Hi All,

ALL 417A/5842 POWER AMP (15 tubes)

This power amplifier uses a very simple topology, although the parallel
output tubes make things look complex. The topology is:
1. phase inverter is the first stage (split load)
2. SRPP drivers for each phase (4 tubes) directly coupling,
3. Push Pull cathode follower output stage (5 per side = 10 tubes)

I used "SUPPLY3" to power the unit. HT voltage was 310 volts (don't
worry, the 417's don't have more than 170 volts on them).

Measured Specs:
Sensitivity: 1.4V for 1 watt output.
Maximum PO before clipping: 10 watts.
Frequency response: .2 dB from 9Hz to 28kHz.
THD = 0.3% at 1W, about 1.2% at 10W.
Feedback: Local cathode degeneration only.
Output impedance: 0.22 ohms on the 8 ohm tap!

The quality of the OPT really limits the specs in this unit. For
test purposes, I used a Hammond 1650P, using the screen taps for
the primary (plate taps not connected). The topology is extremely
stable, even with the VERY HI transconductance of 10 417As in the
final. With some subtle circuit mods and a regulated 300-330 volt
supply, I'd guess you could squeeze 20 watts from this thing. The
frequency response at the upper end is *strictly* controlled by the
OPT characteristics, and the bottom end is *strongly* controlled by
the OPT characteristics.

Listening tests: Very quiet, accurate sounding. Bass is *very* tight
as you'd guess from the very high damping factor. Reasonably musical.
Distortion is primarily 2nd with a little 3rd.

If you listen closely to the 417's you can slightly hear the audio!

Circuit:
There are really only 3 stages to this amplifier, even though the
output stage is a cathode follower.This is due to the very high
transconductance of the devices used. The input stage is a standard
split load phase splitter. The two outputs are fed into separate
drivers. The drivers are SRPP connected and are the only "gain"
stage in this amplifier. The SRPP stages directly couple into the
cathode follower. This keeps the signal path really pretty simple.

The SRPP stage is *slightly* unusual in that the input grid is raised
a couple of volts above ground. This provides 2 benefits... the
cathode resistor can be raised a little more for the same "bias"
(providing a little more local feedback and lowering stage distortion),
and it provides a convenient way of adjusting one od the SRPP stages
so that the "quiescent" voltage is the same for each half (DC balance).

The second adjustment is for AC balance, and is effected by adjusting
the AC impedance in the plate lead of the phase splitter.

The "common" point of the output transformer center tap is connected
to ground through a bypassed resistor. The value of this resistor sets
the idle current of this amplifier. I set it for about 140 mA, which
idles each of the 417's at 14mA. There are probably several other
ways of doing this, but this method is very stable and repeatable.
It is, however, not very efficient. There's almost 20 watts dissipated
in the resistor.

With a VERY low driving point impedance, the transformer is kept
very much in control, so the distortion produced by the transformer
is much lower than you would expect. This also produces a very low
output impedance (high damping factor) even with no global feedback.

The individual small valued resistors in the cathode followers tend
to equalize the currents in the paralleled tubes, so strict matching
is not required. These also aid in overall stability, which otherwise
might be a problem with so many paralleled high gm tubes.

I did rewire the output stage about 3 times to confirm freedom from
oscillation. All 3 rewire jobs worked well, with no signs of
instability. I do always use the 4 separate grid "stopper" resistors,
connecting the resistors DIRECTLY at the grids (about 1/4 inch lead
length, all of which is consumed at the socket.) These are then
wired together "above" the tube socket, which then is used for the
actual grid connection, as implied by the schematic. Constructors note:
wire the cathode connection before you do this, or you'll never be
able to get to it.

The H-K breakdown voltage of the 417As is not spectacular, and the
output cathode followers and "top" of the SRPP's are referenced to
about 150 volts (see SUPPLY3 schematic). The other 3 tubes' heaters
are referenced to ground, as there is no significant cathode voltage
on them. (45-50 volts on the phase splitter).

Distortion performance could probably be improved by using a different
phase splitter tube, but then, it wouldn't be an "all 417" amp.

Steve