It's time to restore the 11/45.

[tony duell]

> 
> 4) This would also be a good time to broach the subject of whether or not
> the capacitors need special love, or if that just adds more risk of
> destroying something.  Those familiar, please chime in as the archive is

My 11/45 which was working last year until I moved house (I haven't reassembled
it yet, or attempted to undo any damage caused by the movers) still has all the 
original capacitors in the PSU. 

I find this 'witch hunt' of capacitors curious. Yes, they do fail, but they are not the
only, or even most comon, thing to so so.

BUT!

There is a capacitor in the front section of the H742 PSU -- not in the regulator
'bricks' that, if it fails, causes ACLO and DCLO to pulse at power line frequency (or
probsbly twice that frequency). This really gives the 11/45 headaches, the microcode 
keeps on trying to handle the power failure and won't let you do anything else. So while
I would not replace any capacitors 'on spec' I would check that the ACLO and DCLO lines
are doing the right things.


> currently missing so I can't refer to the recent threads on this subject.
> Datapoint: I didn't mess at all with power supply / caps on my 11/34, which
> was stored right beside the '45 for the same timespan and it's running
> great now.  Datapoint 2: with my amateur electronics skills / status,
> there's admittedly some risk just having me poking around in there,
> "learning" things :-\   Plus, I'm afraid of death from esoteric high
> voltages that I hear are present in these old power supplies.  Is the 11/45
> one of them?  In a nutshell, this analog / power stuff is FAR FROM my

The 11/45 PSU is fairly friendly. There is, of course, mains inside, but there is a big
(looks to be about 1000VA) transformer in the front section, giving out lots of 
30V AC outputs. The regulator bricks are fed from 30V, so there are no lethal voltages
inside. But given that it's 30V at 10's of amps, there is enough power there to do damage. 
I have had transistors blown off the PCB when I had a nasty failure in one of the regulator
units.

> forte; some hand-holding here would be most appreciated.

I still can't comprehend how anyone can properly understand digital electronics and not
understand anlogue. But anyway..

The 11/45  power supply is a bit odd. Firstly there are multiple 5V rails, the regulators
could supply 25A each, so IIRC there are 2 for the CPU, one for the FPU (which I think you
have), one for the Unibus backplanes, etc. The upper PSU is supposed to be controlled by
the console switch, it powers the CPU and peripherals. The lower run is supposed to run
all the time, it powers the memory. No, it is not a problem if the lower PSU has empty slots,
these were for regulators for the MOS or bipolar memory in the CPU backplane. Don't
assume something has been taken out.

Another nasty... The front section of the PSU supplies ACLO, DCLO, LTC (line time clock) and
+15V. Or at least the upper PSU does. On the lower one, the +15V line is in fact grounded so
what is called 'ground' on that PSU becomes -15V. You need to look carefully at all the diagrams
in the printset.

The PSU should be checked with dummy loads (6V bulbs, for example), bearing in mind that with
several 5V lines you need to check all of them.

-tony