Gregg wrote:

Behold, robert casey signalled from keyed 4-1000A filament:

Been playing around in simulation with a SRPP circuit using a triode on
bottom and a pentode up top. Reason for the pentode is so I can filter
its screen supply to get better power supply ripple rejection. Otherwise
it would be a triode connected pentode. Simulations suggest that I can
get about 50dB power supply rejection with the pentode with a well
filtered screen supply. Say the triode is half a 12AU7 and pentode a
6AU6, or a 6AU6 triode connected on the bottom, and another 6AU6 up top.

Thaks Robert!

/me model addict :-)

I did the SRPP triode/pentode thing (6GH8A) and found great PSRR by tying
the screen bypass cap to ground, rather than cathode. Is that the way
you've gone?

Another way is to have a CCS transistor to supply the top tube G2 current,
and bypass the G2 to the K with a 100 uF.

This then makes the top tube work in true pentode cathode follower mode,
with good supply noise rejection, and the load
imposed by a screen supply dropping R is avoided since it is normally coupled
to the
B+ via the Cbypass.
Alternatively, have a higher B+ so a larger value dropping R can be used.
The EG2 should still be well below Ea for the top tube.

Patrick Turner.