I just posted a schematic of this preamp using triodes
and CCS loading at alt.binaries.schematics.electronic.

Enjoy,

Patrick Turner.

There is an error in the schematic at the balance control,
so i posted another with the error corrected.

Voltage gains are as follows :-

CD player, with 2vrms input, no need to have gain stage in the circuit,
and 1.16v appears at the volume control.

At the 12 o'clock position of the volume control, voltage applied to the

V2 CF will be 0.116 vrms, and 0.11 vrms will appear at the output if the

power amp Rin = 100k, and this will be plenty for most amps.

With the gain stage switched in there is a gain of 5, or 14 dB, which
would be plenty
to suit most insensitive power amps or if the source voltage
is an old fashioned 200mV.

Distortion.

If there is a signal of 200mV at the inputs,
with the gain stage switched in, 1v appears at V1 output, and thd is
less than 0.01%,
mainly 2H with 3H and other H at 15 dB below the 2H levels.

The CF buffer stage will normally run with about 0.1v output,
and thd is also a lot less than 0.01%, and makes an insignificant
contribution to the overall thd.

If there is a 2v input signal, there will be about 0.15% thd at the
10vrms V1 output,
and the V2 CF will still make an insignificant contribution to thd.

Noise.

Depends partially on the quality of input tubes chosen, and 2uV is a
good enough grid input noise figure for chosen tubes.
The shunt FB resistances will cause some noise but because there are in
a FB circuit
the noise is reduced to low enough levels.

DC should be used for the heaters, and the heater circuits for the CF
tube should be biased at
about +100v.

The B+ should be a well filtered supply using CLCRC with C1 = 470uF, L =
10H at 15mA,
C2 = 470uF, and then each channel supplied separately from branches off
with RC
of 2.2k and 470 uF. That should quell all hum and mains jitter.
B+ could be between 225v and 275v, and those addicted to regulators
may indulge themselves, but I am not one of them.

Note that to get Ea for the tubes = 120v, choose R6 and R11 correctly.
I am not sure exactly what values would be correct.

The CCS loads of the tubes reduces thd about 10 db at least below what
would be achieved
using say 25k loads to be able to get 5mA to the tubes and have the Ea
at 120v, a
linear working region.
The CCS act as compliant and passive slaves to the current demand of the
triodes,
and save them from having to provide a current change in an R which
supplies DC to the anode.
So since the signal current changes in the triode are reduced to a
minimum they
act with better linearity than with R loads that bring in the dc, and so
the use of the 5 component CCS
is well justified if we really want to hear the triodes sing properly.

NFB.

NFB is used to limit the amount of gain, and only about 10db of shunt FB
is applied
in a simple blameless application.
This reduces the Ro of V1 from about 10k to about 3k, and bandwidth will
at least be 100kHz.
thd/imd is also reduced by 10dB, and should be well below the
conservatively
quoted figures above.
The gain stage will work fine if the NFB is abolished by removing R7,
the FB resistor.

I built a preamp like this circuit using 12AU7 for a client about
2 years ago, and he is very happy and it has run faultlessly.
6CG7 NOS Telfunkens will give the best sound imho.
And probably the lowest thd/imd.
6H30 are also nice, but will have slightly lower gain than the
6CG7/6SN7.

6922 or 6DJ8 are also usable, but I think they sound a bit hard.

Remote control for the volume control is not shown.

Patrick Turner.

"Note 1" doesn't make sense either, could there be another mistake there?

Also, why didn't you arrange S2, the "gain stage bypass" switch so that
the feedback resistor R7 isn't in the circuit when S2 is in the bypass
position? All it would take is connecting R7 directly to the grid side of
S2A, rather than to the input side of S2A.

I notice you used both a grid stopper and a cathode stopper resistor on
the cathode follower stage, sort of a belt and suspenders approach.


Regards,

John Byrns


In article , Patrick Turner
wrote:

There is an error in the schematic at the balance control,
so i posted another with the error corrected.

Voltage gains are as follows :-

CD player, with 2vrms input, no need to have gain stage in the circuit,
and 1.16v appears at the volume control.

At the 12 o'clock position of the volume control, voltage applied to the

V2 CF will be 0.116 vrms, and 0.11 vrms will appear at the output if the

power amp Rin = 100k, and this will be plenty for most amps.

With the gain stage switched in there is a gain of 5, or 14 dB, which
would be plenty
to suit most insensitive power amps or if the source voltage
is an old fashioned 200mV.

Distortion.

If there is a signal of 200mV at the inputs,
with the gain stage switched in, 1v appears at V1 output, and thd is
less than 0.01%,
mainly 2H with 3H and other H at 15 dB below the 2H levels.

The CF buffer stage will normally run with about 0.1v output,
and thd is also a lot less than 0.01%, and makes an insignificant
contribution to the overall thd.

If there is a 2v input signal, there will be about 0.15% thd at the
10vrms V1 output,
and the V2 CF will still make an insignificant contribution to thd.

Noise.

Depends partially on the quality of input tubes chosen, and 2uV is a
good enough grid input noise figure for chosen tubes.
The shunt FB resistances will cause some noise but because there are in
a FB circuit
the noise is reduced to low enough levels.

DC should be used for the heaters, and the heater circuits for the CF
tube should be biased at
about +100v.

The B+ should be a well filtered supply using CLCRC with C1 = 470uF, L =
10H at 15mA,
C2 = 470uF, and then each channel supplied separately from branches off
with RC
of 2.2k and 470 uF. That should quell all hum and mains jitter.
B+ could be between 225v and 275v, and those addicted to regulators
may indulge themselves, but I am not one of them.

Note that to get Ea for the tubes = 120v, choose R6 and R11 correctly.
I am not sure exactly what values would be correct.

The CCS loads of the tubes reduces thd about 10 db at least below what
would be achieved
using say 25k loads to be able to get 5mA to the tubes and have the Ea
at 120v, a
linear working region.
The CCS act as compliant and passive slaves to the current demand of the
triodes,
and save them from having to provide a current change in an R which
supplies DC to the anode.
So since the signal current changes in the triode are reduced to a
minimum they
act with better linearity than with R loads that bring in the dc, and so
the use of the 5 component CCS
is well justified if we really want to hear the triodes sing properly.

NFB.

NFB is used to limit the amount of gain, and only about 10db of shunt FB
is applied
in a simple blameless application.
This reduces the Ro of V1 from about 10k to about 3k, and bandwidth will
at least be 100kHz.
thd/imd is also reduced by 10dB, and should be well below the
conservatively
quoted figures above.
The gain stage will work fine if the NFB is abolished by removing R7,
the FB resistor.

I built a preamp like this circuit using 12AU7 for a client about
2 years ago, and he is very happy and it has run faultlessly.
6CG7 NOS Telfunkens will give the best sound imho.
And probably the lowest thd/imd.
6H30 are also nice, but will have slightly lower gain than the
6CG7/6SN7.

6922 or 6DJ8 are also usable, but I think they sound a bit hard.

Remote control for the volume control is not shown.

Patrick Turner.


Surf my web pages at, http://users.rcn.com/jbyrns/

John Byrns wrote:

"Note 1" doesn't make sense either, could there be another mistake there?

R6 and R11 are the Rk for the triodes which determine Ia.
R10 is a stopper R for V2.

Sorry for the errors.
I am normally a little brighter at 3am after a night out playing chess,
but I have been busy lately.



Also, why didn't you arrange S2, the "gain stage bypass" switch so that
the feedback resistor R7 isn't in the circuit when S2 is in the bypass
position? All it would take is connecting R7 directly to the grid side of
S2A, rather than to the input side of S2A.

When the switch is in bypass mode, V1 grid is grounded, so the output of the
stage is silent,
and sice Ra = approx10k and R7 = 560k, it makes no difference to have 560k
to the anode circuit. In bypass mode, only one switch contact point is being
used.
But feel free to do it some other way.

The point of the exercize is to keep people awake at r.a.t.



I notice you used both a grid stopper and a cathode stopper resistor on
the cathode follower stage, sort of a belt and suspenders approach.

Somebody may try to plug in a 12AT7, which could oscillate since the open loop

gain is about 50, and the reduction of gain to 5 in V1 and to less than 1 in
V2
can lead to RF oscillations, so some grid stopping and isolation of phase
jerking
cable capacitance is only wise.

Patrick Turner.





Regards,

John Byrns

In article , Patrick Turner
wrote:

There is an error in the schematic at the balance control,
so i posted another with the error corrected.

Voltage gains are as follows :-

CD player, with 2vrms input, no need to have gain stage in the circuit,
and 1.16v appears at the volume control.

At the 12 o'clock position of the volume control, voltage applied to the

V2 CF will be 0.116 vrms, and 0.11 vrms will appear at the output if the

power amp Rin = 100k, and this will be plenty for most amps.

With the gain stage switched in there is a gain of 5, or 14 dB, which
would be plenty
to suit most insensitive power amps or if the source voltage
is an old fashioned 200mV.

Distortion.

If there is a signal of 200mV at the inputs,
with the gain stage switched in, 1v appears at V1 output, and thd is
less than 0.01%,
mainly 2H with 3H and other H at 15 dB below the 2H levels.

The CF buffer stage will normally run with about 0.1v output,
and thd is also a lot less than 0.01%, and makes an insignificant
contribution to the overall thd.

If there is a 2v input signal, there will be about 0.15% thd at the
10vrms V1 output,
and the V2 CF will still make an insignificant contribution to thd.

Noise.

Depends partially on the quality of input tubes chosen, and 2uV is a
good enough grid input noise figure for chosen tubes.
The shunt FB resistances will cause some noise but because there are in
a FB circuit
the noise is reduced to low enough levels.

DC should be used for the heaters, and the heater circuits for the CF
tube should be biased at
about +100v.

The B+ should be a well filtered supply using CLCRC with C1 = 470uF, L =
10H at 15mA,
C2 = 470uF, and then each channel supplied separately from branches off
with RC
of 2.2k and 470 uF. That should quell all hum and mains jitter.
B+ could be between 225v and 275v, and those addicted to regulators
may indulge themselves, but I am not one of them.

Note that to get Ea for the tubes = 120v, choose R6 and R11 correctly.
I am not sure exactly what values would be correct.

The CCS loads of the tubes reduces thd about 10 db at least below what
would be achieved
using say 25k loads to be able to get 5mA to the tubes and have the Ea
at 120v, a
linear working region.
The CCS act as compliant and passive slaves to the current demand of the
triodes,
and save them from having to provide a current change in an R which
supplies DC to the anode.
So since the signal current changes in the triode are reduced to a
minimum they
act with better linearity than with R loads that bring in the dc, and so
the use of the 5 component CCS
is well justified if we really want to hear the triodes sing properly.

NFB.

NFB is used to limit the amount of gain, and only about 10db of shunt FB
is applied
in a simple blameless application.
This reduces the Ro of V1 from about 10k to about 3k, and bandwidth will
at least be 100kHz.
thd/imd is also reduced by 10dB, and should be well below the
conservatively
quoted figures above.
The gain stage will work fine if the NFB is abolished by removing R7,
the FB resistor.

I built a preamp like this circuit using 12AU7 for a client about
2 years ago, and he is very happy and it has run faultlessly.
6CG7 NOS Telfunkens will give the best sound imho.
And probably the lowest thd/imd.
6H30 are also nice, but will have slightly lower gain than the
6CG7/6SN7.

6922 or 6DJ8 are also usable, but I think they sound a bit hard.

Remote control for the volume control is not shown.

Patrick Turner.

Surf my web pages at, http://users.rcn.com/jbyrns/

Hi Patrick,

thank you for this practical example of SS CC plate load
and CCS.

I am wondering if there are HV PNP transistors out there
(and obtanium) which will stand at least, say, 500 volts.

The MJE350 you use in your line stage does 300V, which
obviously is sufficient with the B+ of 250V you use in
any case.

But other tubes (in other apps) like considerable higher
voltages across them, like, say, 6SL7. While the plate
load transistor theoretically only needs to withstand the
B+ voltage minus the voltage dropped across the tube, I
would feel much better if I could use a transistor which
just in case would be able to stand the full B+.

So, got any suggestions at hand for HV PNP BJTs?

Tnx2u,

Tom

--
Do not meddle in the affairs of wizards,
for you are crunchy and good with ketchup.

in article , Tom Schlangen at
wrote on 6/5/05 3:59 PM:

Hi Patrick,

thank you for this practical example of SS CC plate load
and CCS.

I am wondering if there are HV PNP transistors out there
(and obtanium) which will stand at least, say, 500 volts.

The MJE350 you use in your line stage does 300V, which
obviously is sufficient with the B+ of 250V you use in
any case.

But other tubes (in other apps) like considerable higher
voltages across them, like, say, 6SL7. While the plate
load transistor theoretically only needs to withstand the
B+ voltage minus the voltage dropped across the tube, I
would feel much better if I could use a transistor which
just in case would be able to stand the full B+.

So, got any suggestions at hand for HV PNP BJTs?

Tnx2u,

Tom


Tom,

I scanned an NTE catalog and didn't see much over 300 Vceo.

If you can stomach a P channel power mosfet, there are some available that
will withstand 500 V.

Jon

Tom Schlangen wrote:

Hi Patrick,

thank you for this practical example of SS CC plate load
and CCS.

I am wondering if there are HV PNP transistors out there
(and obtanium) which will stand at least, say, 500 volts.

KSE350 is ok to 350v I think; there are a few others out there, but
MJE350 is OK.
Otherwise use a hv mosfet



The MJE350 you use in your line stage does 300V, which
obviously is sufficient with the B+ of 250V you use in
any case.

But other tubes (in other apps) like considerable higher
voltages across them, like, say, 6SL7.

But they only draw 1 or 2 mA, so maybe some other device is better.

While the plate
load transistor theoretically only needs to withstand the
B+ voltage minus the voltage dropped across the tube, I
would feel much better if I could use a transistor which
just in case would be able to stand the full B+.

So, got any suggestions at hand for HV PNP BJTs?

You don't need to have a high B+ with a pnp CCS transistor.

The circuit I have posted nominates Ea = 120V and Ia = 5mA.
Now if you draw the 120v + 5mA Q point on the curves for 6SN7, you will
find the
max swing down of Ea is maybe 40pk volts because its limited by
the Ra line for Eg1 = 0V, and the swing up however is from +120 to maybe
+237V, since B+
is +250V and there is about 12V minimum voltage across the transistor.

So the B+ could be +200V, and the tube working will still give 30vrms
output,
and no difference at all in thd since the DC supply device has about 20M
dynamic impedance, ie,
it is a virtual CCS regardless of where the supply B+ voltage is.

One could set the Ia at 4 mA, to allow a wider V swing, or more even in
each direction
but its not needed, since we will never need more than a few volts from
V1 anode.
The 120v and 5 mA of current is a fair way to set up a 6CG7/6SN7,
and will also well suit a 6H30, because its Ra is lower, and
it is like 3 x 12AU7 paralleled within the one tube, so the Ra lines
are steeper, and the swing down before grid current is more.

The beauty of the CCS load is that you can set how you like,
then juggle Rk values, and the anode voltage
will be where you want it for any given value of Ia, and there
is no need to calculated the values of the DC feed resistor,
and no lessening of the effective B+ due to the combined effect
of the DC RL and the AC RL.

CCS allow tubes to sing.

Patrick Turner.





Tnx2u,

Tom

--
Do not meddle in the affairs of wizards,
for you are crunchy and good with ketchup.