Could be my ISP has fixed a problem but I think I have missed a day's
postings.

Anyway, in case I have, and in case ppl are unaware, I repeat that
I posted 5 line stage schematics at a.b.s.e 2 days ago.
Maybe they are gone but they are at
http://www.usenet-replayer.com/cgi/content/archive

Do a search there on a.b.s.e, and save as usual.

Patrick Turner.

On Tue, 26 Apr 2005 12:45:47 GMT, Patrick Turner
wrote:

Could be my ISP has fixed a problem but I think I have missed a day's
postings.

Anyway, in case I have, and in case ppl are unaware, I repeat that
I posted 5 line stage schematics at a.b.s.e 2 days ago.
Maybe they are gone but they are at
http://www.usenet-replayer.com/cgi/content/archive

Do a search there on a.b.s.e, and save as usual.

Patrick Turner.

A gain of 4x from just three valves. And at a distortion level of only
..02% at 1V out (from a rail of 300V).

Well, I'm impressed - state of the art indeed.

Even on a tubes news group, this needs some explaining.

d

Pearce Consulting
http://www.pearce.uk.com

Don Pearce wrote:

On Tue, 26 Apr 2005 12:45:47 GMT, Patrick Turner
wrote:

Could be my ISP has fixed a problem but I think I have missed a day's
postings.

Anyway, in case I have, and in case ppl are unaware, I repeat that
I posted 5 line stage schematics at a.b.s.e 2 days ago.
Maybe they are gone but they are at
http://www.usenet-replayer.com/cgi/content/archive

Do a search there on a.b.s.e, and save as usual.

Patrick Turner.

A gain of 4x from just three valves. And at a distortion level of only
.02% at 1V out (from a rail of 300V).

Well, I'm impressed - state of the art indeed.

Even on a tubes news group, this needs some explaining.

This is the typical performance of SET preamps at a volt output
with 6SN7, and using mild NFB.

But in practice 1v is what will produce full clipping power in most
amps of say 50 watts so average levels are where the preamp
has only to make 0.2 v output.
thd will be far lower than 0.02%.
Almost unmeasurable thd is gained using all balanced circuitry
at a volt of output, since nearly all the abobe mentioned thd of 2H
is cancelled without any NFB application.


Just do the work of building it, and you will familiarise yourself
fully with how little distortion triodes circuits generate at low signal
levels.
Noise is the only other concern, but a good 6SN7
has 2 uV at its input of grid input noise with
G1 grounded. Add 2 more uV for the gain control series R,
and if the gain is 4, then there under 10 uV of noise at the output,
and if the signal is 0.2v, the SNR = -106dB.
In practice -90 dB is routine.
Using higher gm triodes such as 6H30 and 6DJ8/6922
will give lower equivalent input noise resistance,
and using a 20k gain pot will reduce pot noise,
so this might make noise figures 6 dB better.

Patrick Turner.






d

Pearce Consulting
http://www.pearce.uk.com

"Patrick Turner" wrote in message
...
Could be my ISP has fixed a problem but I think I have missed a day's
postings.

Anyway, in case I have, and in case ppl are unaware, I repeat that
I posted 5 line stage schematics at a.b.s.e 2 days ago.
Maybe they are gone but they are at
http://www.usenet-replayer.com/cgi/content/archive

Do a search there on a.b.s.e, and save as usual.

Patrick Turner.


Patrick, I could not find these schematics. Could you
e-mail them to me, please.

Iain

On Tue, 26 Apr 2005 14:15:02 GMT, Patrick Turner
wrote:



Don Pearce wrote:

On Tue, 26 Apr 2005 12:45:47 GMT, Patrick Turner
wrote:

Could be my ISP has fixed a problem but I think I have missed a day's
postings.

Anyway, in case I have, and in case ppl are unaware, I repeat that
I posted 5 line stage schematics at a.b.s.e 2 days ago.
Maybe they are gone but they are at
http://www.usenet-replayer.com/cgi/content/archive

Do a search there on a.b.s.e, and save as usual.

Patrick Turner.

A gain of 4x from just three valves. And at a distortion level of only
.02% at 1V out (from a rail of 300V).

Well, I'm impressed - state of the art indeed.

Even on a tubes news group, this needs some explaining.

This is the typical performance of SET preamps at a volt output
with 6SN7, and using mild NFB.

But in practice 1v is what will produce full clipping power in most
amps of say 50 watts so average levels are where the preamp
has only to make 0.2 v output.
thd will be far lower than 0.02%.
Almost unmeasurable thd is gained using all balanced circuitry
at a volt of output, since nearly all the abobe mentioned thd of 2H
is cancelled without any NFB application.


Just do the work of building it, and you will familiarise yourself
fully with how little distortion triodes circuits generate at low signal
levels.
Noise is the only other concern, but a good 6SN7
has 2 uV at its input of grid input noise with
G1 grounded. Add 2 more uV for the gain control series R,
and if the gain is 4, then there under 10 uV of noise at the output,
and if the signal is 0.2v, the SNR = -106dB.
In practice -90 dB is routine.
Using higher gm triodes such as 6H30 and 6DJ8/6922
will give lower equivalent input noise resistance,
and using a 20k gain pot will reduce pot noise,
so this might make noise figures 6 dB better.

Patrick Turner.


There are two kinds of device that require preamps - microphones and
phono stages. Everything else should feed directly into a power amp
via nothing more than a volume control.

These circuits are suited to neither of those - so what are they for?

d

Pearce Consulting
http://www.pearce.uk.com

"Don Pearce" wrote in message
...
On Tue, 26 Apr 2005 12:45:47 GMT, Patrick Turner
wrote:

Could be my ISP has fixed a problem but I think I have missed a day's
postings.

Anyway, in case I have, and in case ppl are unaware, I repeat that
I posted 5 line stage schematics at a.b.s.e 2 days ago.
Maybe they are gone but they are at
http://www.usenet-replayer.com/cgi/content/archive

Do a search there on a.b.s.e, and save as usual.

Patrick Turner.

A gain of 4x from just three valves. And at a distortion level of only
.02% at 1V out (from a rail of 300V).

Well, I'm impressed - state of the art indeed.

Even on a tubes news group, this needs some explaining.

d

With the high levels from CD players, high gain pre-amps are
not really needed. People on this group were interested to know
about a low gain stage. x4 is useful.

Iain

Don Pearce wrote:

On Tue, 26 Apr 2005 14:15:02 GMT, Patrick Turner
wrote:



Don Pearce wrote:

On Tue, 26 Apr 2005 12:45:47 GMT, Patrick Turner
wrote:

Could be my ISP has fixed a problem but I think I have missed a day's
postings.

Anyway, in case I have, and in case ppl are unaware, I repeat that
I posted 5 line stage schematics at a.b.s.e 2 days ago.
Maybe they are gone but they are at
http://www.usenet-replayer.com/cgi/content/archive

Do a search there on a.b.s.e, and save as usual.

Patrick Turner.

A gain of 4x from just three valves. And at a distortion level of only
.02% at 1V out (from a rail of 300V).

Well, I'm impressed - state of the art indeed.

Even on a tubes news group, this needs some explaining.

This is the typical performance of SET preamps at a volt output
with 6SN7, and using mild NFB.

But in practice 1v is what will produce full clipping power in most
amps of say 50 watts so average levels are where the preamp
has only to make 0.2 v output.
thd will be far lower than 0.02%.
Almost unmeasurable thd is gained using all balanced circuitry
at a volt of output, since nearly all the abobe mentioned thd of 2H
is cancelled without any NFB application.


Just do the work of building it, and you will familiarise yourself
fully with how little distortion triodes circuits generate at low signal
levels.
Noise is the only other concern, but a good 6SN7
has 2 uV at its input of grid input noise with
G1 grounded. Add 2 more uV for the gain control series R,
and if the gain is 4, then there under 10 uV of noise at the output,
and if the signal is 0.2v, the SNR = -106dB.
In practice -90 dB is routine.
Using higher gm triodes such as 6H30 and 6DJ8/6922
will give lower equivalent input noise resistance,
and using a 20k gain pot will reduce pot noise,
so this might make noise figures 6 dB better.

Patrick Turner.


There are two kinds of device that require preamps - microphones and
phono stages. Everything else should feed directly into a power amp
via nothing more than a volume control.

These circuits are suited to neither of those - so what are they for?

They are for the same use that so many other makers provide preamps for.
You could consider challenging your own belief systems sometimes.

Patrick Turner.



d

Pearce Consulting
http://www.pearce.uk.com

On Tue, 26 Apr 2005 16:52:55 +0200, François Yves Le Gal
wrote:

On Tue, 26 Apr 2005 14:26:49 GMT, (Don Pearce) wrote:

Everything else should feed directly into a power amp
via nothing more than a volume control.

BS. Power amps love low Z drive : at least a buffer stage is required after
the volume control, excet if you use one built around xformers...


CD players, radio tuners etc all provide low Z drive. And power amps
need whatever drive the designer calls for. What exactly do you
believe happens to a power amp if the source impedance is not
perfectly low?

d

Pearce Consulting
http://www.pearce.uk.com

"François Yves Le Gal" wrote:

On Tue, 26 Apr 2005 14:26:49 GMT, (Don Pearce) wrote:

Everything else should feed directly into a power amp
via nothing more than a volume control.

BS. Power amps love low Z drive : at least a buffer stage is required after
the volume control, excet if you use one built around xformers...

It all depends on the set up.

If I had an insensitive power amp, a preamp
used with a 200 mV line source would be sensible, ya?

A volume control to suit CD output could be down to 20k
without overloading the outout of a CD player which is usually
an opamp with a shirtload of NFB used and then with a series 390 ohms approx
to prevent a shorted output blowing the opamp.

A 20k pot has a maximum Ro = 5k in the -6Db position
if the source is a low Ro, such as a CD player.
But with a volt from a CD player, most pots for most power amps are set for say

0.2v average, at -14 db and where the 20k pot has R0 about = 4k.
Even a 50k pot has Ro = 8k approx at -14 dB.

Most power amps have an input of minimum 20k,
and although having say 10k Ro of a pot driving 20k iRi
of a power amp seems not right, it is actually quite alright, just crank the
volume
a bit for the signal you want.

Good power amps have very low input capacitance.
Maybe only 220pF max.
So if Ri of a power amp = 1 M, then 220pF
with 50k pot in the -6 dB position where pot Ro = 12.5k
will cause a HF pole at 57 kHz, somewhat ok imho,
and as I said most listening is done at -14 dB,
and Ro = 8k so the HF pole is at 90 kHz.

But I try to have power amp Rin at about 100k, and Cin = 30 pF,
since I rarely strap a cap across the input of a tube amp,
so the amp Cin comes mainly from the miller C of the input tube,
usually a mild amount because the global FB used lowers the
effective input triode tube gain and thus Cin is very low.
Therefore the only other C we need to worry about is the cable from a pot
in a box for a passive intefgrated control unit.
The cable should of course have less than the industry standard of
68 pF per metre, so 2 metres of cable and Cin
can be effectively 200pF and this causes the pole at HF at 100 kHz
for the -14 dB setting of the passive 50k gain control.
Things are better with a 20k pot.

Listening tests with a decent pot in a box instead of an active preamp
don't indicate to me that passive input circuits used with power amps are bad.


A solid state horror with 1,000 pF across the input to
deliberately reduce HF input and assure stability would need a lower input
drive impedance and a pot followed by a cathode follower will do,
especially if its a tube such as 6DJ8, with both halves parallel,
or a small trioded power tube like 6AQ5, 6BQ5, 6AR5, EL95, etc.

CF input is low Cin, low Ro out.
EL84 in CF give Ro = about 110 ohms.
The pot output sees a low Cin, so bw with a pot plus CF is over 200 kHz.

Not to worry, whatever the problem, there is a solution.

Patrick Turner.

Don Pearce wrote:

On Tue, 26 Apr 2005 16:52:55 +0200, François Yves Le Gal
wrote:

On Tue, 26 Apr 2005 14:26:49 GMT, (Don Pearce) wrote:

Everything else should feed directly into a power amp
via nothing more than a volume control.

BS. Power amps love low Z drive : at least a buffer stage is required after
the volume control, excet if you use one built around xformers...


CD players, radio tuners etc all provide low Z drive.

CD players may have universally low Z outputs.
But a pile of older gear has Ro
about equal to the plate resistance of a triode, since the that's
where the radio output etc is from. And signals were often at between
70 mV and 300 mV for older gear
So some sort of buffering and or buffering is a good idea.


And power amps
need whatever drive the designer calls for. What exactly do you
believe happens to a power amp if the source impedance is not
perfectly low?

The ppl involved work it all out.

Patrick Turner.



d

Pearce Consulting
http://www.pearce.uk.com

"Iain M Churches" a écrit dans le message de news:
...

"Don Pearce" wrote in message
...
On Tue, 26 Apr 2005 12:45:47 GMT, Patrick Turner
wrote:

Could be my ISP has fixed a problem but I think I have missed a day's
postings.

Anyway, in case I have, and in case ppl are unaware, I repeat that
I posted 5 line stage schematics at a.b.s.e 2 days ago.
Maybe they are gone but they are at
http://www.usenet-replayer.com/cgi/content/archive

Do a search there on a.b.s.e, and save as usual.

Patrick Turner.

A gain of 4x from just three valves. And at a distortion level of only
.02% at 1V out (from a rail of 300V).

Well, I'm impressed - state of the art indeed.

Even on a tubes news group, this needs some explaining.

d

With the high levels from CD players, high gain pre-amps are
not really needed. People on this group were interested to know
about a low gain stage. x4 is useful.

Iain

Just one more glowing thing on the shelf.
Some loves ;)

Yves.

Don Pearce wrote:

On Tue, 26 Apr 2005 12:45:47 GMT, Patrick Turner
wrote:

Could be my ISP has fixed a problem but I think I have missed a day's
postings.

Anyway, in case I have, and in case ppl are unaware, I repeat that
I posted 5 line stage schematics at a.b.s.e 2 days ago.
Maybe they are gone but they are at
http://www.usenet-replayer.com/cgi/content/archive

Do a search there on a.b.s.e, and save as usual.

Patrick Turner.

A gain of 4x from just three valves. And at a distortion level of only
.02% at 1V out (from a rail of 300V).

Well, I'm impressed - state of the art indeed.

Even on a tubes news group, this needs some explaining.

d

Pearce Consulting
http://www.pearce.uk.com

Nothing new here. And it may be a surprise to some, but none of the five
circuits tendered for our inspection qualify as a real line stage. To
drive an audio line these days means a 600 ohm line. None of the circuits
shown can do that at a decent level, since all are running low current. I
wonder what kind of line Patrick means to drive.

And you might want to ask Patrick T how he measured that 0.02% distortion!

Could be wishful thinking, even into an open circuit.

Cheers, John Stewart

John Stewart wrote:

Don Pearce wrote:

On Tue, 26 Apr 2005 12:45:47 GMT, Patrick Turner
wrote:

Could be my ISP has fixed a problem but I think I have missed a day's
postings.

Anyway, in case I have, and in case ppl are unaware, I repeat that
I posted 5 line stage schematics at a.b.s.e 2 days ago.
Maybe they are gone but they are at
http://www.usenet-replayer.com/cgi/content/archive

Do a search there on a.b.s.e, and save as usual.

Patrick Turner.

A gain of 4x from just three valves. And at a distortion level of only
.02% at 1V out (from a rail of 300V).

Well, I'm impressed - state of the art indeed.

Even on a tubes news group, this needs some explaining.

d

Pearce Consulting
http://www.pearce.uk.com

Nothing new here. And it may be a surprise to some, but none of the five
circuits tendered for our inspection qualify as a real line stage. To
drive an audio line these days means a 600 ohm line. None of the circuits
shown can do that at a decent level, since all are running low current. I
wonder what kind of line Patrick means to drive.

The professional industry standards are that preamps and mixing desks and
sources
have around 600 ohms Ro, easily achieved with a cathode follower,
and that input impedances be above 10,000 ohms.
Since tube amps normally have above 100k Rin, the use of a preamp
for **line level** signals does not mean the line impedance always has
to be at 600 ohms impedance.

For common domestic amplifiers and sources, the Ro of a pramp
could be a couple of kohms without causing any problems.

I use a splendid sounding preamp which has a 12AU7 CF output
with 4 mA sinked by a CCS from the cathode.
It isn't meant to drive 600 ohms, never was, never will be,
and no need to have any ability to drive 600 ohms.






And you might want to ask Patrick T how he measured that 0.02%
distortion!Could be wishful thinking, even into an open circuit.

Cheers, John Stewart

I routinely measure 0.015% at a volt of output from my preamp stages.

I get 0.1% at 10 vrms from a typical µ follower stage.

At 1 volt the thd has fallen to 0.01%, and at the intended levels used
of 0.1 vrms, its difficult to measure the thd, since
my oscillator produces 0.002%.

The signal from any preamp goes to a darlington complementary
emitter follower pair wired as a high Z input class A buffer with thd lower
than 0.001%, even at 5 vrms input/output.
Then I have an LC bridged T to null out the 1 kHz signal by about 70+ dB,
followed by a high pass filter to remove hum, then
then an opamp to lift the distortion voltage 20 dB, then a hi-pass filter
with a very steep cut off below 2 kHz, so that an additional
18 dB of filtering is done at 1 kHz, and the
distortion voltage is taken to a CRO, where I have a calibrated scale beside
the screen
for low signal voltage measurements when the CRO is turned to its maximum
sensitivity.

I also have a volt meter with 0-10mV full scale reading, so that
1 mV measurement actually is reading 0.1 mV of actual signal
distortion because I have amplified the distortion signal to be able to see it

and measure it.
I used cheap NE5534 low noise opamps.

The 1 kHz oscillator and filter unit is all self designed and made.
Its been running trouble free for 6 years.
The oscillator is a wien bridge type with a lamp bulb in the
NFB loop and opamp based, followed by a voltage range switch
to select 4 ranges of voltages followed by a another voltage amp
with bandpass character to pass 1 kHz but with attenuation of
500Hz and 2 kHz of about 20 dB to thus filter the oscilator's
harmonics and produce a final output of 0.002% thd.
A 5 k pot adjust the output level to use with any amplifier.

The purity of the oscillator is checked by passing the oscilator signal
through the bridged T filter.

The same LC bridged T filter can be connected to
up to 100v rms if the input buffer is switched out
for testing outputs of power amps.
Its Zin is about 10k, too low to test preamps direct.

If the test signal is 0.5 vrms, and distortion is 0.01%,
then the distortion voltage 50 uV.
When this is amplified 10 times by the distortion amp after the
bridged T filter we get 0.5 mV, just enough to measure and see.
When the graph of thd is drawn on a suitably large scale and with logarithmic
scale
for both output voltage and distortion, its easy to plot the distortion of an
amp
from say 50 vrms output down to 0.5 vrms output, and
with all class A triode amps the thd slopes down towards o.o% at o.o vo,
and since we have a log graph you will never see 0.0% or 0.0v of output
but you can project an interpolated line to get the thd fairly reliably
down to where it must be 0.001%.

However, quite a few solid state amps I have built yeild
0.005% thd at 250 watts of output, ( 44.7 vrms into 8 ohms ) with thd
declining towards zero
with declining output voltage and by 4.47 vrms, or 2.5 watts, it
has become difficult to measure thd because it has reduced to such a low
level.
This level of performance is quite routine, the product of carefully
thought out gain amps, careful earth pathing, and a shirtload of applied NFB.
Its usually much easier to build a low thd measuring mosfet based amp than a
bjt based amp.
It is impossible to build a tube power amp that gives much less that 0.01% thd
at
2 watts.

I have to thank the authors of Wireless World for the wealth of seriously
concise information about distortion measurements and anyone with access
to library archives with all the back issues to 1917 would be well advised to
read the lot
for any articles on audio engineering, and photocopy as much as possible
to get the kind of detail not so easily found on the web these days.

Most folks these days couldn't give a **** about thd measuring done the way I
do it,
and they just connect the amp to a sound card on a PC and use a program like
SpecLab
to analyse the signal for all the harmonics at any F, and that's what I would
do
if I had a PC out in my grubby workshop, but I see little real need
to set up an old PC out there, I can get a thd figure in 10 minutes on just
about
any power or preamp with the existing set up.

When 0.02% thd from a preamp or power amp at say 0.5 vrms, it is
difficult to measure the spectra within such a small signal.
But the CRO allows us to see if we have 2H and 3H if in equal proportions,
and usually these are the main thd spectra measured.
If 3H is more than 10 dB below 2H in a 0.02% thd signal it becomes hard to see

the typical waveform effect in a low level test signal because noise
across the bandwidth of the test filter
is about 10 kHz, and it blinds us to the thd content below -80 dB.

I also can check IMD with another box full of opamps and high Q tunable
bandpass filter, but I rarely use it.

After building such test gear one realizes how difficult it is to
create stable reliable low noise test circuitry.

Patrick Turner.