Gentlemen,

when paralleling output tetrodes in a PSE design using
auto bias, each tube having its own bypassed cathode
resistor, would a single coupling cap and common grid
resistor be okay, or should each tube get its own
grid resistor and coupling cap?

Tom

--
Anyone who considers protocol unimportant has
never dealt with a cat. - R. Heinlein

Tom Schlangen wrote:

Gentlemen,

when paralleling output tetrodes in a PSE design using
auto bias, each tube having its own bypassed cathode
resistor, would a single coupling cap and common grid
resistor be okay, or should each tube get its own
grid resistor and coupling cap?

Its OK to have one RC coupling network to the driver tube
with separate cathode bias arrangements.

With my 4 x 6CA7 pse amp, I have 2 uF and 68k Rg to all 4 grids, but
each
with 1.8k stoppers. The driver is a 12BH7, both halves paralleled.

If an output tube goes beserko, and draws a large grid current,
all 4 tubes increase their current, but the active protect circuit
will respond.

Patrick Turner



Tom

--
Anyone who considers protocol unimportant has
never dealt with a cat. - R. Heinlein

Hello Patrick,

Its OK to have one RC coupling network to the driver tube
with separate cathode bias arrangements.

Okay.

With my 4 x 6CA7 pse amp, I have 2 uF and 68k Rg to
all 4 grids, but each with 1.8k stoppers. The driver
is a 12BH7, both halves paralleled.

I am considering a direct coupled driver arrangement.

V1 = 1/2 6SN7, B+ 390V, cathode res (not bypassed)
470 ohms, plate res 68k. This gives ~95 volts at
the plate. gNFB injection point at cathode.

Direct coupled to two halfs of 6SN7 paralleled,
cathode resistor 12k bypassed with 68uF, plate
resistor also 12k. B+ for this stage is 390V, too.

Is there anything objectionable with such an
arrangement?

Tom

--
Live is too short to be taken seriously.
- Oscar Wilde

Tom Schlangen wrote:

Hello Patrick,

Its OK to have one RC coupling network to the driver tube
with separate cathode bias arrangements.

Okay.

With my 4 x 6CA7 pse amp, I have 2 uF and 68k Rg to
all 4 grids, but each with 1.8k stoppers. The driver
is a 12BH7, both halves paralleled.

I am considering a direct coupled driver arrangement.

V1 = 1/2 6SN7, B+ 390V, cathode res (not bypassed)
470 ohms, plate res 68k. This gives ~95 volts at
the plate. gNFB injection point at cathode.

Direct coupled to two halfs of 6SN7 paralleled,
cathode resistor 12k bypassed with 68uF, plate
resistor also 12k. B+ for this stage is 390V, too.

Is there anything objectionable with such an
arrangement?

You are considering a Loftin White style of driver,
and I don't see the merit.
All you save is one R&C to couple V1 to V2.
The 2 x 1/2 6SN7 in parallel needs only one RC in the cathode circuit.
They work reliably that way, but output stages should all be
separately cathode biased.

Since the cathode of V2 is at about +100v, because g1 is direct to V1
anode,
then its effective B+ is only 290v,
and its v swing output is limited, and
has higher thd than if B+ was 390v.

So I have never done a driver this way of yours.
Direct coupling has its place between V1 and V2 in a Williamson,
but I have never used it elsewhere.


Tom

--
Live is too short to be taken seriously.
- Oscar Wilde

But a short circuit could seriously threaten life,

Patrick Turner.

Hello Patrick,

You are considering a Loftin White style of
driver,

Indeed :-)

and I don't see the merit.
All you save is one R&C to couple V1 to V2.

Well, one cap less in the signal path surely
could be considered to be a merit by many folks,
be it only on a pure theoretical basis.

Personally, after having had my share of experience
with "theoretically excellent, but overly complicated"
PP driver arrangements, I am trying to keep things
KISS, especially in an SE design.

The 2 x 1/2 6SN7 in parallel needs only one
RC in the cathode circuit. They work reliably
that way, but output stages should all be
separately cathode biased.

I wholeheartly agree, main point being safety reasons.

Since the cathode of V2 is at about +100v, because
g1 is direct to V1 anode, then its effective B+ is
only 290v, and its v swing output is limited, and
has higher thd than if B+ was 390v.

Yes, I experienced that with a single 6SN7 section
driver stage in such a direct coupled arrangement
experimentally:

At a certain point (say, above 2 watts Po from the two
paralleled 6550 tetrodes - didn't actually measure the
driver voltage swing at that point) the soundstage
collapsed and the music degenerated to mud (okay,
vastly exaggerated, but you surely know what I mean).
But before that certain point, the sound was simply
amazing to me.

That's why I am considering two 6SN7 sections as
an enhanced 2nd stage driver stage.

Direct coupling has its place between V1 and V2
in a Williamson, but I have never used it elsewhere.

Exactly that's where I got the idea from, since my
last project was a Williamson style amp (807PP/UL,
up to ca. 10 watts class A).

Before this Loftin-White like driver arrangement, I
have tried a gain stage / driver circuit made up from
two consecutive mu-follower stages utilizing 6SN7,
but the Loftin-White like arrangement (single 6SN7
section driver version) described above beats that
hands down soundwise - but only for the first few
watts Po. AFAIR, these two consecutive mu-follower
stages (and a voltage devider between them to reduce
gain) needed more than 10 resistors and 6 or 8 more
caps (for a stereo amp) to work than that LW like
design, not to talk about the additional measures
needed PSU-wise to cope with the high heater-cathode
voltages of the "upper" tubes in that mu-follower
arrangement. (I could post the schematic if
interesting.)

So I tried to find a way to eleminate lots of parts,
thus the LW like design ...

Another option I still consider in spite of its
complexity including PSU measures to be taken (I have
not tried it on the breadboard yet) would be a single
"real" (A. Kimmel style) mu-stage using 1/2 6SL7 with
EF184/6EJ7 or even D3a/7721 frame grid pentodes
"on top". This might be on the edge gain-wise (assuming
only a few dBs of global loop NFB needed), but at least
it would be a single stage gain/driver arrangement for
the paralleled 6550 tetrode output tubes ...

Tom

--
Falling in love is a lot like dying.
You never get to do it enough to
become good at it.