Andre Jute wrote:
On Jan 2, 11:46 am, Patrick Turner wrote:
The other thing is that loading the top tube cathode does slightly make
the circuit work as a SRPP, but the effect is minor.
Actually, any circuit which works even partially like an SRPP is a
good circuit, especially if the advantage of the SRPP it mimics is a
low impedance. I'm not always so sure that the mu-follower is truly
worth the additional complication over an SRPP for anyone except the
purist who can, as a consolation for the extra expense and
complication, mutter to himself over and over, "Yes, but an SRPP isn't
a real current source and a mu-follower is an exemplary current
source... It's true, it's true!" An SRPP with a single extra resistor
to arrange a voltage lift can in most instances be arranged to work
for practical purposes as well, and probably more reliably, than a mu-
follower.
The pure SRPP works OK for most folks if they like simplicity, but let's
consider some facts.
For SRPP, where Rk of the top and bottom triode are equal, Rout is less
than the Ra of one triode, but several times that of a pure CF. If you
move to µ-follower where the Rk between top and bottom triode equals at
least half the Ra for one triode then the Rout plummets to be much
closer to that of a CF. The penalty paid for the µ-foll is an extra Rk
of at least Ra/2, and a bias resistor, and a capacitor to couple the
bottom triode anode to the top triode grid. Maybe $2.00.
With SRPP, one can tailor the Rk between the two tubes and the load to
nearly equalize the ac anode current, but there will always be slightly
more Iac in one tube than in the other when a load is connected so full
series PP operation with complete cancelation of 2H never really
happens. And instead, after mucking about with loads and Rk, usually you
find the tubes not very well loaded, and the load each sees is not many
times Ra, and you get considerable 3H instead of the more benign 2H.
With µ-foll, you simply never have to worry about making the damn
circuit work to produce minimal 2H cancelation. Its you have series PP
where the top tube acts as a "bootstrapped follower", ie, quasi CF with
low THD, and the bottom tube has a very high ac load many times Ra and
many more times Ra than the SRPP could ever offer. Thus the bottom tube
develops more gain and lower THD/IMD than most other arrangements and
you get the saving of not having to waste dc power and generate
distortion in cathode load resistors. Let's not try to escape the
miniscule cost of a few R&C parts to get a µ-foll to work better than
its primitive cousin the SRPP. In µ-foll there is mimimized load ac flow
in the bottom gain triode, since it is buffered by the CF above it.
I think the µ-foll sounds better.
I have also used triode gain stages with CCS dc supply, and a following
R load many times Ra and sonic results are the best. I really don't like
resistances used for dc conveyance to a triode in critical circuits
because they create unecessary distortion.
One of my classic examples is this good sound preamp at
http://www.turneraudio.com.au/line-preamp-2003.html
There's a 1/2 12AU7 driving a 50k gain pot. But the anode is a CCS. If
there was a dc load R the value would have been about 33k. With the 50k
pot the load would have been about 20k, and only 2Ra at 4mAdc. But I
have RL at 5Ra approx, and THD is low. It'd even be lower if the pot had
been 100k, which may have been even better, as I have in my later Nemo
line stage circuit at
http://www.turneraudio.com.au/preamp...+psu-2005.html
Okay, I don't want to sound like a cost accountant, but someone must
be realistic.
Sometimes it costs more to employ a bean counter than using a few more
beans :-)
Patrick Turner.
Quite a bit of other good stuff, some of it at least arguable on
grounds of taste, some of it indubitably right, snipped in the
interest of bandwidth.
Andre Jute
The tubes tend to make people believe in a god, and SS leads them to
the
devil. -- Patrick Turner