amps.
> > > >
> > > > And the sound? That's for you to find out for yourself.
> > > >
> > > > I think one should be mentally flexible about drive amp circuits;
> > > > I like to stick to medium U triodes, perhaps with a 12AX/AT/AY7
> > > > at the input.
> > > >
> > > > Patrick Turner.
> > > >
> > > >
> >

[Patrick Turner]

jim wrote:

> > >
> > > No, no, no........
> > >
> > > If you want to spend some money, spend it on the OP transformer. The
> larger
> > > the better. A concertina inverter works with 100% NFB so whether the
> valve
> > > was made in Libya or Birmingham, the result will be the same.
> > > 6SL7 ?? Looks pretty but is inferior to the ****ty little 9 pin 12A??
> > > series things in terms of gain and interelectrode capacitances. In LTP
> > > configuration a 6SL7 or a 12AX7 are fine, but neither are good as a
> > > concertina due to their high anode resistance....... in concertina
> mode,
> > > you go for a 6SN7 or a 12AU/AT7
> >
> > But horses for courses Jim.
> > In my ealier post one of the amps I linked to was a 10 watt thinge with
> 6GW8.
>
> 6GW8 ?? .. Is this one of those little TP things ??

Yes, but two can make the same as 2 x EL84, 6V6.

>
>
> > The triode half of one tube is the CPI, and triode in other tube is
> > the input tube. The circuit with 17dB of NFB is sensitive to about a volt
> > of input.
> > Thousands of such circuits were used, and both bean counters and listeners
> > were winners.
> > The drive voltages are about 7 vrms max to each outptu pentode.
>
> Now, you're taking the ****, here...... I had a Dansette record player with
> an ECL82 in it in 1972.. I thought we were trying to move on.... IV input ??
> Oh, wow !!!
> For what out ?? 10W ?? In terms of efficiency, my electric fire does a
> better job .......

But all the old world amps had a single 12AX7 for a line stage, which included
a passive tone control and volume control.
So, with 0.2v standard value line input level was amped up 50 times
to 10vrms, and the losses in the tone control and volume
was 20dB, so 1 volt came out, and ppl thought it was luverly.

>
> > The output impedance of the CPI is LOW,
>
> No it isn't. It's variable. The OP impedances of the two outputs go in
> opposite directions as the frequency changes. They are however low/equal
> enough to be only of interest to my dog !!!!!!

.....and to anyone using NFB!

>
>
> providing the load
> > the anode and cathode see is the same, and it is, in all circuits,
> > until at HF, some phase shift occurs, and the anode output voltage rolls
> off
> > more than the cathode voltage.
> > Measured separately, anode and cathode output R is very different,
> > but so what, the CPI stays well balanced to well over 50kHz,
> > even with 12AX7/6SL7.
> > The Williamson used 6SN7, the best triode of all time,
>
> At least we agree on something......
>
> > and the CPI acts as a buffer between input tube and the balanced driver
> stage,
> > which results in a terrific bandwidth for the whole four triode driver
> stage.
> >
>
> So why are we not all building Williamson lookalikes ??

Why does the car industry produce more than one model?
Didn't you know, there are redheads, brunnettes, and blondes.
Just think if all women were clones, noboody would know whose wife was who's.


>
>
> > >
> > > Not that you will 'HEAR' whatever you put in as an inverter.
> > > What make ???................. My $3 Chinese 6SN7s sound great, NOS
> Brimars
> > > sound pretty much the same.... you choose.....
> > > regards
> > > jim
> >
> > I don't hear much difference. A PI is just another stage you must have for
> > a PP amp, unless the amp has a balanced input,
> > then no PI is required.
>
> Well. yes.. You want to even try thinking about the earth/earth loop
> problems with seperate turntable, CD player, tape, radio, etc......

I don't have hum problems, so I get by with half the tubes used for all
balanced.

Patrick Turner.

[Ian Iveson]

We agree on nearly everything Jim.

But there are matters of degree.

As perhaps you know by now my own amps are just 4*6CH6 with UL a
distributed load. As I run them now, I get 12W class A which is
plenty into my not-very-sensitive speakers and small room.

Apart from these fairly distortion-free watts, I like a few for what
I believe the experts call headroom. we're dealing with music here,
right? Not just testing with sine waves.

I feel it is important to consider what defines the edges of the
headroom. If I bias the my output valves to the recommended 40mA or
more each, then grid current comes before cutoff. Driven with a
typical cathodyne I am then lucky to get 9W before it totally falls
apart. A LTP copes much better but I don't like them. A paraphase
copes better still and that is what I might change to. Yes I could
make a bigger cathodyne...I don't need you to tell me that. But it
would have to be much bigger.

At the moment I get the 12 clean watts further towards AB1.
However, this means that the 1W I hear most of the time is not as
good as it could be. Not as linear, quite, and with a higher output
impedance. Now headroom is limited by soft clipping at first
before...the cathodyne crashes.

I don't want heaps of grid current, I just want a splitter that can
cope with a little every now and then.
I could add cathode followers or a short-tailed pair like Williamson
but that defeats the object of simplicity.

Yes, yes, yes you can say I just need bigger output valves. There
is more to these things than size though, Jim. I like as far as I
can to use finesse rather than brute force. We appear to have very
different attitudes in this respect.

cheers, Ian


"jim" > wrote in message
...
>
> "Ian Iveson" > wrote in message
> ...
> > Also, most importantly in my case, it depends on the linearity
of
> > the driven impedance, and hence on how the output stage is
biased.
> > Not just on how big it is. One of Van der Veen's designs, for
> > example, uses a cathodyne ECC81 for 6*EL34.
> >
> > But the slightest bit of grid current causes the cathodyne anode
> > output to collapse resulting in a flurry of odd harmonics at the
> > output of the amp.
>
> Hi Ian,
> Is this suprising ?? It is working entirely outside it's design
> envelope. My wife's Peugeot diesel is easily outpaced by a
Ferrari.
> Should I complain. ??
> No inverter we are using is designed to supply current. If you
need a
> current supply to the OP valve grids you need CFs or a transformer
hanging
> on them. This is AB2 or B. Neither are related to high end
audio.
> They are about a big noise, efficiency and bangs per buck
> Nothing wrong with Menno Van der Veens concertina driving 6 x
EL34s. Not
> seen the circuit but assume it's cathode biased to retain a fairly
high grid
> impedance for the concertina to work into. Parallelling 6 x EL34s
provides
> the output power without asking the concertina to provide a big
voltage
> swing and keeps the circuit simple and stable.
> >
> > I think there is an issue of slew rate limiting too, and/oce at
HF.
> Current-derived cathode feedback is good for
> > turning a problem into a crisis.
>
> There is an imbalance between the cathode and anode outputs but
this is
> usually outside the limits of audibility. This is not a problem if
we can't
> hear it, but since the concertina has no gain it is usually
necessary to
> insert another gain stage to get the drive voltages we need for
large
> valves. A Williamson. We now have four stages and since each
stage shifts
> phase, complete inversion can occur at the extremes of frequency.
It is
> unstable. We have an oscillator. Add in global NFB and unless
the
> bandwidth is restricted it may well all get away from you.
>
>
> > So the cathodyne should be restricted to AB1 output stages, and
it
> > should idle with plenty of current. An ECC83 is unlikely to be
a
> > good candidate. An '82 or an '81 is better, as might be an '88.
> > You have to get enough voltage swing and enough current
capability
> > for the task.
>
> All the inverters we use should be restricted to AB1 None of them
are
> current generators.
>
> > A LTP makes a good differential amplifier and you can call it a
> > splitter or an inverter if you want, but it's a very inelegant
> > approach, IMO.
>
> Only the concertina could be classed as elegant...with a direct
coupled
> voltage amp up front. The LTP and paraphase are a compromise of a
gain and
> inverter stage combined.
> Patrick has probably taken the LTP as far as it can go and tells
me I can
> have a 6SN7 as one half and an ECC83 as the other and, with a CCJ
tail, will
> still achieve perfect balance of the two outputs. I cannot
dispute this,
> he knows what he is doing.
> The paraphase ?? Seems even more difficult to balance and keep
balanced and
> I have rarely used it.
> By combining inversion and gain in one stage both the paraphase
and the LTP
> do have the advantage of of reducing the number of stages and
suggest a more
> stable design.
>
> > What we need to do is to create an inverted copy of the original
> > signal. This is more than inverting, or splitting (whatever
that
> > means). The copying part naturally raises the issue of
distortion.
> >
> > The paraphase might seem awkward for those looking for symmetry.
> > But you why should copying and inverting be a symmetrical
process?
> > The paraphase uses a valve to produce an inverted copy.
Perfect.
> > Tricky to get the details right so it stays in balance of
course,
> > but so is the LTP.
> >
> > All in all, the cathodyne used as a driver ensures balance and
low
> > distortion only within a comparatively narrow corridor of
operation,
> > outside of which it goes badly awry. A paraphase is always
slightly
> > awry but maintains this over a much wider proportion of
available HT
> > and load conditions. A LTP falls somewhere between the two.
> >
> > Really you have to consider the amp as a whole.
>
> Exactly. It is possible to build fine 10-15W amplifiers with
simple
> circuitry and few components. These are maximum ratings however
and with
> inefficient speakers you could be running close to overload at
normal
> listening volumes. To design a 15W amp and ask it to sound good at
20W is a
> non starter.
> I like 50W as a design target. Times two for stereo. The whole
question of
> pushing it to the design limits and watching it get out of shape
does not
> then occur. Most of what I am ever going to hear is Class A. No
crossover
> distortion and a driver design which can put 90v Pk-pk on the OP
valve grids
> without really trying.
> The drawback is that including the PSU, it is big, heavy, and
expensive
> kind regards
> jim
>
>
> For example, how
> > much HT you have available depends partly on how much hum you
can
> > tolerate at the driver stage, which also depends on how good the
hum
> > rejection is of the driver.
> >
> > cheers, Ian
> >
> >
> > "jim" > wrote in message
> > ...
> > >
> > > "Armand" > wrote in message
> > > ...
> > > > I know that the sections should be well matched, but how
much
> > does the PI
> > > > affect the overall sound of an amp, gain notwithstanding?
> > Thanks.
> > > >
> > >
> > > Hi, Armand,
> > >
> > > Short answer. It doesn't.!!!!!!
> > > I have often rebuilt the same PP design with different
inverters.
> > The
> > > distortion contributed by the inverter stage is nothing
compared
> > to that
> > > produced by the following stages, and you certainly won't hear
it,
> > provided
> > > the two antiphase outputs are closely balanced.
> > > So why are there different inverter configurations and which
one
> > is best ??
> > > Only two inverters seem to be regularly used now. The
cathodyne
> > and the LTP.
> > > Which you choose is not dictated by which sounds best, it is
> > decided by the
> > > AC output required, the impedance of the output load you are
> > driving, and
> > > the available B+ voltage.
> > > kind regards
> > > jim
> > >
> > >
> >
> >
> >
> >
> >
> >
>
>
>

[TubeGarden]

>At last. Someone else realises a couple of KT88 are always going to do better
than a pair of EL34, or '84.
> - Patrick Turner.

Hi RATs!

Do is not out there as a tidy, independent reality.

It and we share time and space.

Well, temporarily ;)

Happy Ears!
Al


Alan J. Marcy
Phoenix, AZ

PWC/mystic/Earhead

[Patrick Turner]

TubeGarden wrote:

> >At last. Someone else realises a couple of KT88 are always going to do better
> than a pair of EL34, or '84.
> > - Patrick Turner.
>
> Hi RATs!
>
> Do is not out there as a tidy, independent reality.
>
> It and we share time and space.
>
> Well, temporarily ;)
>
> Happy Ears!
> Al
>
> Alan J. Marcy
> Phoenix, AZ

John paul Satre said "to do is to be", something my ex-wife never realised,
somebody also said, "to be is to do"
and maybe he was Dr Doolittle.

But Frank Sinatra had it right, he just sang
"tobe to beto to beto toobe..."
and to remember what to sing, he had to read the words from a
prompt carefully set up.

>
>
> PWC/mystic/Earhead

Feel fee to be as mystical as you wish, its the weekend!

Patrick Turner.

[Patrick Turner]

jim wrote:

> Hi, Ian,
> .
> > We agree on nearly everything Jim.
>
> That's unusual. People rarely agree with anything I say !!
>
> > But there are matters of degree.
>
> Of course. A design target of 1200W for home listening is unnecessary ...
> On the other hand, the 3W from my old Dansette is inadequate.
>
> > As perhaps you know by now my own amps are just 4*6CH6 with UL a
> > distributed load. As I run them now, I get 12W class A which is
> > plenty into my not-very-sensitive speakers and small room.
> >
> > Apart from these fairly distortion-free watts, I like a few for what
> > I believe the experts call headroom. we're dealing with music here,
> > right? Not just testing with sine waves.
>
> Exactly, testing is carried out with ajustable amplitude waveforms.
> Increasing the amplitude finds the point where unacceptable distortion can
> be measured and this is the maximum useful output power of the amplifier
> A music source is not of fixed amplitude, and as we have discussed before,
> there appears to be no standard governing by how much transient peaks can
> exceed the mean level. Since you have to set a lower volume level to allow
> for these peaks, how much rated power do we need ??

It depends on speakers, don't it?
if you had 95 Db/watt speakers, 12 watts id fine.
to get the same dynamic headroom, low thd etc, with modern speakers
of 89 dB, we should have 48 watts.
I still find a 12 watt amp with my 89 dB speakers isn't too bad.

>
>
> > I feel it is important to consider what defines the edges of the
> > headroom.
>
> You cannot do this with any measure of accuracy and put figures to it. We
> have discussed this before and you know what I am going to say. I only
> repeat it because , apart from you, there may be somebody else out there who
> thinks I know what I am talking about !!!
> Our ears hear best at mid frequencies. Depending on the music content the
> volume control has to be adjusted accordingly to maintain the same percieved
> loudness. 50W output at 5Hz is no problem for my ears but the KT88s are
> still working very hard. My wife says I am deaf because I play music at
> high levels. I prefer it that way. Her ears may be more sensitive.

Indeed, and women are queer about men and hi-fi;
they are likely to think we need a hearing aid, rather than more watts,
and besides, they subconsciously hate anything occuring in the home which
competes with attention and expenditure in their direction.
Women exist to drain men of goods and services, and they are like bottomless
chasms to be filled up with all you have, and more, more, more!
Hi-fi represents cheeky male assertions, and like sex, isn't
comprehended by female brains as audiological erotica,
something sensuous, to be savoured. Women use sex for power,
and its the power over men that turns them on, not the dirty act itself.
But music, and musing by men with glasses of cabernet is indulgence,
and independant indulgence, and the interest in tubes and speacial gear
is like a mistress.
My ex-wife truly hated my hi-fi interest, and resented my not planning for a
holiday
under palm trees at horrendous expense. She NEVER brought a cup of coffee
to me in the garage as I worked to build things I thought WE might need
to live.
Other wives I knew of my peer group all those years ago dreaded
their dopey husband's efforts restoring old cars, and working on the yard, or
the house.
A lot of the women sulked about, as we seemed to enjoy all this stuff with our
mates.

One older shiela said of my loungeroom, recently, and disparingly
"its very like a batchelor's room"
I just smiled, and thought " and WTF would you do to make life for US more
better?"


>
> We also hear logarithmically, hence log pots for volume control. Twice as
> loud as 10W is 100W. Add in the variables of speaker efficiency, and room
> size and how much power do you actually need.?? Even the furnishings come
> in here. Take all the furniture and carpets out and anything else which
> absorbes rather than reflects sound energy.

Rooms are very important, and a room without carpets, drapes,
lots of cushions, is indeed an audio horror, besides, no place to sit down....


> But you know all this, but the question ---- is 10W enough ? --- has no
> answer.
> W=VI and is therefore a very poor way to measure sound levels or judge where
> to start a design, and this is unfortunately where the design has to
> start -- at the end
>
> If I bias the my output valves to the recommended 40mA or
> > more each, then grid current comes before cutoff. Driven with a
> > typical cathodyne I am then lucky to get 9W before it totally falls
> > apart.
>
> I thought you had a clean 12W ??

If he has class A, then grid current indeed occurs before cut off,
because the Gm of the tubes is more when the tubes are being turned on, than
off,
and the 3H distortion in class A is from the difference in complementary Gm
and and when grid current does occur, flattening of the sine wave peaks begins
on both top and bottom of the sine wave, and when 3H reaches 3%,
with a sudden increase with a small % in output voltage, clipping is deemed to
have begun.
and that's where maximum AB power is said to be, and for 4 x EL84, depending on
load,
and bias, this could be up to around 75 watts with B+ = 700v.
Maximum class A power it could be 17 watts,
if maximum idle power was 48 watts, and efficiency was 35%.

>
>
> A LTP copes much better but I don't like them. A paraphase
> > copes better still and that is what I might change to. Yes I could
> > make a bigger cathodyne...I don't need you to tell me that. But it
> > would have to be much bigger.
>
> I don't understand this bit. If you are driving the OP valves into grid
> current or cutoff, why would altering the inverter improve things. ??

The cathodyne or CPI clips symetrically when the onset of grid current occurs,
but AFTER innitial clipping things get very asymetical.
The anode circuit has less power output into the charging up of grid coupling
caps to the OPVs, and the clipping is asymetrical, ie, the clipped wave has 2H
in the squared waves.
LTPs clip more evenly on each side of 0V, and seem to recover better.

But in hi-fi situations, we never go near clipping;, its only of interest to
rock
and roll artistes. They adore square waves.

>
> Remember I am a hobby boy, not Patrick, but if you are driving the grids
> positive, AB2, and looking for an inverter which can supply this drive, the
> parahase and LTP have high OP impedances and are not very good at supplying
> current. Undistorted output watts is decided by the available HT voltage,
> OPT load impedance, grid drive voltage and max anode dissipation of the OP
> valves isn't it ?? Load lines ...Everything in front of the OP valves is
> just there to provide the voltage swing to make it happen. Changing the
> inverter/driver won't get 50W out of an EL84....... You do not want to get
> into AB2, Ian, .. all sorts of nasty things happen there .......

Some tubes, like the 807/6L6 thrive in well concieved AB2 designs.
AFAIK, EL84, and EL34, just don't.

One can build a triode amp with 807, and get a lousy 14 watts AB1 fairly easy,
regardless of what phase inverter is used, if it is used as the driver, either
paraphase, LTP
or CPI will do, providing enough undistorted drive voltage is available.

Then indeed, grid current limits the power output.
With beam tetrode or pentode operation, the drive voltage never gets near
causing grid current, as the load swing voltage is limited by the
plate characteristics and the B+ rails.
So we get 25 watts from the same 807 in class A.

But say we use whatever drive amp we want with a a cathode follower
to each OPV grid, directly coupled, then the triode class A1 power out
could still be 14 watts, but we could get an extra 12 watts with the same load.
pushing into class AB2, since the CF input grid will not itself draw grid
current
when driving the output grids. There will be some compression, or increased
3H, but NFB can minimise this, and the triode power is quite high.
30 watts of AB2 triode power isn't unrealisable.
If the B+ for 807 is raised to 750V
and the screens kept at 300v, and the load a-a raised to 10ka-a,
then we can extract a heap more power in class AB2 beam tetrode mode,
maybe 80+ watts. THD tends to be quite high.

>
> > At the moment I get the 12 clean watts further towards AB1.
> > However, this means that the 1W I hear most of the time is not as
> > good as it could be. Not as linear, quite, and with a higher output
> > impedance. Now headroom is limited by soft clipping at first
> > before...the cathodyne crashes.
> >
> > I don't want heaps of grid current, I just want a splitter that can
> > cope with a little every now and then.
> > I could add cathode followers or a short-tailed pair like Williamson
> > but that defeats the object of simplicity.

Gee, why worry?

>
>
> I have a sledgehammer approach to valve audio which actually makes design
> very simple. I have to say that 12W is just not enough. Consider a slightly
> Williamson circuit which can make a pair of 60vRMS signals available on the
> OP valve grids... At that level, I have to go outside..
> If I want to come back inside, the drivers are loafing along at 15-20vRMS,
> the room's full of noise, and I have to look very closely at the 'scope for
> anything out of shape.

If I want quiet backgound sound, on my 89db speakers, I use SE pentode
amps in my ancient Trio AM/FM reciever.
Then if I have ppl around to hear something special, I might use my
100w monoblocs, with their 40 watts class A ability,
but more often I'll use the 22 w SEUL amps, and these
seem to sound loud, clean, detailed, and more like 30 watt amps,
and with the magical warm midrange.

Would I sit in the front row with an orchestra of 100 players?
No way, I'd get way back.
But at age 25 I might have done so happily.

>
>
> > Yes, yes, yes you can say I just need bigger output valves. There
> > is more to these things than size though, Jim. I like as far as I
> > can to use finesse rather than brute force. We appear to have very
> > different attitudes in this respect.

Tubes ain't like sex, I mean, if the willy is small, use it twice a night,
but it don't make sense with tubes, if you need more than a small tube,
you must use a bigger one, or more of the small ones in parallel,
all acting together.
Baird make amps with 6 x EL84, and they have a fine sound.
EL84 is a fine tube, my favourite smally.
Musical Reference run 2 x EL84 to get 36 watts with 14k a-a,
and 700 v for Ea. Definately class AB1. Its possible to use a
CPI driver stage with a single gain stage ahead of it.

Then of course we haven't even touched on the Dynaco ST70
method, which uses the CPI to provide a bootstrapped load for the input pentode,

so the pentode gain is unusually high, so that when 20 dB od NFB
is applied, the amp is a lot more sensitive than a williamson, Quad II,
or the 10-10 w amp at my website, used so often in so many countless
receiver or amp designs.


>
>
> We do, but electricity is cheap and valves are very inefficient. You are an
> academic and will search out efficiency and numbers which add up. I don't
> think valve audio is about that at all...

Thank goodness it isn't.
Using a PC or an electric heater is wasteful.
But it is all hugely innefficient, modern life, as humanity exhausts the earth's

reserves of non renewable energy sources to uphold a way of life.
we will have to genetically alter ouselves to eat garbage at rubbish tips,
and then create alternative energy forms in the thousands of years ahead.

Don't be too concerned about the use of a few tubes within our life times,
and don't scorn those whose interest it is to restore steam locomotives,
or old cars. Don't let the modern technocrats make you feel guilty
running triode amps, when those same *******s drive expensive cars to
airconditioned offices
and comfy chairs to get into a position to rave at us.


> Silicon is, but valves are about a
> big warm open sound and a smile on your face.

Hear,hear!

>
> I think you need a 2 x 100w KT88 stereo pair for a few minutes, and work
> back from there..

Does he?

>
>
> regards as always
> jim

I hope youse have a great weekend.

Patrick Turner.

>
>
> > cheers, Ian
> >
> >
> > "jim" > wrote in message
> > ...
> > >
> > > "Ian Iveson" > wrote in message
> > > ...
> > > > Also, most importantly in my case, it depends on the linearity
> > of
> > > > the driven impedance, and hence on how the output stage is
> > biased.
> > > > Not just on how big it is. One of Van der Veen's designs, for
> > > > example, uses a cathodyne ECC81 for 6*EL34.
> > > >
> > > > But the slightest bit of grid current causes the cathodyne anode
> > > > output to collapse resulting in a flurry of odd harmonics at the
> > > > output of the amp.
> > >
> > > Hi Ian,
> > > Is this suprising ?? It is working entirely outside it's design
> > > envelope. My wife's Peugeot diesel is easily outpaced by a
> > Ferrari.
> > > Should I complain. ??
> > > No inverter we are using is designed to supply current. If you
> > need a
> > > current supply to the OP valve grids you need CFs or a transformer
> > hanging
> > > on them. This is AB2 or B. Neither are related to high end
> > audio.
> > > They are about a big noise, efficiency and bangs per buck
> > > Nothing wrong with Menno Van der Veens concertina driving 6 x
> > EL34s. Not
> > > seen the circuit but assume it's cathode biased to retain a fairly
> > high grid
> > > impedance for the concertina to work into. Parallelling 6 x EL34s
> > provides
> > > the output power without asking the concertina to provide a big
> > voltage
> > > swing and keeps the circuit simple and stable.
> > > >
> > > > I think there is an issue of slew rate limiting too, and/oce at
> > HF.
> > > Current-derived cathode feedback is good for
> > > > turning a problem into a crisis.
> > >
> > > There is an imbalance between the cathode and anode outputs but
> > this is
> > > usually outside the limits of audibility. This is not a problem if
> > we can't
> > > hear it, but since the concertina has no gain it is usually
> > necessary to
> > > insert another gain stage to get the drive voltages we need for
> > large
> > > valves. A Williamson. We now have four stages and since each
> > stage shifts
> > > phase, complete inversion can occur at the extremes of frequency.
> > It is
> > > unstable. We have an oscillator. Add in global NFB and unless
> > the
> > > bandwidth is restricted it may well all get away from you.
> > >
> > >
> > > > So the cathodyne should be restricted to AB1 output stages, and
> > it
> > > > should idle with plenty of current. An ECC83 is unlikely to be
> > a
> > > > good candidate. An '82 or an '81 is better, as might be an '88.
> > > > You have to get enough voltage swing and enough current
> > capability
> > > > for the task.
> > >
> > > All the inverters we use should be restricted to AB1 None of them
> > are
> > > current generators.
> > >
> > > > A LTP makes a good differential amplifier and you can call it a
> > > > splitter or an inverter if you want, but it's a very inelegant
> > > > approach, IMO.
> > >
> > > Only the concertina could be classed as elegant...with a direct
> > coupled
> > > voltage amp up front. The LTP and paraphase are a compromise of a
> > gain and
> > > inverter stage combined.
> > > Patrick has probably taken the LTP as far as it can go and tells
> > me I can
> > > have a 6SN7 as one half and an ECC83 as the other and, with a CCJ
> > tail, will
> > > still achieve perfect balance of the two outputs. I cannot
> > dispute this,
> > > he knows what he is doing.
> > > The paraphase ?? Seems even more difficult to balance and keep
> > balanced and
> > > I have rarely used it.
> > > By combining inversion and gain in one stage both the paraphase
> > and the LTP
> > > do have the advantage of of reducing the number of stages and
> > suggest a more
> > > stable design.
> > >
> > > > What we need to do is to create an inverted copy of the original
> > > > signal. This is more than inverting, or splitting (whatever
> > that
> > > > means). The copying part naturally raises the issue of
> > distortion.
> > > >
> > > > The paraphase might seem awkward for those looking for symmetry.
> > > > But you why should copying and inverting be a symmetrical
> > process?
> > > > The paraphase uses a valve to produce an inverted copy.
> > Perfect.
> > > > Tricky to get the details right so it stays in balance of
> > course,
> > > > but so is the LTP.
> > > >
> > > > All in all, the cathodyne used as a driver ensures balance and
> > low
> > > > distortion only within a comparatively narrow corridor of
> > operation,
> > > > outside of which it goes badly awry. A paraphase is always
> > slightly
> > > > awry but maintains this over a much wider proportion of
> > available HT
> > > > and load conditions. A LTP falls somewhere between the two.
> > > >
> > > > Really you have to consider the amp as a whole.
> > >
> > > Exactly. It is possible to build fine 10-15W amplifiers with
> > simple
> > > circuitry and few components. These are maximum ratings however
> > and with
> > > inefficient speakers you could be running close to overload at
> > normal
> > > listening volumes. To design a 15W amp and ask it to sound good at
> > 20W is a
> > > non starter.
> > > I like 50W as a design target. Times two for stereo. The whole
> > question of
> > > pushing it to the design limits and watching it get out of shape
> > does not
> > > then occur. Most of what I am ever going to hear is Class A. No
> > crossover
> > > distortion and a driver design which can put 90v Pk-pk on the OP
> > valve grids
> > > without really trying.
> > > The drawback is that including the PSU, it is big, heavy, and
> > expensive
> > > kind regards
> > > jim
> > >
> > >
> > > For example, how
> > > > much HT you have available depends partly on how much hum you
> > can
> > > > tolerate at the driver stage, which also depends on how good the
> > hum
> > > > rejection is of the driver.
> > > >
> > > > cheers, Ian
> > > >
> > > >
> > > > "jim" > wrote in message
> > > > ...
> > > > >
> > > > > "Armand" > wrote in message
> > > > > ...
> > > > > > I know that the sections should be well matched, but how
> > much
> > > > does the PI
> > > > > > affect the overall sound of an amp, gain notwithstanding?
> > > > Thanks.
> > > > > >
> > > > >
> > > > > Hi, Armand,
> > > > >
> > > > > Short answer. It doesn't.!!!!!!
> > > > > I have often rebuilt the same PP design with different
> > inverters.
> > > > The
> > > > > distortion contributed by the inverter stage is nothing
> > compared
> > > > to that
> > > > > produced by the following stages, and you certainly won't hear
> > it,
> > > > provided
> > > > > the two antiphase outputs are closely balanced.
> > > > > So why are there different inverter configurations and which
> > one
> > > > is best ??
> > > > > Only two inverters seem to be regularly used now. The
> > cathodyne
> > > > and the LTP.
> > > > > Which you choose is not dictated by which sounds best, it is
> > > > decided by the
> > > > > AC output required, the impedance of the output load you are
> > > > driving, and
> > > > > the available B+ voltage.
> > > > > kind regards
> > > > > jim
> > > > >
> > > > >
> > > >
> > > >
> > > >
> > > >
> > > >
> > > >
> > >
> > >
> > >
> >
> >

[jim]

Hi, Ian,
..
> We agree on nearly everything Jim.

That's unusual. People rarely agree with anything I say !!

> But there are matters of degree.

Of course. A design target of 1200W for home listening is unnecessary ...
On the other hand, the 3W from my old Dansette is inadequate.


> As perhaps you know by now my own amps are just 4*6CH6 with UL a
> distributed load. As I run them now, I get 12W class A which is
> plenty into my not-very-sensitive speakers and small room.
>
> Apart from these fairly distortion-free watts, I like a few for what
> I believe the experts call headroom. we're dealing with music here,
> right? Not just testing with sine waves.

Exactly, testing is carried out with ajustable amplitude waveforms.
Increasing the amplitude finds the point where unacceptable distortion can
be measured and this is the maximum useful output power of the amplifier
A music source is not of fixed amplitude, and as we have discussed before,
there appears to be no standard governing by how much transient peaks can
exceed the mean level. Since you have to set a lower volume level to allow
for these peaks, how much rated power do we need ??

> I feel it is important to consider what defines the edges of the
> headroom.

You cannot do this with any measure of accuracy and put figures to it. We
have discussed this before and you know what I am going to say. I only
repeat it because , apart from you, there may be somebody else out there who
thinks I know what I am talking about !!!
Our ears hear best at mid frequencies. Depending on the music content the
volume control has to be adjusted accordingly to maintain the same percieved
loudness. 50W output at 5Hz is no problem for my ears but the KT88s are
still working very hard. My wife says I am deaf because I play music at
high levels. I prefer it that way. Her ears may be more sensitive.
We also hear logarithmically, hence log pots for volume control. Twice as
loud as 10W is 100W. Add in the variables of speaker efficiency, and room
size and how much power do you actually need.?? Even the furnishings come
in here. Take all the furniture and carpets out and anything else which
absorbes rather than reflects sound energy.
You know all this, but the question ---- is 10W enough ? --- has no answer.
W=VI and is therefore a very poor way to measure sound levels or judge where
to start a design, and this is unfortunately where the design has to
start -- at the end

If I bias the my output valves to the recommended 40mA or
> more each, then grid current comes before cutoff. Driven with a
> typical cathodyne I am then lucky to get 9W before it totally falls
> apart.

I thought you had a clean 12W ??

A LTP copes much better but I don't like them. A paraphase
> copes better still and that is what I might change to. Yes I could
> make a bigger cathodyne...I don't need you to tell me that. But it
> would have to be much bigger.

I don't understand this bit. If you are driving the OP valves into grid
current or cutoff, why would altering the inverter improve things. ??
Remember I am a hobby boy, not Patrick, but if you are driving the grids
positive, AB2, and looking for an inverter which can supply this drive, the
parahase and LTP have high OP impedances and are not very good at supplying
current. Undistorted output watts is decided by the available HT voltage,
OPT load impedance, grid drive voltage and max anode dissipation of the OP
valves isn't it ?? Load lines ...Everything in front of the OP valves is
just there to provide the voltage swing to make it happen. Changing the
inverter/driver won't get 50W out of an EL84....... You do not want to get
into AB2, Ian, .. all sorts of nasty things happen there .......

> At the moment I get the 12 clean watts further towards AB1.
> However, this means that the 1W I hear most of the time is not as
> good as it could be. Not as linear, quite, and with a higher output
> impedance. Now headroom is limited by soft clipping at first
> before...the cathodyne crashes.
>
> I don't want heaps of grid current, I just want a splitter that can
> cope with a little every now and then.
> I could add cathode followers or a short-tailed pair like Williamson
> but that defeats the object of simplicity.

I have a sledgehammer approach to valve audio which actually makes design
very simple. I have to say that 12W is just not enough. Consider a slightly
Williamson circuit which can make a pair of 60vRMS signals available on the
OP valve grids... At that level, I have to go outside..
If I want to come back inside, the drivers are loafing along at 15-20vRMS,
the room's full of noise, and I have to look very closely at the 'scope for
anything out of shape.

> Yes, yes, yes you can say I just need bigger output valves. There
> is more to these things than size though, Jim. I like as far as I
> can to use finesse rather than brute force. We appear to have very
> different attitudes in this respect.

We do, but electricity is cheap and valves are very inefficient. You are an
academic and will search out efficiency and numbers which add up. I don't
think valve audio is about that at all... Silicon is, but valves are about a
big warm open sound and a smile on your face.
I think you need a 2 x 100w KT88 stereo pair for a few minutes, and work
back from there..

regards as always
jim

> cheers, Ian
>
>
> "jim" > wrote in message
> ...
> >
> > "Ian Iveson" > wrote in message
> > ...
> > > Also, most importantly in my case, it depends on the linearity
> of
> > > the driven impedance, and hence on how the output stage is
> biased.
> > > Not just on how big it is. One of Van der Veen's designs, for
> > > example, uses a cathodyne ECC81 for 6*EL34.
> > >
> > > But the slightest bit of grid current causes the cathodyne anode
> > > output to collapse resulting in a flurry of odd harmonics at the
> > > output of the amp.
> >
> > Hi Ian,
> > Is this suprising ?? It is working entirely outside it's design
> > envelope. My wife's Peugeot diesel is easily outpaced by a
> Ferrari.
> > Should I complain. ??
> > No inverter we are using is designed to supply current. If you
> need a
> > current supply to the OP valve grids you need CFs or a transformer
> hanging
> > on them. This is AB2 or B. Neither are related to high end
> audio.
> > They are about a big noise, efficiency and bangs per buck
> > Nothing wrong with Menno Van der Veens concertina driving 6 x
> EL34s. Not
> > seen the circuit but assume it's cathode biased to retain a fairly
> high grid
> > impedance for the concertina to work into. Parallelling 6 x EL34s
> provides
> > the output power without asking the concertina to provide a big
> voltage
> > swing and keeps the circuit simple and stable.
> > >
> > > I think there is an issue of slew rate limiting too, and/oce at
> HF.
> > Current-derived cathode feedback is good for
> > > turning a problem into a crisis.
> >
> > There is an imbalance between the cathode and anode outputs but
> this is
> > usually outside the limits of audibility. This is not a problem if
> we can't
> > hear it, but since the concertina has no gain it is usually
> necessary to
> > insert another gain stage to get the drive voltages we need for
> large
> > valves. A Williamson. We now have four stages and since each
> stage shifts
> > phase, complete inversion can occur at the extremes of frequency.
> It is
> > unstable. We have an oscillator. Add in global NFB and unless
> the
> > bandwidth is restricted it may well all get away from you.
> >
> >
> > > So the cathodyne should be restricted to AB1 output stages, and
> it
> > > should idle with plenty of current. An ECC83 is unlikely to be
> a
> > > good candidate. An '82 or an '81 is better, as might be an '88.
> > > You have to get enough voltage swing and enough current
> capability
> > > for the task.
> >
> > All the inverters we use should be restricted to AB1 None of them
> are
> > current generators.
> >
> > > A LTP makes a good differential amplifier and you can call it a
> > > splitter or an inverter if you want, but it's a very inelegant
> > > approach, IMO.
> >
> > Only the concertina could be classed as elegant...with a direct
> coupled
> > voltage amp up front. The LTP and paraphase are a compromise of a
> gain and
> > inverter stage combined.
> > Patrick has probably taken the LTP as far as it can go and tells
> me I can
> > have a 6SN7 as one half and an ECC83 as the other and, with a CCJ
> tail, will
> > still achieve perfect balance of the two outputs. I cannot
> dispute this,
> > he knows what he is doing.
> > The paraphase ?? Seems even more difficult to balance and keep
> balanced and
> > I have rarely used it.
> > By combining inversion and gain in one stage both the paraphase
> and the LTP
> > do have the advantage of of reducing the number of stages and
> suggest a more
> > stable design.
> >
> > > What we need to do is to create an inverted copy of the original
> > > signal. This is more than inverting, or splitting (whatever
> that
> > > means). The copying part naturally raises the issue of
> distortion.
> > >
> > > The paraphase might seem awkward for those looking for symmetry.
> > > But you why should copying and inverting be a symmetrical
> process?
> > > The paraphase uses a valve to produce an inverted copy.
> Perfect.
> > > Tricky to get the details right so it stays in balance of
> course,
> > > but so is the LTP.
> > >
> > > All in all, the cathodyne used as a driver ensures balance and
> low
> > > distortion only within a comparatively narrow corridor of
> operation,
> > > outside of which it goes badly awry. A paraphase is always
> slightly
> > > awry but maintains this over a much wider proportion of
> available HT
> > > and load conditions. A LTP falls somewhere between the two.
> > >
> > > Really you have to consider the amp as a whole.
> >
> > Exactly. It is possible to build fine 10-15W amplifiers with
> simple
> > circuitry and few components. These are maximum ratings however
> and with
> > inefficient speakers you could be running close to overload at
> normal
> > listening volumes. To design a 15W amp and ask it to sound good at
> 20W is a
> > non starter.
> > I like 50W as a design target. Times two for stereo. The whole
> question of
> > pushing it to the design limits and watching it get out of shape
> does not
> > then occur. Most of what I am ever going to hear is Class A. No
> crossover
> > distortion and a driver design which can put 90v Pk-pk on the OP
> valve grids
> > without really trying.
> > The drawback is that including the PSU, it is big, heavy, and
> expensive
> > kind regards
> > jim
> >
> >
> > For example, how
> > > much HT you have available depends partly on how much hum you
> can
> > > tolerate at the driver stage, which also depends on how good the
> hum
> > > rejection is of the driver.
> > >
> > > cheers, Ian
> > >
> > >
> > > "jim" > wrote in message
> > > ...
> > > >
> > > > "Armand" > wrote in message
> > > > ...
> > > > > I know that the sections should be well matched, but how
> much
> > > does the PI
> > > > > affect the overall sound of an amp, gain notwithstanding?
> > > Thanks.
> > > > >
> > > >
> > > > Hi, Armand,
> > > >
> > > > Short answer. It doesn't.!!!!!!
> > > > I have often rebuilt the same PP design with different
> inverters.
> > > The
> > > > distortion contributed by the inverter stage is nothing
> compared
> > > to that
> > > > produced by the following stages, and you certainly won't hear
> it,
> > > provided
> > > > the two antiphase outputs are closely balanced.
> > > > So why are there different inverter configurations and which
> one
> > > is best ??
> > > > Only two inverters seem to be regularly used now. The
> cathodyne
> > > and the LTP.
> > > > Which you choose is not dictated by which sounds best, it is
> > > decided by the
> > > > AC output required, the impedance of the output load you are
> > > driving, and
> > > > the available B+ voltage.
> > > > kind regards
> > > > jim
> > > >
> > > >
> > >
> > >
> > >
> > >
> > >
> > >
> >
> >
> >
>
>

[jim]

>
> There is little need for high current bias flow in an LTPdriver/CPI driver
were
> it not
> for the input C of the stage being driven.
> The higher the idle current, the higher the bandwidth of the driver stage,
> since the capacitance can be charged and discharged with enough current
> to maintain the voltage swing at HF.

Hiya, Patrick...
> In the case of the CPI, a small C across the Rk of the stage will boost
> the HF response from the anode, which tends to roll off first.
>

Yes, O.K. ....


> > If you need a
> > current supply to the OP valve grids you need CFs or a transformer
hanging
> > on them. This is AB2 or B. Neither are related to high end audio.
> > They are about a big noise, efficiency and bangs per buck
> > Nothing wrong with Menno Van der Veens concertina driving 6 x EL34s. Not
> > seen the circuit but assume it's cathode biased to retain a fairly high
grid
> > impedance for the concertina to work into.
>
> But the cathode bias itself causes input grid impedance to be no higher
> than fixed bias, since grid current occurs only when the peak swing
> input voltage goes higher than the value of grid bias.


I'm looking at the load the driver has to work into.. Cathode bias does
not alter the grid impedance per se, as you say, but the grid leak can be
made higher because of the self regulating effect of the cathode resistor.
The grid leak is the load. We have 6 EL34s here with grid leaks in parallel.
I'm just trying to avoid a coupling cap the size of a 45 gallon oil drum
and a time constant round about last Thursday !!!!

> > Parallelling 6 x EL34s provides
> > the output power without asking the concertina to provide a big voltage
> > swing and keeps the circuit simple and stable.
>
> 12 x EL84 would do about the same thing, and be even easier to drive!.
>

You trying to put one over on me ?? I ain't going to apologise for being a
hobby boy. Go up one paragraph ... 12 EL34s ??? Driver design, please ??
You planning on a Dodge or a Chrysler to move this lot ??

> > > I think there is an issue of slew rate limiting too, and/oce at HF.
> > Current-derived cathode feedback is good for
> > > turning a problem into a crisis.
> >
> > There is an imbalance between the cathode and anode outputs but this is
> > usually outside the limits of audibility. This is not a problem if we
can't
> > hear it, but since the concertina has no gain it is usually necessary to
> > insert another gain stage to get the drive voltages we need for large
> > valves. A Williamson. We now have four stages and since each stage
shifts
> > phase, complete inversion can occur at the extremes of frequency. It is
> > unstable. We have an oscillator. Add in global NFB and unless the
> > bandwidth is restricted it may well all get away from you.
>
> But to STOP the oscillation, which is normally due to a lousy OPT,
> and NOT the williamson driver circuit, it is necessary to place a zobel
network
> across the main RL of V1, and williamson recommended a 470pF and 4.7k
> in parallel with the 47 k load of the 1/2 6SN7 which was V1.

Twelve months ago you took my head off about dumping things and introducing
extra time constants when I suggested restricting the bandwidth by dumping
things you can't hear across V1R and removing them from the feedback chain
....... Whether the OPT is lousy or wonderful, a C across V1R will kill it
stone dead....... Why would I want to listen to anything over 30kHz ??

> The Williamson 3 stage driver circuit has more bandwidth than an LTP using
> the same tubes, since the CPI acts as a buffer against the effects of the
> miller C of the balanced diver stage causing a pole too early at V1,
> whose Ro is high because the cathode R isn't bypassed.

So, bypass it. I usually do....or partially, if I need a feedback point.

> The stepped HF response raises the pole F caused by the driver stage,
> but reduces its gain by 12 dB at this HF, ie, about 200 kHz, and
> therefore the amount of applied NFB at 200 kHz is lowered,
> and stability is increased.

Patrick, we're in the medium waveband here ....... Your designs may have no
effect on Radio Canberra, but personally, I don't give a **** whether Radio
Canberra gets my LPs full on or not. It will probably be better than the
**** they normally play...... Adjust the CR around V1 and 200kHz isn't in
there.
What does 200kHz sound like ?? I may have it and not know it !! Tune in ,
now, I'm going to put a Roy Orbison track on !.......

> Suitable zobel networks are shown also at my website schematics.

You dump the **** you can't hear so it can't come back and bite you ..

> The bandwidth of all my amps are tailored to at least 65 kHz,
> and any sized cap can be connected at the output without
> oscillations occuring. It is unconditional stability.
>
> >
> >
> > > So the cathodyne should be restricted to AB1 output stages, and it
> > > should idle with plenty of current.
>
> Indeed, but if CF bufferes after the cathodyne (CPI) are used, direct
coupled to
> the OPV grids,
> no worries, you can go AB2, if the voltage swing is wide enough.

But Ian is going for elegant simplicity. He doesn't want a an extra ECC 81
stuck in there.. He doesn't want to be into AB2. He doesn't really want to
be in AB1

> > An ECC83 is unlikely to be a
> > > good candidate. An '82 or an '81 is better, as might be an '88.
> > > You have to get enough voltage swing and enough current capability
> > > for the task.
> >
> > All the inverters we use should be restricted to AB1 None of them are
> > current generators.
>
> I think you have that wrong.

I may have. I will rephrase it. None are low impedance current generators.
As I see it, a low impedance driver for AB2 cannot include a 100K anode
load resistor and a coupling cap as in parahase/LTP,
hanging on the OP valve grid. I may even have this wrong. You will probably
put me right...


> None are voltage sources, like a cathode follower.
> Most drivers/inverters have highish Ro, and are limited current sources.
> Pentodes are current sources, since the voltage is determined by the load,
> andout put current
> is more nearly constant, and voltage is anywhere.
>
Eh ??

> >
> > > A LTP makes a good differential amplifier and you can call it a
> > > splitter or an inverter if you want, but it's a very inelegant
> > > approach, IMO.
>
> I think its very elegant.
> What makes it inelegant?
>
> Is is the hobnail boots and hairy legs worn with the pink frock?

No, it's all the bits and pieces dangling off the side of it which makes it
look a complete ****. !!!!
How can you trust anything which looks like that !!! Christ !! It's
American !!! Where's me concertina ??

> > Only the concertina could be classed as elegant...with a direct coupled
> > voltage amp up front. The LTP and paraphase are a compromise of a gain
and
> > inverter stage combined.
>
> Compromise? Have you been drinking too much cheap english ale?

No, it's a compromise. How can we get a valve to do two jobs. Gain and
inversion. The word compromise implies imperfection .. less than perfect..
How well it measures up is for you to decide. I keep an open mind..... I do
not have a favourite inverter.

> > Patrick has probably taken the LTP as far as it can go and tells me I
can
> > have a 6SN7 as one half and an ECC83 as the other and, with a CCJ tail,
will
> > still achieve perfect balance of the two outputs. I cannot dispute
this,
> > he knows what he is doing.
>
> Well its true. Once the CCS and equal RLs is established, it is impossible
> for the increase in current in one RL to be different in value
> to the reduction of current in the other RL.
>
> So balance is assured by RL accuracy, easy in today's world
> of plentiful 1% R values.
>
> One wouldn't dream of using different tubes for each 1/2 of the LTP.
> But you could if you wanted to. Maybe a triode pentode, like the 6U8.
> Then it'd be high gain.

Theoretically, I cannot deny you are right, and, no, of course you wouldn't
use dissimilar triodes for the two halves. Furthermore you have practical
circuits up and running which prove the point. As I said, you have taken the
LTP about as far as it will go.


> > The paraphase ?? Seems even more difficult to balance and keep balanced
and
> > I have rarely used it.
> > By combining inversion and gain in one stage both the paraphase and the
LTP
> > do have the advantage of of reducing the number of stages and suggest a
more
> > stable design.
>
> Not necessarily more stable; see above.

No comment .. it has always looked like a mongrel to me .. It could be
wonderful.. or not.. I do not know..


erted copy of the original
> > > signal. This is more than inverting, or splitting (whatever that
> > > means). The copying part naturally raises the issue of distortion.
> > >
> > > The paraphase might seem awkward for those looking for symmetry.
> > > But you why should copying and inverting be a symmetrical process?
> > > The paraphase uses a valve to produce an inverted copy. Perfect.
> > > Tricky to get the details right so it stays in balance of course,
> > > but so is the LTP.
> > >
> > > All in all, the cathodyne used as a driver ensures balance and low
> > > distortion only within a comparatively narrow corridor of operation,
> > > outside of which it goes badly awry. A paraphase is always slightly
> > > awry but maintains this over a much wider proportion of available HT
> > > and load conditions. A LTP falls somewhere between the two.
> > >
> > > Really you have to consider the amp as a whole.
> >
> > Exactly. It is possible to build fine 10-15W amplifiers with simple
> > circuitry and few components. These are maximum ratings however and with
> > inefficient speakers you could be running close to overload at normal
> > listening volumes. To design a 15W amp and ask it to sound good at 20W
is a
> > non starter.
> > I like 50W as a design target. Times two for stereo. The whole
question of
> > pushing it to the design limits and watching it get out of shape does
not
> > then occur.
>
> At last. Someone else realises a couple of KT88 are always going to do
better
> than a pair of EL34, or '84.


At last ?? I don't build girls amps !!

of what I am ever going to hear is Class A. No crossover
> > distortion and a driver design which can put 90v Pk-pk on the OP valve
grids
> > without really trying.
> > The drawback is that including the PSU, it is big, heavy, and expensive
> > kind regards
> > jim
>
> PS don't have to be huge if class A is all you want, and +380v
> for class A triode/pentode/UL/acoustical is easy.

No, no no.... Never below 450V and octals... 40 years tells me it's a
waste of time... Nowadays we're into rinky dink girly things however. Amps
ought to bite your arse ..
Regards as always
jim

> Patrick Turner.
>

[Patrick Turner]

jim wrote:

> >
> > There is little need for high current bias flow in an LTPdriver/CPI driver
> were
> > it not
> > for the input C of the stage being driven.
> > The higher the idle current, the higher the bandwidth of the driver stage,
> > since the capacitance can be charged and discharged with enough current
> > to maintain the voltage swing at HF.
>
> Hiya, Patrick...
> > In the case of the CPI, a small C across the Rk of the stage will boost
> > the HF response from the anode, which tends to roll off first.
> >
>
> Yes, O.K. ....
>
> > > If you need a
> > > current supply to the OP valve grids you need CFs or a transformer
> hanging
> > > on them. This is AB2 or B. Neither are related to high end audio.
> > > They are about a big noise, efficiency and bangs per buck
> > > Nothing wrong with Menno Van der Veens concertina driving 6 x EL34s. Not
> > > seen the circuit but assume it's cathode biased to retain a fairly high
> grid
> > > impedance for the concertina to work into.
> >
> > But the cathode bias itself causes input grid impedance to be no higher
> > than fixed bias, since grid current occurs only when the peak swing
> > input voltage goes higher than the value of grid bias.
>
> I'm looking at the load the driver has to work into.. Cathode bias does
> not alter the grid impedance per se, as you say, but the grid leak can be
> made higher because of the self regulating effect of the cathode resistor.
> The grid leak is the load.

Indeed, the grid R can be higher value when using cathode bias.

> We have 6 EL34s here with grid leaks in parallel.
> I'm just trying to avoid a coupling cap the size of a 45 gallon oil drum
> and a time constant round about last Thursday !!!!

In my 300 watt amps, the 12 x 6550 each have 0.47 and 120k,
so the load seen by the driver stage is 20k per side, driven by the equivalent
of
6 x 0.47 uF in parallel, or about 3 uF, which is easy to arrange in such large
amps.

To reduce the loading effects on the driver, I use a choke with CT
to supply the DC to the LTP, which uses 2 x EL84 in triode, which is as
effective
as 4 x 6SN7 in parallel. But since you like octals, try a 6V6, or even an EL34.

The 40k a-a load seen by the LTP driver means the thd is negligible.
the low Ro means the miller C of the OPVs have little effect until
well over 100 kHz.

The circuit is at
http://www.turneraudio.com.au/webschema02/300wdriverschema550w335h.gif


>
>
> > > Parallelling 6 x EL34s provides
> > > the output power without asking the concertina to provide a big voltage
> > > swing and keeps the circuit simple and stable.
> >
> > 12 x EL84 would do about the same thing, and be even easier to drive!.
> >
>
> You trying to put one over on me ?? I ain't going to apologise for being a
> hobby boy. Go up one paragraph ... 12 EL34s ??? Driver design, please ??
> You planning on a Dodge or a Chrysler to move this lot ??

See above, its easy to drive multiple output tubes.

>
>
> > > > I think there is an issue of slew rate limiting too, and/oce at HF.
> > > Current-derived cathode feedback is good for
> > > > turning a problem into a crisis.
> > >
> > > There is an imbalance between the cathode and anode outputs but this is
> > > usually outside the limits of audibility. This is not a problem if we
> can't
> > > hear it, but since the concertina has no gain it is usually necessary to
> > > insert another gain stage to get the drive voltages we need for large
> > > valves. A Williamson. We now have four stages and since each stage
> shifts
> > > phase, complete inversion can occur at the extremes of frequency. It is
> > > unstable. We have an oscillator. Add in global NFB and unless the
> > > bandwidth is restricted it may well all get away from you.
> >
> > But to STOP the oscillation, which is normally due to a lousy OPT,
> > and NOT the williamson driver circuit, it is necessary to place a zobel
> network
> > across the main RL of V1, and williamson recommended a 470pF and 4.7k
> > in parallel with the 47 k load of the 1/2 6SN7 which was V1.
>
> Twelve months ago you took my head off about dumping things and introducing
> extra time constants when I suggested restricting the bandwidth by dumping
> things you can't hear across V1R and removing them from the feedback chain
> ...... Whether the OPT is lousy or wonderful, a C across V1R will kill it
> stone dead....... Why would I want to listen to anything over 30kHz ??

Tim De Paravicini once said of Peter Walker's idea of current dumping,
" the worst thing you can do with a current, is dump it..."

I don't recall performing a headectomy on you, over "dumping" anything.

I do believe in critical damping, in FB amps, to make coexistance with our amps
easier without smoky listening rooms a more definite possibility.
I have been trying to get the public to understand what I mean,
and which is described in RDH4, but in a completely obscure manner.

Basically, every amp is a bandpass filter, with a fairly narrow
bandwidth without FB.

When FB is added, BW increases dramatically, and unless
the BW AND OPEN LOOP GAIN at extremes of the BW is reduced
in some way without adding extra phase shift, then oscillation is likely
with some types of loads, and depending on the quality of the OPT.

This message is lost on so many ppl, since they refuse to allow their brains
to realise that 6 things are simultaneously contributing to an outcome
in a very simple audio device, an amplifier.
They want it all to be so simple, but alas, it just cannot be like that.

You don't want to listen to anything over 30 kHz, and probably, like me,
struggle with anything over 14 kHz, but we do want the amp to stay stable
and not misbehave regardless of load,
at frequencies above 30 kHz, and below 100 Hz!


>
> > The Williamson 3 stage driver circuit has more bandwidth than an LTP using
> > the same tubes, since the CPI acts as a buffer against the effects of the
> > miller C of the balanced diver stage causing a pole too early at V1,
> > whose Ro is high because the cathode R isn't bypassed.
>
> So, bypass it. I usually do....or partially, if I need a feedback point.

I do that to, as my circuits indicate, and the Ro of V1 is lowered about
2.5 times, but I still need to use a zobel network to attain
critical damping, and I use a radio tuning cap and a series pot mounted on it to

establish the best values of C&R, with about 0.22uF on the output, no R load,
and with a 5 kHz square wave, where I nearly minimise ringing..

>
>
> > The stepped HF response raises the pole F caused by the driver stage,
> > but reduces its gain by 12 dB at this HF, ie, about 200 kHz, and
> > therefore the amount of applied NFB at 200 kHz is lowered,
> > and stability is increased.
>
> Patrick, we're in the medium waveband here ....... Your designs may have no
> effect on Radio Canberra, but personally, I don't give a **** whether Radio
> Canberra gets my LPs full on or not. It will probably be better than the
> **** they normally play...... Adjust the CR around V1 and 200kHz isn't in
> there.
> What does 200kHz sound like ?? I may have it and not know it !! Tune in ,
> now, I'm going to put a Roy Orbison track on !.......
>
> > Suitable zobel networks are shown also at my website schematics.
>
> You dump the **** you can't hear so it can't come back and bite you ..

I build amplifiers, not oscillators, and if an amp is merrily oscillating
at 200 kHz, there is a power loss, and possibly the amplitude could be high
enough to cause tweeter damage, or cause interference, or saturate the
drive amp, labouring under its presence.

I am a firm believer that only the audio frequencies we want be present in
the output of an audio amp of any kind, regardless of the loading conditions.


>
>
> > The bandwidth of all my amps are tailored to at least 65 kHz,
> > and any sized cap can be connected at the output without
> > oscillations occuring. It is unconditional stability.
> >
> > >
> > >
> > > > So the cathodyne should be restricted to AB1 output stages, and it
> > > > should idle with plenty of current.
> >
> > Indeed, but if CF bufferes after the cathodyne (CPI) are used, direct
> coupled to
> > the OPV grids,
> > no worries, you can go AB2, if the voltage swing is wide enough.
>
> But Ian is going for elegant simplicity. He doesn't want a an extra ECC 81
> stuck in there.. He doesn't want to be into AB2. He doesn't really want to
> be in AB1

He like class A1, and so do I.
Class A1 in most of the gear I build covers 99.9% of what I listen to.

>
>
> > > An ECC83 is unlikely to be a
> > > > good candidate. An '82 or an '81 is better, as might be an '88.
> > > > You have to get enough voltage swing and enough current capability
> > > > for the task.
> > >
> > > All the inverters we use should be restricted to AB1 None of them are
> > > current generators.
> >
> > I think you have that wrong.
>
> I may have. I will rephrase it. None are low impedance current generators.

Engineering terminology regards a "low impedance current generator" as a
voltage source, ie, the Ro of the voltage source is low compared to the load
being driven.
Triode drivers provide this. Pentode drivers ususlly don't,
unless RL is low, and the low RL in comparison with the following
Cap coupled high value Rg, might mean the pentode circuit as a whole is a
voltage source.
When ever blokes say "current generators" it means the Ro of the that current
is very high in comparison with the load, which is true of pentode, or cascoded
triode
outputs, or outputs from triodes with large value Rk, as is the case with ther
CPI
at its anode circuit, if measured singly.

>
> As I see it, a low impedance driver for AB2 cannot include a 100K anode
> load resistor and a coupling cap as in parahase/LTP,
> hanging on the OP valve grid. I may even have this wrong. You will probably
> put me right...

One could build an equivalent circuit of an output grid circuit.
Imagine we have a typically designed cathode bais circuit,
with the voltage at the cathode sittng at say +30v for an EL34.
Imagine having a diode with its anode connected to the tube grid,
and its cathode connected to the +30 at the cathode, but via say a 1 k resistor.

Imagine for a moment, that grid current was impossible, regardless of a high +ve
grid voltage swing.
But with the above arrangement of diode and R, there is no current input to the
grid
from the driver, but when the grid voltage swings +ve enough, ie, about 31 peak
volts,
the diode starts to conduct, and the the impedance seen by the driver
is not just the 100k grid bias R, but it suddenly becomes 1k, the R in series
with
our imaginery diode.
The gain of the driver suddently stalls, because gain = U x RL / ( Ra + RL ),
which is true for all tubes.
Also, the voltage difference across the coupling cap, which is its charge
which is maintained at a constant value during non grid current operation,
suddenly starts to change to a higher value of volts.
So when the driver voltage swings negative, the diode has switched off, but the
coupling
cap carries the extra charge which slowly tries to discharge through Rg.
Sustained over driving of the output tube into AB2 conditions
causes considerable extra direct voltage to be built up
in the coupling cap, resulting in the tube being operated at an effectively
lower bias point.

In effect, our imaginery diode plus certain low value resistor DOES exist,
because as soon as a grid goes +ve,
it attracts and absorbs negatively charged electrons, which are not fussy where
they go, as long as its to a positive something.

When music signals become large enough to cause occasional rail clipping,
then ususally, there is grid current being drawn, and the idle bias condition
shifts,
and thd becomes several times higher. when the levels are increased futher,
the amp sounds like its choking, which is due to paralysis caused by the charge
in the coupling caps, and musiscians with electric geetahs lerve this sound

If a sine wave is used, and grid current is drawn, you have the OPVs
simply acting as switches for a square wave, and biased fairly off,
and the power in the load can be huge compared to the rated undistorted power.

With hi-fi, we never go there.

>
>
> > None are voltage sources, like a cathode follower.
> > Most drivers/inverters have highish Ro, and are limited current sources.
> > Pentodes are current sources, since the voltage is determined by the load,
> > andout put current
> > is more nearly constant, and voltage is anywhere.
> >
> Eh ??

See above.

>
>
> > >
> > > > A LTP makes a good differential amplifier and you can call it a
> > > > splitter or an inverter if you want, but it's a very inelegant
> > > > approach, IMO.
> >
> > I think its very elegant.
> > What makes it inelegant?
> >
> > Is is the hobnail boots and hairy legs worn with the pink frock?
>
> No, it's all the bits and pieces dangling off the side of it which makes it
> look a complete ****. !!!!
> How can you trust anything which looks like that !!! Christ !! It's
> American !!! Where's me concertina ??

But even an american sheila with nice tits a dangle dingling about
is OK, even with hobnail boots, and nearly wearing her pink frock!

>
>
> > > Only the concertina could be classed as elegant...with a direct coupled
> > > voltage amp up front. The LTP and paraphase are a compromise of a gain
> and
> > > inverter stage combined.
> >
> > Compromise? Have you been drinking too much cheap english ale?
>
> No, it's a compromise. How can we get a valve to do two jobs. Gain and
> inversion. The word compromise implies imperfection .. less than perfect..
> How well it measures up is for you to decide. I keep an open mind..... I do
> not have a favourite inverter.

I don't either, but what I use depends on the output stage.

>
>
> > > Patrick has probably taken the LTP as far as it can go and tells me I
> can
> > > have a 6SN7 as one half and an ECC83 as the other and, with a CCJ tail,
> will
> > > still achieve perfect balance of the two outputs. I cannot dispute
> this,
> > > he knows what he is doing.
> >
> > Well its true. Once the CCS and equal RLs is established, it is impossible
> > for the increase in current in one RL to be different in value
> > to the reduction of current in the other RL.
> >
> > So balance is assured by RL accuracy, easy in today's world
> > of plentiful 1% R values.
> >
> > One wouldn't dream of using different tubes for each 1/2 of the LTP.
> > But you could if you wanted to. Maybe a triode pentode, like the 6U8.
> > Then it'd be high gain.
>
> Theoretically, I cannot deny you are right, and, no, of course you wouldn't
> use dissimilar triodes for the two halves. Furthermore you have practical
> circuits up and running which prove the point. As I said, you have taken the
> LTP about as far as it will go.

Nope, I am working on one with current mirrors, using
PNP transitors. I'd like to use PNP toobes,
but toobes are all NPN formatted.

>
> > > The paraphase ?? Seems even more difficult to balance and keep balanced
> and
> > > I have rarely used it.
> > > By combining inversion and gain in one stage both the paraphase and the
> LTP
> > > do have the advantage of of reducing the number of stages and suggest a
> more
> > > stable design.
> >
> > Not necessarily more stable; see above.
>
> No comment .. it has always looked like a mongrel to me .. It could be
> wonderful.. or not.. I do not know..
>
> erted copy of the original
> > > > signal. This is more than inverting, or splitting (whatever that
> > > > means). The copying part naturally raises the issue of distortion.
> > > >
> > > > The paraphase might seem awkward for those looking for symmetry.
> > > > But you why should copying and inverting be a symmetrical process?
> > > > The paraphase uses a valve to produce an inverted copy. Perfect.
> > > > Tricky to get the details right so it stays in balance of course,
> > > > but so is the LTP.
> > > >
> > > > All in all, the cathodyne used as a driver ensures balance and low
> > > > distortion only within a comparatively narrow corridor of operation,
> > > > outside of which it goes badly awry. A paraphase is always slightly
> > > > awry but maintains this over a much wider proportion of available HT
> > > > and load conditions. A LTP falls somewhere between the two.
> > > >
> > > > Really you have to consider the amp as a whole.
> > >
> > > Exactly. It is possible to build fine 10-15W amplifiers with simple
> > > circuitry and few components. These are maximum ratings however and with
> > > inefficient speakers you could be running close to overload at normal
> > > listening volumes. To design a 15W amp and ask it to sound good at 20W
> is a
> > > non starter.
> > > I like 50W as a design target. Times two for stereo. The whole
> question of
> > > pushing it to the design limits and watching it get out of shape does
> not
> > > then occur.
> >
> > At last. Someone else realises a couple of KT88 are always going to do
> better
> > than a pair of EL34, or '84.
>
> At last ?? I don't build girls amps !!
>
> of what I am ever going to hear is Class A. No crossover
> > > distortion and a driver design which can put 90v Pk-pk on the OP valve
> grids
> > > without really trying.
> > > The drawback is that including the PSU, it is big, heavy, and expensive
> > > kind regards
> > > jim
> >
> > PS don't have to be huge if class A is all you want, and +380v
> > for class A triode/pentode/UL/acoustical is easy.
>
> No, no no.... Never below 450V and octals... 40 years tells me it's a
> waste of time... Nowadays we're into rinky dink girly things however. Amps
> ought to bite your arse ..

I am choosy about how I like my arse being bit, and she has to be good looking
before I'll let her.

But for class A, 450v is sometimes too much, and 375v is plenty for
low THD and class A, with octals in tetrode, pentode, UL, acoustical.
But for triode connected tubes, 450v is OK.

Patrick Turner.

>
> Regards as always
> jim

[Patrick Turner]

jim wrote:

> Hiya, Patrick,
> We've all been on your website. Thanks for contributing. I love you,
> desperately !!!! ... You have no faith in women...... I may be prepared
> to become homosexual if it furthers my knowledge of a KT88.... Come back to
> me... I could be yours !!

Oh Dahling, I thought you'd never care for me,
but youse is so ugly, and I is so fussay......

>
> Part 2 of this header is irrelevant. I cannot-----
> HEAR----- a phase inverter. If you can, it needs a rethink.!!!

I never said I couldn't hear it, or that I could, or could have heard it,
if the amp wasn't switched on....

> We may be
> able to measure things way past the point at which it is supposed to
> operate, but normally, we cannot hear an inverter.

Well, every tube, or amplifying device inverts the phase.
the signal goes in negative, and comes out positive, so what's the fuss?
Does an inverted signal sound like a man talking while he stands on his head?
Maybe.

>
> Your comments on driving multiple pairs are relevant, and accurate, but I
> think most of
> us, here, are listening to music at home, and 2 x50W is about as far as we
> need to go.
> 2 x 50W is fairly easy. .... 6L6GC/5881/KT88/6550 .... 450v HT/B+...
> Somebody out there is going to use 48 CRTs or GZ34s, in parallel because
> the load lines look good, but I just like to think that valves designed for
> audio are the best bet.

Indeed I agree.

>
> Easy to drive too... LTP or concertina/Williamson....
> I have been into LTP/concertina and ocassionally paraphase. I still like
> the
> symmetry, coupling simplicity, and low imp drive of a CT transformer, but
> it will probably not happen in my lifetime.

It looks simple on paper, but one needs iron, wire, and patience, a lathe, wax,
bobbins,
and no swearing when you break the fine wire when turning it up.
Most ppl use resistors, which are cheaper, easier,
but the thd is 3 times that of the long way around the iron.

>
> There is an answer to the ultimate driver/power amp. Most of us have
> realized that it is silicon. Fortunately, none of is going to admit it......
> I will be with you, all the way, ..we are all behind you ..to plumb the
> depths and hit the highs of an ECC83, no matter what Texas Instruments
> throws at us......
> regards
> jim.

Oh Jimmy, I was nearly going to fall in love, but then you said that;
about SS, an I'm allergic to some sillycon, and now I've gone all bashful
and shy.

Patrick Turner.

[Ross Matheson]

"jim" > wrote:

: My main point, however, is that you are blatantly advertizing Turnip Audio
: again !!!
: I have reported this to Andre, Arny and Stewart
: They will be contacting you shortly
: regards
: jim

I think dat Turnip Audio has already been hexed into jive, man:=))
Dig dat design philsophy! Slap mah fro! :=})
http://tinyurl.com/itxl

But the previous discussion has been enlightening ...
Cheers, Pat!

[Fred Nachbaur]

Ross Matheson wrote:
> "jim" > wrote:
>
> : My main point, however, is that you are blatantly advertizing Turnip Audio
> : again !!!
> : I have reported this to Andre, Arny and Stewart
> : They will be contacting you shortly
> : regards
> : jim
>
> I think dat Turnip Audio has already been hexed into jive, man:=))
> Dig dat design philsophy! Slap mah fro! :=})
> http://tinyurl.com/itxl
>
> But the previous discussion has been enlightening ...
> Cheers, Pat!

LOL! Where do you find this stuff ???

I especially liked "Vacuum Tube Amplifiers and Loudrapa' Systems" Hehehe..

Even the links to other pages are "dialectisized." Clever bit of coding,
this Dialectizer!

Cheers,
Fred
--
+--------------------------------------------+
| Music: http://www3.telus.net/dogstarmusic/ |
| Projects, Vacuum Tubes & other stuff: |
| http://www.dogstar.dantimax.dk |
+--------------------------------------------+

[Ross Matheson]

"jim" > feverishly worked up new ad copy:-
in >,
on Stardate Mon, 4 Aug 2003 14:30:05 -0700,
:
: "Patrick Turner" > wrote in message
: ...
: >
: >
: > jim wrote:
: >
: > > Hiya, Patrick,
: > > We've all been on your website. Thanks for contributing. I love you,
: > > desperately !!!! ... You have no faith in women...... I may be
: > > prepared to become homosexual if it furthers my knowledge of a KT88....
: > > Come back to me... I could be yours !!
: >
: > Oh Dahling, I thought you'd never care for me,
: > but youse is so ugly, and I is so fussay......
:
: **** off, you old ****.!!!!!!
: (sorry, Uselessnet-- I won't swear again)
: I got a picture of you standing behind a 6550 barby looking as if the Queen
: and Prince Philip were going to turn up at any minute........... Don't give
: me that 'choosy' line... I look like you.. You take anything on
: offer...............!!! It's got a pulse.... you go for it !!!!
:
: > >
: > > Part 2 of this header is irrelevant. I cannot-----
: > > HEAR----- a phase inverter. If you can, it needs a rethink.!!!

<{([SNIP])}>

We clearly need a marketing re-think here!!! Let's do some X-RESEARCH! :)
Those funky dialects are just tame compared to entering a favourite URL at

[[[[[ R 18 WARNING!!! ]]]]]

http://www.pornolize.com/

<{([ HIT Refresh! ]})>

[Hit refresh and it changes again ... might be even for the better!]

: > > I will be with you, all the way, ..we are all behind you ..to plumb the
: > > depths and hit the highs of an ECC83, no matter what Texas Instruments
: > > throws at us......
: > > regards
: > > jim.
: >
: > Oh Jimmy, I was nearly going to fall in love, but then you said that;
: > about SS, an I'm allergic to some sillycon, and now I've gone all bashful
: > and shy.
:
: Bugger off, Patty !! May I call you Patty, darling ?? Mmmmmmmmmm ..
:
: > Patrick Turner.

Forget Venus! Mars is heading straight for Canberra, right this minute!!!
http://tinyurl.com/iyvl

RdM::==})

[jim]

Hiya, Patrick,
We've all been on your website. Thanks for contributing. I love you,
desperately !!!! ... You have no faith in women...... I may be prepared
to become homosexual if it furthers my knowledge of a KT88.... Come back to
me... I could be yours !!
Part 2 of this header is irrelevant. I cannot-----
HEAR----- a phase inverter. If you can, it needs a rethink.!!! We may be
able to measure things way past the point at which it is supposed to
operate, but normally, we cannot hear an inverter.
Your comments on driving multiple pairs are relevant, and accurate, but I
think most of
us, here, are listening to music at home, and 2 x50W is about as far as we
need to go.
2 x 50W is fairly easy. .... 6L6GC/5881/KT88/6550 .... 450v HT/B+...
Somebody out there is going to use 48 CRTs or GZ34s, in parallel because
the load lines look good, but I just like to think that valves designed for
audio are the best bet.
Easy to drive too... LTP or concertina/Williamson....
I have been into LTP/concertina and ocassionally paraphase. I still like
the
symmetry, coupling simplicity, and low imp drive of a CT transformer, but
it will probably not happen in my lifetime.
There is an answer to the ultimate driver/power amp. Most of us have
realized that it is silicon. Fortunately, none of is going to admit it......
I will be with you, all the way, ..we are all behind you ..to plumb the
depths and hit the highs of an ECC83, no matter what Texas Instruments
throws at us......
regards
jim.

[jim]

"Patrick Turner" > wrote in message
...
>
>
> jim wrote:
>
> > Hiya, Patrick,
> > We've all been on your website. Thanks for contributing. I love you,
> > desperately !!!! ... You have no faith in women...... I may be
prepared
> > to become homosexual if it furthers my knowledge of a KT88.... Come back
to
> > me... I could be yours !!
>
> Oh Dahling, I thought you'd never care for me,
> but youse is so ugly, and I is so fussay......

**** off, you old ****.!!!!!!
(sorry, Uselessnet-- I won't swear again)
I got a picture of you standing behind a 6550 barby looking as if the Queen
and Prince Philip were going to turn up at any minute........... Don't give
me that 'choosy' line... I look like you.. You take anything on
offer...............!!! It's got a pulse.... you go for it !!!!

> >
> > Part 2 of this header is irrelevant. I cannot-----
> > HEAR----- a phase inverter. If you can, it needs a rethink.!!!
>
> I never said I couldn't hear it, or that I could, or could have heard it,
> if the amp wasn't switched on....

I sometimes hear it when the amp is off. I often hear it after a few cans of
something or other .......If the amp is on, or not, it could be better, or
not, or on, or not, I don't really know.......

> > We may be
> > able to measure things way past the point at which it is supposed to
> > operate, but normally, we cannot hear an inverter.

Now, let's be serious...

> Well, every tube, or amplifying device inverts the phase.
> the signal goes in negative, and comes out positive, so what's the fuss?
> Does an inverted signal sound like a man talking while he stands on his
head?
> Maybe.

Try as they might, Americans and Belgians cannot speak English. Australians
only get near.. Why am I trying to reduce the THD when people can't even
talk properly ?? Why are we talking about distortion with people with
speech defects ??

> > Your comments on driving multiple pairs are relevant, and accurate, but
I
> > think most of
> > us, here, are listening to music at home, and 2 x50W is about as far as
we
> > need to go.
> > 2 x 50W is fairly easy. .... 6L6GC/5881/KT88/6550 .... 450v HT/B+...
> > Somebody out there is going to use 48 CRTs or GZ34s, in parallel
because
> > the load lines look good, but I just like to think that valves designed
for
> > audio are the best bet.
>
> Indeed I agree.

What !!!!!!!!!! You agree ?? With the Hobby Boy ?????

> >
> > Easy to drive too... LTP or concertina/Williamson....
> > I have been into LTP/concertina and ocassionally paraphase. I still
like
> > the
> > symmetry, coupling simplicity, and low imp drive of a CT transformer,
but
> > it will probably not happen in my lifetime.
>
> It looks simple on paper, but one needs iron, wire, and patience, a lathe,
wax,
> bobbins,
> and no swearing when you break the fine wire when turning it up.
> Most ppl use resistors, which are cheaper, easier,
> but the thd is 3 times that of the long way around the iron.

I have not gone very far down the split phase driver transfomer route. I
actually haven't gone anywhere at all. It is a dream and you may well
invent it. I have to go to work every day.


> > There is an answer to the ultimate driver/power amp. Most of us have
> > realized that it is silicon. Fortunately, none of is going to admit
it......
> > I will be with you, all the way, ..we are all behind you ..to plumb the
> > depths and hit the highs of an ECC83, no matter what Texas Instruments
> > throws at us......
> > regards
> > jim.
>
> Oh Jimmy, I was nearly going to fall in love, but then you said that;
> about SS, an I'm allergic to some sillycon, and now I've gone all bashful
> and shy.

Bugger off, Patty !! May I call you Patty, darling ?? Mmmmmmmmmm ..

> Patrick Turner.
>
>

[jim]

Patrick,
I just know you are going to tell me that my dog is ****ing up the wrong
tree..... My picture is a paralelled 6SN7 -- about 15mA...you can start a
truck with it ..... and a 1:2 CT split phase driver ..
Anode or cathode loaded. ?? You choose...... But don't come back and poke
me in the eye and tell me I have it wrong, . Unless you know you are
right.............
..regards
jim

[Patrick Turner]

jim wrote:

> Patrick,
> I just know you are going to tell me that my dog is ****ing up the wrong
> tree..... My picture is a paralelled 6SN7 -- about 15mA...you can start a
> truck with it ..... and a 1:2 CT split phase driver ..
> Anode or cathode loaded. ?? You choose...... But don't come back and poke
> me in the eye and tell me I have it wrong, . Unless you know you are
> right.............
> .regards
> jim

Last time I poked someone in the eye, the big bully bashed me up,
and mussed up my nice new hairdo, and made me cry.

If the tranny has wide bandwidth, and a flat response which gets more unlikely
as the step up ratio increases, then the 6SN7 is a nice driver for this,
or even better, a 6V6 in triode.
The use of a CF to drive the input to the primary
means the bandwidth and response is more likely to be better,
and thd of the iron will be lower, since that also reduces with
the lower drive impedance. But the tranny does have parasitic inductances, and
shunt C,
and leakage L, and hence resonances, and use of FB is limited, since extra
phase shift
occurs.
A second input 6SN7 needs to be used for gain.

But you like them and they look nice on the chassis....


But in triode amps with no FB, IST inverters are fine.
Lundahl make a nice IST 1:2 for an inverter.

Patrick Turner.

[Patrick Turner]

jim wrote:

> "Patrick Turner" > wrote in message
>
> > Last time I poked someone in the eye, the big bully bashed me up,
> > and mussed up my nice new hairdo, and made me cry.
>
> Hiya,
> So, my lad did find you, then !!

It was so late at night, and dark in that alley way,
And I don't know what your whippersnapper lad was looking for, but he wasn't
going to take liberties with ME!

>
>
> > If the tranny has wide bandwidth, and a flat response which gets more
> unlikely
> > as the step up ratio increases, then the 6SN7 is a nice driver for this,
> > or even better, a 6V6 in triode.
> > The use of a CF to drive the input to the primary
> > means the bandwidth and response is more likely to be better,
> > and thd of the iron will be lower, since that also reduces with
> > the lower drive impedance. But the tranny does have parasitic
> inductances, and
> > shunt C,
> > and leakage L, and hence resonances, and use of FB is limited, since extra
> > phase shift
> > occurs.
> > A second input 6SN7 needs to be used for gain.
> >
> > But you like them and they look nice on the chassis....
>
> We've had this conversation before. I only keep dragging it up again in the
> hope that you will think about it rather than keep giving me those negative
> waves. We are agreed that dedicated audio valves are probably the best
> things to use for audio. Since the KT88/90/6550 are available and just a
> pair of them will frighten most people, possibly we have the nucleus of the
> ultimate valve amp, here. ?? Multiple pairs of EL84s are not a step forward
> because the impedances go in opposite directions. Paralleling them might
> result in a simpler OPT, but the driver design becomes more complex.

Nah, paralleling EL84 makes the driver desigm MORE simple,
since all one needs is a gain tube, maybe just a CPI, which only has to
make about 6vrms to each side of the OPV grids, and
the input tube only has to make this maximum lowish voltage.
Baird use a sixpack of EL84, good for 60 watts,
which is too much, and 36 watts would be better,
or use 8 tubes for 50 watts.
4 x EL84 is the equivalent of a tube with Gm = 44 mA/v, Ra = 10k,
and U of 440, in pentode mode.
This means that with acoustical connection, the use of 10
dB of OPT CFB will work very effectively,
and give an output stage giving 50 watts, with triode like character,
and sound, and still be no worse to drive than KT88 in plain UL.
The paralleled tubes have a "higher figure of merit"
than a single KT88, for about the same figure of maximum plate dissipation.
Costs are similar. OPT for 8 x EL84 would be the same size as for 2 X KT88.

>
> So, let's go down the road a bit... The KT88/90/6550 as triodes or
> pentodes/beam tets are pretty good. Turnipaudio.com will have no problem
> matching up to any speaker you choose with a full range of, hand crafted,
> individually signed, limited edition, personally comissioned, OPT,s (now
> available in gold, silver or brushed walnut finish) All OPTs now come with a
> signed photo of the main man standing behind his 6550 barbecue, and are
> fully approved by Andre Jute, Neddy Seagoon and Kylie Minogue.... (
> Patrick's big one is the best I have ever come across..

That's what Kylie and Nicole told me, but they say that to all the fellas.....

> quote.. Rangoon Auto
> Trader 29.07.2003)
> The back end is fairly easy, then.
> Up the front, it is just a matter of getting a bit of voltage to excite the
> inverter. Then we all go off in opposite directions with lots of Rs and Cs
> and neg supplies and silicon for a gain of bugger all, ... just to get two
> opposite phase drives..
> The back and front end are easy. Put your head around the middle bit.
> Actually, throw the LTP over the hedge and get a blank sheet of paper.
> My kindest regards, big boy......
> jim

We have such tall hedges here in Oz, its difficult to throw things over...

I will take on board your comments re the foregoing matters, and forward them
to the specialist in our organisation.

Patrick Turner.

[jim]

"Patrick Turner" > wrote in message
...
>
>
> jim wrote:
>
> > Patrick,
> > I just know you are going to tell me that my dog is ****ing up the wrong
> > tree..... My picture is a paralelled 6SN7 -- about 15mA...you can start
a
> > truck with it ..... and a 1:2 CT split phase driver ..
> > Anode or cathode loaded. ?? You choose...... But don't come back and
poke
> > me in the eye and tell me I have it wrong, . Unless you know you are
> > right.............
> > .regards
> > jim
>
> Last time I poked someone in the eye, the big bully bashed me up,
> and mussed up my nice new hairdo, and made me cry.

Hiya,
So, my lad did find you, then !!

> If the tranny has wide bandwidth, and a flat response which gets more
unlikely
> as the step up ratio increases, then the 6SN7 is a nice driver for this,
> or even better, a 6V6 in triode.
> The use of a CF to drive the input to the primary
> means the bandwidth and response is more likely to be better,
> and thd of the iron will be lower, since that also reduces with
> the lower drive impedance. But the tranny does have parasitic
inductances, and
> shunt C,
> and leakage L, and hence resonances, and use of FB is limited, since extra
> phase shift
> occurs.
> A second input 6SN7 needs to be used for gain.
>
> But you like them and they look nice on the chassis....

We've had this conversation before. I only keep dragging it up again in the
hope that you will think about it rather than keep giving me those negative
waves. We are agreed that dedicated audio valves are probably the best
things to use for audio. Since the KT88/90/6550 are available and just a
pair of them will frighten most people, possibly we have the nucleus of the
ultimate valve amp, here. ?? Multiple pairs of EL84s are not a step forward
because the impedances go in opposite directions. Paralleling them might
result in a simpler OPT, but the driver design becomes more complex.
So, let's go down the road a bit... The KT88/90/6550 as triodes or
pentodes/beam tets are pretty good. Turnipaudio.com will have no problem
matching up to any speaker you choose with a full range of, hand crafted,
individually signed, limited edition, personally comissioned, OPT,s (now
available in gold, silver or brushed walnut finish) All OPTs now come with a
signed photo of the main man standing behind his 6550 barbecue, and are
fully approved by Andre Jute, Neddy Seagoon and Kylie Minogue.... (
Patrick's big one is the best I have ever come across.. quote.. Rangoon Auto
Trader 29.07.2003)
The back end is fairly easy, then.
Up the front, it is just a matter of getting a bit of voltage to excite the
inverter. Then we all go off in opposite directions with lots of Rs and Cs
and neg supplies and silicon for a gain of bugger all, ... just to get two
opposite phase drives..
The back and front end are easy. Put your head around the middle bit.
Actually, throw the LTP over the hedge and get a blank sheet of paper.
My kindest regards, big boy......
jim



> But in triode amps with no FB, IST inverters are fine.
> Lundahl make a nice IST 1:2 for an inverter.
>
> Patrick Turner.
>
>
>

[jim]

"Patrick Turner" > wrote in message
...
>
>
> jim wrote:
>
> > "Patrick Turner" > wrote in message
> >
> > > Last time I poked someone in the eye, the big bully bashed me up,
> > > and mussed up my nice new hairdo, and made me cry.
> >
> > Hiya,
> > So, my lad did find you, then !!
>
> It was so late at night, and dark in that alley way,
> And I don't know what your whippersnapper lad was looking for, but he
wasn't
> going to take liberties with ME!

I told him to be gentle..... This guy is going to give us the ultimate
inverter, I said !!

> > > If the tranny has wide bandwidth, and a flat response which gets more
> > unlikely
> > > as the step up ratio increases, then the 6SN7 is a nice driver for
this,
> > > or even better, a 6V6 in triode.
> > > The use of a CF to drive the input to the primary
> > > means the bandwidth and response is more likely to be better,
> > > and thd of the iron will be lower, since that also reduces with
> > > the lower drive impedance. But the tranny does have parasitic
> > inductances, and
> > > shunt C,
> > > and leakage L, and hence resonances, and use of FB is limited, since
extra
> > > phase shift
> > > occurs.
> > > A second input 6SN7 needs to be used for gain.
> > >
> > > But you like them and they look nice on the chassis....
> >
> > We've had this conversation before. I only keep dragging it up again in
the
> > hope that you will think about it rather than keep giving me those
negative
> > waves. We are agreed that dedicated audio valves are probably the best
> > things to use for audio. Since the KT88/90/6550 are available and just
a
> > pair of them will frighten most people, possibly we have the nucleus of
the
> > ultimate valve amp, here. ?? Multiple pairs of EL84s are not a step
forward
> > because the impedances go in opposite directions. Paralleling them might
> > result in a simpler OPT, but the driver design becomes more complex.
>
> Nah, paralleling EL84 makes the driver desigm MORE simple,
> since all one needs is a gain tube, maybe just a CPI, which only has to
> make about 6vrms to each side of the OPV grids, and
> the input tube only has to make this maximum lowish voltage.

Thought we agreed that KT88/90/6550 was a good idea ?? What's with this
nine pin stuff ??

> Baird use a sixpack of EL84, good for 60 watts,
> which is too much, and 36 watts would be better,
> or use 8 tubes for 50 watts.


Times two for a stereo pair equals 16 x EL84s...... It's no more difficult
to drive than 4 x 6550s.. OK ... but why don't we just go for the 6550s and
save all those holes, sockets and heater amps ??
Get simple, just use the 6550s !!


> 4 x EL84 is the equivalent of a tube with Gm = 44 mA/v, Ra = 10k,
> and U of 440, in pentode mode.
> This means that with acoustical connection, the use of 10
> dB of OPT CFB will work very effectively,
> and give an output stage giving 50 watts, with triode like character,
> and sound, and still be no worse to drive than KT88 in plain UL.
> The paralleled tubes have a "higher figure of merit"

????? What's that mean ??

> than a single KT88, for about the same figure of maximum plate
dissipation.
> Costs are similar. OPT for 8 x EL84 would be the same size as for 2 X
KT88.

We're just going to fall out, here... As far as I am concerned, the fact
that I cannot connect my Shure M75 directly to the speakers is a
disadvantage and means that I have to put something between them. What goes
in the middle should be as simple and uncomplicated as possible. Everything
I put between them will degrade, corrupt or distort what my Shure is trying
to say.
Why would I want to drive 12/16 nine pins when 4 octals will do the same job
??
Every cap and resistor is going to do something. How few can I get away
with ??? Ian is trying to go the same way. Elegant simplicity are his
words. We just have different output levels.

> > So, let's go down the road a bit... The KT88/90/6550 as triodes or
> > pentodes/beam tets are pretty good. Turnipaudio.com will have no
problem
> > matching up to any speaker you choose with a full range of, hand
crafted,
> > individually signed, limited edition, personally comissioned, OPT,s
(now
> > available in gold, silver or brushed walnut finish) All OPTs now come
with a
> > signed photo of the main man standing behind his 6550 barbecue, and are
> > fully approved by Andre Jute, Neddy Seagoon and Kylie Minogue.... (
> > Patrick's big one is the best I have ever come across..
>
> That's what Kylie and Nicole told me, but they say that to all the
fellas.....


Nicole ??? You haven't filled us in on Nicole ?? Hey !!


> > quote.. Rangoon Auto
> > Trader 29.07.2003)
> > The back end is fairly easy, then.
> > Up the front, it is just a matter of getting a bit of voltage to excite
the
> > inverter. Then we all go off in opposite directions with lots of Rs
and Cs
> > and neg supplies and silicon for a gain of bugger all, ... just to get
two
> > opposite phase drives..
> > The back and front end are easy. Put your head around the middle bit.
> > Actually, throw the LTP over the hedge and get a blank sheet of paper.
> > My kindest regards, big boy......
> > jim
>
> We have such tall hedges here in Oz, its difficult to throw things over...

I have a hedge trimmer design which uses 120 EL84s in parallel I could let
you have ......

> I will take on board your comments re the foregoing matters, and forward
them
> to the specialist in our organisation.

> Patrick Turner.
>
>