Fabio Berutti wrote:

I always wandered: is a grounded cathode directly followed by a DC coupled
cathode follower really different from a symmetric SRPP? In both cases the
triode producing the gain is not loaded by the following stage and the
output impedance is low. The SRPP, being an active load, should be slightly
more linear...

SRPP is very different to a common cathode stage followed by a direct coupled
CF.

The SRPP has Ro = about Ra / 4 if you are lucky,
and the thd differs according to the load value, becoming lower
as one loads the cathode output, and then rising again
when below a critical RL value.
Its queer, but they don't call SRPP shunt regulated **push pull**
for nothing.

If you have 6SN7 in SRPP with 1.8k between top and bottom tubes and 1.8k
for bottom Rk, then the top grid can be direct connected to the
botton tube anode.

Without any output load whatever, the top tube plus its Rk = 1.8k
appear as a load for the bottom tube = Ra + [ ( u + 1 ) x Rk ] = about 10k +
38k = 48k.
This active load is no better than 47k used as a fixed resistor
in terms of gain and thd.

When a load is connected to the top tube's cathode the load affects the thd in
that
there is more current drawn though the top tube, but the Ro isn't
as low as the case of gain + CF set up.
The careful choice of load allows fairly good reduction of
2H at some critical load, since each triode can be coaxed into passing
the nearly but never equal signal currents in a quasis PP manner, so that 2H
cancelation
occurs to a useful amount. But usually the 3H slightly increases,
but it still is a good way to use a pair of triodes,
since the total input power needed for the SRPP circuit is half that for the
gain + CF circuit..

The SRPP does not completely buffer its gain tube.

The gain + CF does completely buffer the output from the gain tube.

The CF is subject to the NFB of its follower topology, so that it will
produce only 1 / A times the thd of the gain tube where A = gain.
So the CF is a pretty blameless customer concerning thd.

Better still imho is to use a µ follower, where the top
tube is biased from a fixed supply of about 0.6 x B+,
and there is a 12k R between top k and bottom a,
then the top triode acts more like a real CF, doing most of the work
on the load, and the bottom tube then sees an active load
far higher than 47k mentioned above, and its thd is far lower as a result and
gain is
closer to µ, or about 18 in the case of 6SN7.
The bottom triode anode must be cap coupled to the top grid,
and the Rk of the bottom triode sets the current flow through both seriesed
tubes.
Rk of the bottom tube needn't be bypassed because the current NFB
is rendered ineffective because such a small current change occurs in the
µ follower bottom gain triode.
Little 3H is ever created or provoked by the µ follower config.

The µ follower will have Ro much lower than SRPP,
and the buffering action is on a par
with gain+CF stage.

Either way it is done, it sounds well.

Swapping brands of 6SN7 probably will have more effect
on the sound than the topology.
There is utterly no logic to this claim since thd measurements in
triode preamps with only 0.1 vrms output usually
have quite low thd, less than 0.015%, all 2H,
regardless of the topologies mentioned, so the sound changes heard
have nothing to do with thd/imd, and are due to the mystery of tubacious
phenomena.
We really should just thank the God of Triodes for his audio
wonders, and just accept the fidelity that triodes offer, rather than bicker and
complain,
since he is a God in charge of a heck of a lotta high voltage........

Patrick Turner.







(BTW my "high gain" line preamp uses the first solution, with two ECC82s.
The 6N30 could be a great tube for SRPP 'cos it allows 300V between heater
and cathode.)

Ciao

Fabio

"Patrick Turner" ha scritto nel messaggio
...


Fabio Berutti wrote:

It is difficult to get such a low gain out of tubes without using
negative
feedback, To keep it simple I'd use a 12SN7 in symmetric SRPP without
cathode resistor bypass cap. It will give an open loop gain of about 7 -
8
and provide almost "DC to light" bandpass.
Otherwise, just make a plate follower and adjust FB to suit Your needs.
Be it a SRPP or a PF, I'd operate this unit out of a 250V tube stabilised
PS, there's a simple schematic of one of those using the 6GV8 tube (which
costs about 0.8$) plus some zeners (just google a bit).

The 12SL7 is a high gain tube. Great sound, but better used for other
purposes.

Ciao

Fabio

He wants a buffer he says, so maybe he needs a cathode follower output.

But first he can have the gain pot right after the input source select
switch, and 50k would be OK, then have simple R loaded gain stage using
1/2 a 6SN7 say 47k RL and 1.8k Rk, no need to bypass it
and that will give a gain of about 11.
The CF can be direct coupled to the anode of the gain stage and gain will
then
be about
10 overall.
A 2 uF cap off the CF should be used to drive the output with a 1.0 M to
0V
to keep the output biased at 0V.

The wiper of the input pot can have 80k to the grid of the
gain tube, and then a feedback resistor of 470k from input grid to the
output.

This will reduce the gain to around 4, or 12 dB.

The preamp is then an inverting amp.

For a non inverting amp, a different approach is required.

The source switch feeds the volume pot as before.

Then a differential pair is set up with the first tube a CF, second is a
gain
tube with 47k RL
The cathodes are commoned, and taken to -200v via a 47k.
The input CF has its grid taken to the pot wiper, with a 1M safety biasing
R to
0V.

The anode of the gain tube has the same phase as the input.

This anode is direct coupled to a second CF and the output is taken from
the
cathode
and via a 2 uF cap. The load on the CF is a 47 k to 0V.

But to reduce the expected gain of around 12, the is a resistance divider
from
the
output terminal to 0V using say 100k and 20k, and the junction of the
two R is taken to the gain triode's grid, where the voltage acts as NFB
to reduce gain fron say 12 to about 4.
The guessed value of the 20k could be adjusted easily with a temporary pot
to exactly adjust gain, then a fixed R soldered in when the value is
found.
Such a preamp will have good BW, and have Ro less than a CF,
since the CF naked Ro is further reduced by the NFB applied to trim the
gain.

Tubes are all 6SN7, no need to bother with 6SL7, since the
gain reduction needed will be high, so a lot of loop NFB is needed.
I am not saying 6SL7 cannot or must not be used.
In both ideas mentioned here the gain tubes could be 6SL7, and the
buffer CF can be a 6SN7.
Loadings for the 6SL7 would need to be around 3 times the 6SN7 values.

Placing the gain pot before the gain tubes in this case
results in slightly more noise in the preamp but since the gain wanted is
only
3 to 5, noise increase will be insignificant.
With this arrangement the distortion is the lowest since the
only the low level output signal from the gain pot is amplified.
If the signal was first amplified before being sent to the
gain pot the level from a CD player would raise the signal to
10vrms before attenuation down to say 0.1vrms for
loud play into a power amp, and although the 6SN7 would easily
cope with making 10vrms, the thd would have crept up.

The second option using a CF cathode coupled driving a gain tube could
have the

output CF eliminated if the Rout of around 3k was acceptable
and where the NFB loop resistors didn't load down the gain tube.

Thought about a 6H30 as a preamp tube?
Its rather good sounding, and has nice low Ra, so option 2 without the
second
OPT is a real goer.
Simple and effective.



Patrick Turner.



"BHV" ha scritto nel messaggio
. ..
Hello

I working on a tube preamp.

I need a buffer with a gain of 3 to 5, based on a 12SN7 or 12SL7 valve.
Can
anybody help a newbee.

Regards

Benny