Maker Pro
Maker Pro

BEYOND Push Pull Parallel (Tubes)

Just curious. Many old tube amplifiers, especially guitar and PA amps
had four audio output tubes, such as four 6L6 tubes. Was there ever
an amplifier designed with MORE tubes for the output stage? For
example, using EIGHT 6L6 tubes? I'm not sure what that would be
called? (push pull parallel xxxxxx)?
 
L

Les Cargill

Jan 1, 1970
0
Just curious. Many old tube amplifiers, especially guitar and PA amps
had four audio output tubes, such as four 6L6 tubes. Was there ever
an amplifier designed with MORE tubes for the output stage? For
example, using EIGHT 6L6 tubes? I'm not sure what that would be
called? (push pull parallel xxxxxx)?


Yes. The Ampeg SVT and the Fender 400 both use six. They're just
push-pull ( Class AB ) with more tubes in parallel.
 
Just curious. Many old tube amplifiers, especially guitar and PA amps
had four audio output tubes, such as four 6L6 tubes. Was there ever
an amplifier designed with MORE tubes for the output stage? For
example, using EIGHT 6L6 tubes? I'm not sure what that would be
called? (push pull parallel xxxxxx)?

Four (2x2) or six (2x3) EL34 pentodes was pretty standard way in push
pull configuration for generating the "magical" 100 W in guitar
amplifiers with 400-450 V anode voltage.

I have also seen schematic diagrams, but never worked with amplifiers
using six power triodes in a totem pole configuration driving a 16 ohm
loudspeaker without an output transformer. The layout was quite
similar to current NPN/PNP complementary (emitter follower) stage,
with three triodes in parallel in the same position as the upper NPN
emitter follower and three triodes in parallel in the same position
which is currently occupied by the PNP emitter follower. Triodes have
lower internal resistance than pentodes and putting several triodes in
parallel further reduced the resistance and thus, the amplifier could
drive "standard" 16 ohm loudspeakers directly.

While two EL86 pentodes could be used in totem pole configuration
without output transformers, special loudspeakers were required with
an impedance about 1000 ohms.
 
A

Adrian Tuddenham

Jan 1, 1970
0
Just curious. Many old tube amplifiers, especially guitar and PA amps
had four audio output tubes, such as four 6L6 tubes. Was there ever
an amplifier designed with MORE tubes for the output stage? For
example, using EIGHT 6L6 tubes? I'm not sure what that would be
called? (push pull parallel xxxxxx)?

The 1946 BBC 'Disk Recording Equipment Type D' used six AL/60 valves in
parallel push-pull in the recording amplifier (Type DRA/4). With an HT
of 365V, this gave the 75 to 100VA needed to drive the inductive
cutterhead with the sufficient current for full low frequency modulation
and the enough voltage to take the response up to 10 Kc/s with recording
characteristic boost and radius compensation.

In the late 1950s, Wireless World published a design which used about
eight paralleled pentodes each side. The layout was a strange
breadboard affair with bakelite surface-mounting valveholders.
 
D

DonMack

Jan 1, 1970
0
wrote in message
Just curious. Many old tube amplifiers, especially guitar and PA amps
had four audio output tubes, such as four 6L6 tubes. Was there ever
an amplifier designed with MORE tubes for the output stage? For
example, using EIGHT 6L6 tubes? I'm not sure what that would be
called? (push pull parallel xxxxxx)?
 
A

Adrian Tuddenham

Jan 1, 1970
0
Michael A. Terrell said:
RCA used 16 parallel 6146 as the video amplifier to modulate a 4CX250
in their TTU-25 series TV transmitters.

It's sometimes more economical to use multiple small devices, rather
than have to find some single big device to do the job.

With semiconductors, it can often be advantageous to use several cheap
transistors and run them gently, rather than push a big expensive one to
its limits. If there is a lot of dissipation, the thermal paths from
multiple chips to the heatsink are parallelled, which allows a smaller
heatsink that can safely be run hotter.

In the case of valves, there was often a big jump in price between a
'consumer' output pentode and the next size up, which fell into the
'industrial' class. For guitar amplifiers, which were, in essence,
'consumer' equipment, it made sense to stick to consumer valves, even if
it meant using more of them.

It also made the power supplies cheaper because bigger valves usually
needed higher voltages, which meant special rectifiers and paper
smoothing capacitors. A smoothing system with paper capacitors worked
out more economical if the capacitors could be relatively small and the
extra smoothing was provided by chokes. The chokes made the equipment
bigger and heavier- and the higher voltages made it very dangerous for
semi-skilled 'roadies' to service (and more spectacular when it went
wrong).
 
It's sometimes more economical to use multiple small devices, rather
than have to find some single big device to do the job.

Yes, but this is an odd one. A 6146 should be good for 100W, or so, IIRC. A
4CX250 is 250W, no? 16x100 >> 250. Seems like a bit of overdrive. ;-)
 
wrote in message
Just curious. Many old tube amplifiers, especially guitar and PA amps
had four audio output tubes, such as four 6L6 tubes. Was there ever
an amplifier designed with MORE tubes for the output stage? For
example, using EIGHT 6L6 tubes? I'm not sure what that would be
called? (push pull parallel xxxxxx)?

I located a schematic on the web for a Fender 400 with 6 output tubes.
But 12, that's insane..... Damn, I got to find a schematic for this
beast. Any idea of the model number?

Thanks
 
A

Adrian Tuddenham

Jan 1, 1970
0
Michael A. Terrell said:
It was a distributed amplifer, and flat from DC to 5 MHz.

[Nit pick] It is a debateable point whether a distributed amplifier
truly counts as parallel - but D.C. to 5 Mc/s at that power level is
quite an achievement.

I've always been in awe of distributed amplifier design. The theory is
straightforward enough, but what happens if you build one and it goes
unstable? Where do you begin to sort it out?
 
B

Baron

Jan 1, 1970
0
Les Cargill Inscribed thus:
Yes. The Ampeg SVT and the Fender 400 both use six. They're just
push-pull ( Class AB ) with more tubes in parallel.

I belive some of the old "Trix" cinema amplifiers used push/pull
triples.
 
They are, with a network between each tube. What I thought was
interesting was the output tubes. Water cooled power tetrodes at 12.5
or 25 KW across the entire UHF TV band.

How does one water cool a tube without the glass shattering?
 
S

Sjouke Burry

Jan 1, 1970
0
How does one water cool a tube without the glass shattering?
Use metal and ceramics?
And make the anode (almost)ground, using the coils as
water feed and make the cathode negative supply voltage.

That's just one solution.
 
F

Fred Abse

Jan 1, 1970
0
I've always been in awe of distributed amplifier design. The theory is
straightforward enough, but what happens if you build one and it goes
unstable? Where do you begin to sort it out?

Down at the pub?
 
A

Adrian Tuddenham

Jan 1, 1970
0
Michael A. Terrell said:
Adrian said:
Michael A. Terrell said:
:

Yes, but this is an odd one. A 6146 should be good for 100W, or so,
IIRC .A
4CX250 is 250W, no? 16x100 >> 250. Seems like a bit of overdrive. ;-)


It was a distributed amplifer, and flat from DC to 5 MHz.

[Nit pick] It is a debateable point whether a distributed amplifier
truly counts as parallel - but D.C. to 5 Mc/s at that power level is
quite an achievement.


They are, with a network between each tube. What I thought was
interesting was the output tubes. Water cooled power tetrodes at 12.5
or 25 KW across the entire UHF TV band.

I've always been in awe of distributed amplifier design. The theory is
straightforward enough, but what happens if you build one and it goes
unstable? Where do you begin to sort it out?


One tube at a time. it would be easier to troubleshoot, than trying
to match the 16 6146 that are needed for proper operation.

The trouble with that method would be that if you unplugged most of the
valves, the capacitance at each step of the line would change, so the
line would no longer match the terminating resistor. Then it would
develop standing waves and add to the problem.
 
Adrian said:
:

Yes, but this is an odd one. A 6146 should be good for 100W, or so,
IIRC .A
4CX250 is 250W, no? 16x100 >> 250. Seems like a bit of overdrive. ;-)


It was a distributed amplifer, and flat from DC to 5 MHz.

[Nit pick] It is a debateable point whether a distributed amplifier
truly counts as parallel - but D.C. to 5 Mc/s at that power level is
quite an achievement.


They are, with a network between each tube. What I thought was
interesting was the output tubes. Water cooled power tetrodes at 12.5
or 25 KW across the entire UHF TV band.

I've always been in awe of distributed amplifier design. The theory is
straightforward enough, but what happens if you build one and it goes
unstable? Where do you begin to sort it out?


One tube at a time. it would be easier to troubleshoot, than trying
to match the 16 6146 that are needed for proper operation.

The trouble with that method would be that if you unplugged most of the
valves, the capacitance at each step of the line would change, so the
line would no longer match the terminating resistor. Then it would
develop standing waves and add to the problem.[/QUOTE]

Are we talking about a baseband (0..6 MHz) signal or UHF channel ?

For a single channel UHF operation, why not use some Wilkinson
dividers to combine the output power from multiple tubes/transistors ?
If one module failed, no big issue.
 
Top