Had an Allen & Heath PA20-CP 2 x 500 watt powered mixer in today. Has two
completely separate and identical power amps, one of which was running very
hot very quickly. No schematics, but problem was quickly traced to an open
circuit resistor, by comparison ohms checks between the two amps. In order
to replace the resistor, it was necessary to remove the offending ("B")
power amp to get to the board underside. The power rails and ribbon from the
mixer desk section were on removable connectors, but the 4 pole speakon
connector was bolted to the rear panel and the wires were soldered directly
to the board. When I came to unsolder these wires from the socket, I noticed
that the "B" channel one was wired the opposite way round to the "A" channel
one. I noted this down to check on later.

When the amp was tested ok, I bolted it back in, and reconnected the speakon
socket the same as it was when I unsoldered it. The wires lay exactly where
they originally were, and looked absolutely factory-original, even down to
the one which went to the uppermost connector tag, having a rubber sleeve on
it, exactly the same as on the "A" channel. Except the two channels were
definitely wired arse-about-face to one another. On both channels' speakons,
the "1-" and "2-" pins were linked, as were "1+" and "2+". But on channel
"A" the ground wire was connected to the "+" pair, and on the "B" channel,
the ground was on the "-" pair.

I fired it up again, and put a sine wave in, then hooked up two channels of
my 'scope, one to the "A" output and one to the "B", both with the same
polarity, and was surprised to find that the two signals were completely
anti-phase, which meant that with the reversed connections to the speakon
sockets, they would be back in phase again to the outside world.

I then went back to the power amps' front ends near where the ribbon
connector was, and found the same thing. At the same point on each amp, the
signals were in opposite phase to one another. I then spotted three little 3
pin header blocks, with one pin pair on each designated "B", and the other
pair "A" and a little 1 or maybe an I next to the "A" in a sort of 'to the
power of' position. On the B amp, these three jumpers were set to "B", and
on the A amp they were set to "A", so clearly, these jumpers reverse the
phase of the signal coming in from the desk section. The two amps then
process the signals in antiphase to one another right to the final output,
where the 'correct' phase relationship is again restored, by wiring the
speakon connector 'wrongly'.

I have puzzled over this, and can think of no good reason for doing it that
way. I did wonder at one point if it was anything to do with being able to
make the amps bridgeable, but I downloaded a copy of the user manual, and
there is a dire warning paragraph towards the end, specifically saying that
under no circumstances should any attempt be made to operate the amps in
bridge. It also warns against making any connections between the two amps or
any speakers connected to them, although in actual fact, the 'ground' side
of each output is truly ground, and is common to both amps, the whole (dual)
power supply, and all metalwork.

What am I missing here ? Not of any real consequence, as the unit is mended
and back in its original condition, but for academic interest, it would be
good to understand the designer's thinking.

Arfa