Patrick Turner's preamp circuit looks great.

How does it sound? Patrick?

TIA,

Jon

Behold, Jon Yaeger signaled to us from keyed
4-1000A filament:

Patrick Turner's preamp circuit looks great.

How does it sound? Patrick?

Patrick is biased (pun intended).

I say Pat should send each one of us a sample preamp for independent
review ;-)

--
Gregg
http://geek.scorpiorising.ca
"It's probably useful - even if it can't be SPICE'd."

Jon Yaeger wrote:

Patrick Turner's preamp circuit looks great.

How does it sound? Patrick?

TIA,

Jon

I found there were no other preamps which did any better, and I tried a
few,
so I stuck with the basic design shown for 3 years, but for 5 years
before that
it was very similar, with SRPP basic stages throughout.
The only bother was than my pre had provision for only MM,
or high output MC, such as the Sumiko BPS.

Recently, I wanted to accommodate the use of low output MC for the
phono,
so the 12AX7 input tube wasn't quiet enough for the low output
Ortophons, which put out only 0.125 mV, 32 dB less than
the 5 mV of an MM Shure V15, which I have been using for awhile now.
I'd like to get a Denon 103, 304, 303, etc, since I think they are THE
best value
in carts for musicality, and are a fraction of the price of more
pretentious MC, often costing
grands and grands..
So I revised the whole amp, so a 2SJ369 fet is at the phono input,
and it drives a 6EJ7 (trioded), in a cascode input stage.
The following stages are all bootstrapped follower, ( or if you must be
modern,
then call it a mu follower, ) like the existing
12AX7 phono input stage at my website, and all have fixed bias applied
to the top
tube, and a larger value R between the top and bottom tubes, and cap
coupling
between the bottom tube anode and top tube grid.
The higher value R increases the ohmic load seen by the bottom gain
tube,
and the load seen by the top tube follower, thus reducing thd about 6 to
12 dB
over what is shown at the existing circuit, and this allows a lower RL
to be connected,
such as a 50k log volume control, instead of a 100k.
The Ro of the bootstrapped follower is nearly as low as a pure CF, but
much lower than a SRPP, so recording off the phono output is better.
The extra RC couplings don't ruin the sound;
I just don't agree well done RC grid couplings ever do.
( flame suit being put on right now).
This keeps the response below 2 Hz tamed, but still gives a pole at the
amp output of 5 Hz,
despite even more RC couplings.
The sound is slightly better imho.
The thd in phono is less than 0.1% at 10vrms output, with the gain at
the 12o'clock position.

I still have a deletable gain block, which has the balance control as
part of
what is now a 12AU7 line stage's shunt FB network.
The balance pot is 100k linear, in series with 33k, and so
I have 103k between input and grid, then 390k between grid and anode
cap,
so the FB is mild, and gain overall is only about 4 for the line stage,
which seems enough for me.
When the balance control is used, one channel is boosted, the other is
cut,
by varying the input resistance to the shunt FB network.

When the line stage is now switched out, the balance control is also
deleted,
which is one more thing I don't really need most of the time.
Most amps I have are sensitive to 0.5vrms input, so
the usual 0.2vrms inputs from line sources still give high enough
listening levels.

The same deleatable tone control stage has been retained, except that
the
maximum amount of boost and cut has been limited to
+/- 8 dB at 100 Hz and 10 kHz, which is plenty.
Dull LP can be brightened slightly, harsh CDs cut a bit.
The tone control stage is slightly simpler, as attempts to make the
gain exactly unity, so switching the stage in or out gives no signal
change,
has been abandoned, but although there is now a 1 dB insertion loss,
with this baxandall feedback type of tone stage, there is no detectable
sound quality
change in the flat position, which allows square waves to go through OK.

The CF outputs now all have transistor CCS loads for the DC, using
MJE340.

The power supply has no B+ regulation.
I was using a HV darlington pair transistor to make an emitter follower
buffer
for the B+, but when I examined its operation after 3 years, it seemed
the input transistor
had died, which made its effectiveness quite lousy, so into the bin went
that idea,
and I now have a CLCRCRC type of supply,
with 470 uF, 5 H choke, 470 uF, then separate B+ rails
with 235 ohms, 470 uF, to a star point for each B+, then various R
values between 1k8 and 4k7 to the 100 and 47 uF caps
for each stage's B+, except the phono, which has a 470 uF cap.

The 470 uF caps are cheap, and great value, since they are used in their

millions in SMPS, etc, and plain generic types will do,
but these still have some esr, or series R and inductance, which is
effectively
in series with the value of capacitance, so in effect, at RF,
the electros in a PSU do little to keep the RF from the PS
out of the amp.
I have a filtered IEC chassis plug now fitted at the mains input,
and of course there is the 2 uF rail caps at each stage,
and RF above the BW where the electros are used is kept out.
2 uF has only 0.159 ohms of impedance at 500 kHz, so when used
where a PS circuit has plenty of series R from the rectifier,
and with anode input R of thousands of ohms, RF entry is barred
effectively.


In phono, with the tone control bass boost turned up, gain stage in,
and volume turned up full, there is no sign
of the yo-yo of the cro trace even with the fet gain set for MC,
so the passive filtering works as well as most regulators I have tried,
ie, when they worked.
The mains at my place changes level
due to the supply being shared by one and all, and the switching
noise from neighbours and myself changes the mains levels,
and hence the B+ rail changes up and down continuously.
It actually dances about at about +/- 30 mV, so the regulator
or passive filter + amp LF response needs to be able to reject this LF
instability.

The heater DC is still regulated with SS. All the heaters run on DC.
The heaters to the 6EJ7 are fed in through RF chokes, and heaters well
bypassed
to 0V locally with 0.05uF caps. Using the 6EJ7 was
not entirely a dream run; I got oscillations at 50 MHz? until I reduced
the circuit area size to about 50 mm square, and bypassed everything
with both electros and plastic caps.
The rails of each stage have their electro cap mounted away from the
board,
but near the anode supply on the board, there is a 2 uF from B+ to 0V,
to bypass any RF in the anode leads, up to 200 mm long, from the electro
caps.
But the phono stage has its B+ caps of 470 uF mounted nearby,
and with 2 uF extra plastic bypass caps.
This didn't stop the oscillations, which seemed to occur due to
insufficient bypassing of the
6EJ7 grid to ground, as it is a grounded grid tube.
Don't be a smartarse, and try to use the 6EJ7 in pentode, with grounded
grid,
its asking for serious RF instabilities and horrendous weird
microphonics,
like the bells of St Marys have been mounted right there inside the lil
ol toob.
I found a couple of good condition wafer switches which are about 40
years old,
and when cleaned up, gave better performance than the new Radio Shack
style modern switches for deleting gain and tone control stages.

I had to slightly revise the 0V layouts to keep hum to a low level,
since this is a non balanced circuit, and the chassis is connected
to the 0V rail at one point only near the line stage and phono inputs,
and then
via 100 ohms, and a 0.1 uF cap in parallel. This seemed to control hum,
and RF pick up, since the BW is 100 kHz.

I made a new box for the PS, which now has a mild steel sheet bottom, as
well
as non perforated top, except for a row of vent holes top and bottom.
The transformers for heaters and B+ are both potted, and all diodes are
inside the box.
0.008 uF caps across each half primary of the B+ winding, and
0.1 uF across the heater winding removed the tiny blips I got
at 100 Hz intervals at the output, due I guess to the silicon diodes
radiating their poisonous
signals to the rest of the circuit, despite all the screening,
and screened wire in couplings I used.
With fet front end for phono, keeping mains wires away from
leads form TT is essential, and I have 400 mm between the PS at one end
of the chassis,
and fets and phono input at the other, barely enough distance, but the
better
PS box allows it. During breadboard trials of the fet stage, I used
a breadboard supply, with no steel box, or potted trannies, and it
had to be kept 1.2 metres away during tests, lest it spoil the testing
with
induced hum.

Unless you have a chassis 500 mm long, I suggest using only remote
power supplies, and then always in a steel box with potted trannies.
A rough mild steel box, close fitted around the heater tranny, filled
with dry sand,
then sealed with silicone, is all I use.
The B+ tranny I have was a NOS military grade potted and sealed tranny.
Its not as good as what I make, as its losses are high, so it don't have
GOSS
lams, and in summer, it gets quite warm, so just because something is
mil spec, it
don't mean its darn perfect, oh no, those contractors knew how
to get away with supplying crap to the military.
But it was sealed, and only $16 when I bought it. I have used it since
1995,
no worries.

Internal wiring in the amp uses plain copper stranded wire fort B+ rail
supplies, but
all signal wiring is single strand solid core, 0.6 mm dia ex telephone
exchange
wiring, and used carefully to avoid pick up of unwanted signals,
or to allow cross talk between channels.

Pictures of the completed revised amp will be published at the website
when I get some time,
maybe november, since I have lots to do.

Recently, I compared this preamp of mine to the Allen Wright designed
FVP which he first produced back in 1988, which I recall got favourable
reviews at the time,
and ever since.
Mine was quieter, and had twice the undistorted 1 kHz square wave output
in phono,
and had equally low thd.

Patrick Turner.

Patrick,

Thank you for your elegant and generous explanation . . . . that was really
nice of you to take the time!

Couple of quick questions:

1) Do you sell a PCB for any of the circuit, esp. the phono section?

2) Not sure I understand the reason for two CF outputs . . .

Best,

Jon

From: Patrick Turner
Organization: Turner Audio
Newsgroups: rec.audio.tubes
Date: Mon, 11 Aug 2003 10:42:37 +1000
Subject: Any "reviews" of Patrick Turner's preamp?



Jon Yaeger wrote:

Patrick Turner's preamp circuit looks great.

How does it sound? Patrick?

TIA,

Jon

I found there were no other preamps which did any better, and I tried a
few,
so I stuck with the basic design shown for 3 years, but for 5 years
before that
it was very similar, with SRPP basic stages throughout.
The only bother was than my pre had provision for only MM,
or high output MC, such as the Sumiko BPS.

Recently, I wanted to accommodate the use of low output MC for the
phono,
so the 12AX7 input tube wasn't quiet enough for the low output
Ortophons, which put out only 0.125 mV, 32 dB less than
the 5 mV of an MM Shure V15, which I have been using for awhile now.
I'd like to get a Denon 103, 304, 303, etc, since I think they are THE
best value
in carts for musicality, and are a fraction of the price of more
pretentious MC, often costing
grands and grands..
So I revised the whole amp, so a 2SJ369 fet is at the phono input,
and it drives a 6EJ7 (trioded), in a cascode input stage.
The following stages are all bootstrapped follower, ( or if you must be
modern,
then call it a mu follower, ) like the existing
12AX7 phono input stage at my website, and all have fixed bias applied
to the top
tube, and a larger value R between the top and bottom tubes, and cap
coupling
between the bottom tube anode and top tube grid.
The higher value R increases the ohmic load seen by the bottom gain
tube,
and the load seen by the top tube follower, thus reducing thd about 6 to
12 dB
over what is shown at the existing circuit, and this allows a lower RL
to be connected,
such as a 50k log volume control, instead of a 100k.
The Ro of the bootstrapped follower is nearly as low as a pure CF, but
much lower than a SRPP, so recording off the phono output is better.
The extra RC couplings don't ruin the sound;
I just don't agree well done RC grid couplings ever do.
( flame suit being put on right now).
This keeps the response below 2 Hz tamed, but still gives a pole at the
amp output of 5 Hz,
despite even more RC couplings.
The sound is slightly better imho.
The thd in phono is less than 0.1% at 10vrms output, with the gain at
the 12o'clock position.

I still have a deletable gain block, which has the balance control as
part of
what is now a 12AU7 line stage's shunt FB network.
The balance pot is 100k linear, in series with 33k, and so
I have 103k between input and grid, then 390k between grid and anode
cap,
so the FB is mild, and gain overall is only about 4 for the line stage,
which seems enough for me.
When the balance control is used, one channel is boosted, the other is
cut,
by varying the input resistance to the shunt FB network.

When the line stage is now switched out, the balance control is also
deleted,
which is one more thing I don't really need most of the time.
Most amps I have are sensitive to 0.5vrms input, so
the usual 0.2vrms inputs from line sources still give high enough
listening levels.

The same deleatable tone control stage has been retained, except that
the
maximum amount of boost and cut has been limited to
+/- 8 dB at 100 Hz and 10 kHz, which is plenty.
Dull LP can be brightened slightly, harsh CDs cut a bit.
The tone control stage is slightly simpler, as attempts to make the
gain exactly unity, so switching the stage in or out gives no signal
change,
has been abandoned, but although there is now a 1 dB insertion loss,
with this baxandall feedback type of tone stage, there is no detectable
sound quality
change in the flat position, which allows square waves to go through OK.

The CF outputs now all have transistor CCS loads for the DC, using
MJE340.

The power supply has no B+ regulation.
I was using a HV darlington pair transistor to make an emitter follower
buffer
for the B+, but when I examined its operation after 3 years, it seemed
the input transistor
had died, which made its effectiveness quite lousy, so into the bin went
that idea,
and I now have a CLCRCRC type of supply,
with 470 uF, 5 H choke, 470 uF, then separate B+ rails
with 235 ohms, 470 uF, to a star point for each B+, then various R
values between 1k8 and 4k7 to the 100 and 47 uF caps
for each stage's B+, except the phono, which has a 470 uF cap.

The 470 uF caps are cheap, and great value, since they are used in their

millions in SMPS, etc, and plain generic types will do,
but these still have some esr, or series R and inductance, which is
effectively
in series with the value of capacitance, so in effect, at RF,
the electros in a PSU do little to keep the RF from the PS
out of the amp.
I have a filtered IEC chassis plug now fitted at the mains input,
and of course there is the 2 uF rail caps at each stage,
and RF above the BW where the electros are used is kept out.
2 uF has only 0.159 ohms of impedance at 500 kHz, so when used
where a PS circuit has plenty of series R from the rectifier,
and with anode input R of thousands of ohms, RF entry is barred
effectively.


In phono, with the tone control bass boost turned up, gain stage in,
and volume turned up full, there is no sign
of the yo-yo of the cro trace even with the fet gain set for MC,
so the passive filtering works as well as most regulators I have tried,
ie, when they worked.
The mains at my place changes level
due to the supply being shared by one and all, and the switching
noise from neighbours and myself changes the mains levels,
and hence the B+ rail changes up and down continuously.
It actually dances about at about +/- 30 mV, so the regulator
or passive filter + amp LF response needs to be able to reject this LF
instability.

The heater DC is still regulated with SS. All the heaters run on DC.
The heaters to the 6EJ7 are fed in through RF chokes, and heaters well
bypassed
to 0V locally with 0.05uF caps. Using the 6EJ7 was
not entirely a dream run; I got oscillations at 50 MHz? until I reduced
the circuit area size to about 50 mm square, and bypassed everything
with both electros and plastic caps.
The rails of each stage have their electro cap mounted away from the
board,
but near the anode supply on the board, there is a 2 uF from B+ to 0V,
to bypass any RF in the anode leads, up to 200 mm long, from the electro
caps.
But the phono stage has its B+ caps of 470 uF mounted nearby,
and with 2 uF extra plastic bypass caps.
This didn't stop the oscillations, which seemed to occur due to
insufficient bypassing of the
6EJ7 grid to ground, as it is a grounded grid tube.
Don't be a smartarse, and try to use the 6EJ7 in pentode, with grounded
grid,
its asking for serious RF instabilities and horrendous weird
microphonics,
like the bells of St Marys have been mounted right there inside the lil
ol toob.
I found a couple of good condition wafer switches which are about 40
years old,
and when cleaned up, gave better performance than the new Radio Shack
style modern switches for deleting gain and tone control stages.

I had to slightly revise the 0V layouts to keep hum to a low level,
since this is a non balanced circuit, and the chassis is connected
to the 0V rail at one point only near the line stage and phono inputs,
and then
via 100 ohms, and a 0.1 uF cap in parallel. This seemed to control hum,
and RF pick up, since the BW is 100 kHz.

I made a new box for the PS, which now has a mild steel sheet bottom, as
well
as non perforated top, except for a row of vent holes top and bottom.
The transformers for heaters and B+ are both potted, and all diodes are
inside the box.
0.008 uF caps across each half primary of the B+ winding, and
0.1 uF across the heater winding removed the tiny blips I got
at 100 Hz intervals at the output, due I guess to the silicon diodes
radiating their poisonous
signals to the rest of the circuit, despite all the screening,
and screened wire in couplings I used.
With fet front end for phono, keeping mains wires away from
leads form TT is essential, and I have 400 mm between the PS at one end
of the chassis,
and fets and phono input at the other, barely enough distance, but the
better
PS box allows it. During breadboard trials of the fet stage, I used
a breadboard supply, with no steel box, or potted trannies, and it
had to be kept 1.2 metres away during tests, lest it spoil the testing
with
induced hum.

Unless you have a chassis 500 mm long, I suggest using only remote
power supplies, and then always in a steel box with potted trannies.
A rough mild steel box, close fitted around the heater tranny, filled
with dry sand,
then sealed with silicone, is all I use.
The B+ tranny I have was a NOS military grade potted and sealed tranny.
Its not as good as what I make, as its losses are high, so it don't have
GOSS
lams, and in summer, it gets quite warm, so just because something is
mil spec, it
don't mean its darn perfect, oh no, those contractors knew how
to get away with supplying crap to the military.
But it was sealed, and only $16 when I bought it. I have used it since
1995,
no worries.

Internal wiring in the amp uses plain copper stranded wire fort B+ rail
supplies, but
all signal wiring is single strand solid core, 0.6 mm dia ex telephone
exchange
wiring, and used carefully to avoid pick up of unwanted signals,
or to allow cross talk between channels.

Pictures of the completed revised amp will be published at the website
when I get some time,
maybe november, since I have lots to do.

Recently, I compared this preamp of mine to the Allen Wright designed
FVP which he first produced back in 1988, which I recall got favourable
reviews at the time,
and ever since.
Mine was quieter, and had twice the undistorted 1 kHz square wave output
in phono,
and had equally low thd.

Patrick Turner.

Jon Yaeger wrote:

Patrick,

Thank you for your elegant and generous explanation . . . . that was really
nice of you to take the time!

Couple of quick questions:

1) Do you sell a PCB for any of the circuit, esp. the phono section?

Nope, dedicated DIYers will make their own board, which should be easy.



2) Not sure I understand the reason for two CF outputs . . .

I sometimes hook up two power amps, and to test them,
I simply switch the speakers from one to the other, and when I don't notice
there
is a level change, I have the levels equal enough, and a listening test can be
done.

Normally, one could have both halves of the 12AU7 CF paralleled,
to make a nice low Ro. One tube per stage, per channel.
Tubes are cheap, so it makes sense.
And the amp can be that much more spectacular.

Maybe I should add yet another triode stage to filter signal coming raw from the

DA converter in a CD player.
I don't think anyone does this, they try to put a tube into an existing player,
where it looks a bit silly, and often isn't set up very well.

Patrick Turner.



Best,

Jon

From: Patrick Turner
Organization: Turner Audio
Newsgroups: rec.audio.tubes
Date: Mon, 11 Aug 2003 10:42:37 +1000
Subject: Any "reviews" of Patrick Turner's preamp?



Jon Yaeger wrote:

Patrick Turner's preamp circuit looks great.

How does it sound? Patrick?

TIA,

Jon

I found there were no other preamps which did any better, and I tried a
few,
so I stuck with the basic design shown for 3 years, but for 5 years
before that
it was very similar, with SRPP basic stages throughout.
The only bother was than my pre had provision for only MM,
or high output MC, such as the Sumiko BPS.

Recently, I wanted to accommodate the use of low output MC for the
phono,
so the 12AX7 input tube wasn't quiet enough for the low output
Ortophons, which put out only 0.125 mV, 32 dB less than
the 5 mV of an MM Shure V15, which I have been using for awhile now.
I'd like to get a Denon 103, 304, 303, etc, since I think they are THE
best value
in carts for musicality, and are a fraction of the price of more
pretentious MC, often costing
grands and grands..
So I revised the whole amp, so a 2SJ369 fet is at the phono input,
and it drives a 6EJ7 (trioded), in a cascode input stage.
The following stages are all bootstrapped follower, ( or if you must be
modern,
then call it a mu follower, ) like the existing
12AX7 phono input stage at my website, and all have fixed bias applied
to the top
tube, and a larger value R between the top and bottom tubes, and cap
coupling
between the bottom tube anode and top tube grid.
The higher value R increases the ohmic load seen by the bottom gain
tube,
and the load seen by the top tube follower, thus reducing thd about 6 to
12 dB
over what is shown at the existing circuit, and this allows a lower RL
to be connected,
such as a 50k log volume control, instead of a 100k.
The Ro of the bootstrapped follower is nearly as low as a pure CF, but
much lower than a SRPP, so recording off the phono output is better.
The extra RC couplings don't ruin the sound;
I just don't agree well done RC grid couplings ever do.
( flame suit being put on right now).
This keeps the response below 2 Hz tamed, but still gives a pole at the
amp output of 5 Hz,
despite even more RC couplings.
The sound is slightly better imho.
The thd in phono is less than 0.1% at 10vrms output, with the gain at
the 12o'clock position.

I still have a deletable gain block, which has the balance control as
part of
what is now a 12AU7 line stage's shunt FB network.
The balance pot is 100k linear, in series with 33k, and so
I have 103k between input and grid, then 390k between grid and anode
cap,
so the FB is mild, and gain overall is only about 4 for the line stage,
which seems enough for me.
When the balance control is used, one channel is boosted, the other is
cut,
by varying the input resistance to the shunt FB network.

When the line stage is now switched out, the balance control is also
deleted,
which is one more thing I don't really need most of the time.
Most amps I have are sensitive to 0.5vrms input, so
the usual 0.2vrms inputs from line sources still give high enough
listening levels.

The same deleatable tone control stage has been retained, except that
the
maximum amount of boost and cut has been limited to
+/- 8 dB at 100 Hz and 10 kHz, which is plenty.
Dull LP can be brightened slightly, harsh CDs cut a bit.
The tone control stage is slightly simpler, as attempts to make the
gain exactly unity, so switching the stage in or out gives no signal
change,
has been abandoned, but although there is now a 1 dB insertion loss,
with this baxandall feedback type of tone stage, there is no detectable
sound quality
change in the flat position, which allows square waves to go through OK.

The CF outputs now all have transistor CCS loads for the DC, using
MJE340.

The power supply has no B+ regulation.
I was using a HV darlington pair transistor to make an emitter follower
buffer
for the B+, but when I examined its operation after 3 years, it seemed
the input transistor
had died, which made its effectiveness quite lousy, so into the bin went
that idea,
and I now have a CLCRCRC type of supply,
with 470 uF, 5 H choke, 470 uF, then separate B+ rails
with 235 ohms, 470 uF, to a star point for each B+, then various R
values between 1k8 and 4k7 to the 100 and 47 uF caps
for each stage's B+, except the phono, which has a 470 uF cap.

The 470 uF caps are cheap, and great value, since they are used in their

millions in SMPS, etc, and plain generic types will do,
but these still have some esr, or series R and inductance, which is
effectively
in series with the value of capacitance, so in effect, at RF,
the electros in a PSU do little to keep the RF from the PS
out of the amp.
I have a filtered IEC chassis plug now fitted at the mains input,
and of course there is the 2 uF rail caps at each stage,
and RF above the BW where the electros are used is kept out.
2 uF has only 0.159 ohms of impedance at 500 kHz, so when used
where a PS circuit has plenty of series R from the rectifier,
and with anode input R of thousands of ohms, RF entry is barred
effectively.


In phono, with the tone control bass boost turned up, gain stage in,
and volume turned up full, there is no sign
of the yo-yo of the cro trace even with the fet gain set for MC,
so the passive filtering works as well as most regulators I have tried,
ie, when they worked.
The mains at my place changes level
due to the supply being shared by one and all, and the switching
noise from neighbours and myself changes the mains levels,
and hence the B+ rail changes up and down continuously.
It actually dances about at about +/- 30 mV, so the regulator
or passive filter + amp LF response needs to be able to reject this LF
instability.

The heater DC is still regulated with SS. All the heaters run on DC.
The heaters to the 6EJ7 are fed in through RF chokes, and heaters well
bypassed
to 0V locally with 0.05uF caps. Using the 6EJ7 was
not entirely a dream run; I got oscillations at 50 MHz? until I reduced
the circuit area size to about 50 mm square, and bypassed everything
with both electros and plastic caps.
The rails of each stage have their electro cap mounted away from the
board,
but near the anode supply on the board, there is a 2 uF from B+ to 0V,
to bypass any RF in the anode leads, up to 200 mm long, from the electro
caps.
But the phono stage has its B+ caps of 470 uF mounted nearby,
and with 2 uF extra plastic bypass caps.
This didn't stop the oscillations, which seemed to occur due to
insufficient bypassing of the
6EJ7 grid to ground, as it is a grounded grid tube.
Don't be a smartarse, and try to use the 6EJ7 in pentode, with grounded
grid,
its asking for serious RF instabilities and horrendous weird
microphonics,
like the bells of St Marys have been mounted right there inside the lil
ol toob.
I found a couple of good condition wafer switches which are about 40
years old,
and when cleaned up, gave better performance than the new Radio Shack
style modern switches for deleting gain and tone control stages.

I had to slightly revise the 0V layouts to keep hum to a low level,
since this is a non balanced circuit, and the chassis is connected
to the 0V rail at one point only near the line stage and phono inputs,
and then
via 100 ohms, and a 0.1 uF cap in parallel. This seemed to control hum,
and RF pick up, since the BW is 100 kHz.

I made a new box for the PS, which now has a mild steel sheet bottom, as
well
as non perforated top, except for a row of vent holes top and bottom.
The transformers for heaters and B+ are both potted, and all diodes are
inside the box.
0.008 uF caps across each half primary of the B+ winding, and
0.1 uF across the heater winding removed the tiny blips I got
at 100 Hz intervals at the output, due I guess to the silicon diodes
radiating their poisonous
signals to the rest of the circuit, despite all the screening,
and screened wire in couplings I used.
With fet front end for phono, keeping mains wires away from
leads form TT is essential, and I have 400 mm between the PS at one end
of the chassis,
and fets and phono input at the other, barely enough distance, but the
better
PS box allows it. During breadboard trials of the fet stage, I used
a breadboard supply, with no steel box, or potted trannies, and it
had to be kept 1.2 metres away during tests, lest it spoil the testing
with
induced hum.

Unless you have a chassis 500 mm long, I suggest using only remote
power supplies, and then always in a steel box with potted trannies.
A rough mild steel box, close fitted around the heater tranny, filled
with dry sand,
then sealed with silicone, is all I use.
The B+ tranny I have was a NOS military grade potted and sealed tranny.
Its not as good as what I make, as its losses are high, so it don't have
GOSS
lams, and in summer, it gets quite warm, so just because something is
mil spec, it
don't mean its darn perfect, oh no, those contractors knew how
to get away with supplying crap to the military.
But it was sealed, and only $16 when I bought it. I have used it since
1995,
no worries.

Internal wiring in the amp uses plain copper stranded wire fort B+ rail
supplies, but
all signal wiring is single strand solid core, 0.6 mm dia ex telephone
exchange
wiring, and used carefully to avoid pick up of unwanted signals,
or to allow cross talk between channels.

Pictures of the completed revised amp will be published at the website
when I get some time,
maybe november, since I have lots to do.

Recently, I compared this preamp of mine to the Allen Wright designed
FVP which he first produced back in 1988, which I recall got favourable
reviews at the time,
and ever since.
Mine was quieter, and had twice the undistorted 1 kHz square wave output
in phono,
and had equally low thd.

Patrick Turner.

Robert Casey wrote:

Patrick Turner wrote:




Maybe I should add yet another triode stage to filter signal coming raw from the

DA converter in a CD player.
I don't think anyone does this, they try to put a tube into an existing player,
where it looks a bit silly, and often isn't set up very well.



I did that in one of my CD players. See
http://home.netcom.com/~wa2ise/radios/tubedac.htm

I have 3 cd players, and all have the second generation type pcbs,
with extremely tiny circuit tracks, of enormous complexity, especially the Denon
dcd1015.
The schematic of densely drawn lines spreads right across the kitchen table.
Only fools go where angels fear to tread, and going in there with a soldering iron
would be like the bull in the china shop.
The Denon chips have no numbers, and they keep secret what goes on inside the chips.

I have zero idea about whether its worth fiddling around with the oscillators, or
"clocks"
as nobody can show me the artifacts of jitter extracted shown clearly on an
oscilliscope,
and quanitified as a meaningful % of N&D.

It would be so nice if a cd player was a simple train of parts, and the
whole process could be explained in linear circuit type explanations,
but its never like this, wherever I go on the Web.

I wouldn't ever try to place a tube inside the Denon, with the circuitry which you
have crammed inside your CD player; its just not what I would do, its too messy.
I would have thought it prefable to have a cable transfer of the raw unfiltered DA
output to connect to an external filter elements mounted on a chassis, with a tube,
to get done to the signal what a tube has to do, which would seem to simply be a
sharp cut off LPF, with a pole at 20 kHz.

I found it rather difficult to follow much of the text in your webpage,
perhaps you ought to draft it with the know-nothing-about-digital tube enthusiast
in mind, since the basic info about digital is not outlined in the RDH4.

When one does find something which explains digital and the alegedly horrible things
it does badly ( jitter, errors ), much is spelled out in totally incomprehensible
language,
and surrounded by maths that is queer, with little emphasis on the basic concepts.

As enthusiasts, we like to see WHY we do things at every single tiny step,
and arguments rage here about the values of a single resistor, or a cap, and such
like.
Digital? clocks? oscillators? samples? huh? Jitter? wtf is all that?
How do we know its worth attacking the dreaded jitter if we
never see an oscillagram which says " and now folks, here is a bad case of the jitter
"
And then the follow up, "and here's what do do about it, follow these simple
39 steps," needs to be fairly precise,
for folks prone to getting confused over tube bias circuitry.

I have bookmaked your pages, and will need more time to
read them more closely, if I can understand what is the purpose for it all.

Patrick Turner.