hams on classiccmp
pat at computer-refuge.org
Sun Jan 18 08:57:58 CST 2009
> > On Saturday 17 January 2009, Ian King wrote:
> > > This machine was one of a pair bought from a collector who bought
> > > from a collector who bought from...? It was interesting noting
> > > the differences between the two: for instance, on one the black
> > > foam was in fairly good shape (replaced it anyway), but on the
> > > other it was falling off in chunks. Both seemed in good
> > > mechanical condition. The next step was to ensure electrical
> > > integrity.
> I wrote:
> > FWIW, if you can be sure you're going to operate the machine in a
> > relatively clean environment (eg, no carpeting or other things to
> > generate dust), I'd forgo the filter foam, as all it'll do is
> > crumble and block airflow.
> > > So I replaced every electrolytic capacitor of consequence in the
> > > switching power supplies - five in one machine, six in the other
> > > - plus the LSI-11 boot machine and RX01 boot floppy. This was
> > > expensive!
> > This also seems unnecessary if the capacitors were still good. The
> > 11/780 I got up and running has one bad psu (de-asserts DCOK every
> > once in a while, and causes the machine to reboot) which may be bad
> > capacitors, but everything else was ok, despite being stored for
> > over 10 years.
On Sunday 18 January 2009, Ian King wrote:
> Industry folks say that the lifetime of electrolytic capacitors is
> about fourteen years. Since the filter caps in our machines were
> significantly (i.e. TWICE) older than that (and some showed physical
> symptoms of degradation), we replaced them rather than deal with them
> one by one as they failed. In our experience with our PDP-10
> machines, this was (despite its cost) cheap insurance against
> periodic failures, which can occasionally be catastrophic.
That's an "average" lifetime, which doesn't take into account things
like the quality, brand, etc, of the capacitors in question. Yes, some
of them may fail, and replacing ones that are actually bad (ether
through physical or ESR-meter examination) is a good idea, but just
replacing them randomly seems like a bad idea. The replacements are
likely of a lower quality than the ones made 20+ years ago, and may
actually fail sooner than the ones in the machine.
Also, those power supplies aren't unobtanium, and blowing up a cap is
unlikely to kill the rest of the machine. If you're going to go
replacing caps, why stop there? Why not replace all of the tantalum or
electrolytic bypass caps on all of the boards, which are as likely to
have issues and cause a "catastrophic" failure.
> The foam on the doors not only serves to cut down on dust into the
> machine, it also cuts down on the noise in the machine room. That
> makes me happy. :-) Yes, the machines can be run without it. But
> since our goal is restoration, we'll replace it - and yes, replace it
> again in ten or fifteen years.
Noise abatement is a good thing, but I guess I'd be surprised if they
make a noticeable difference. I've never tried it.
> We *are* running the machine off three separate 110V single-phase
> circuits - apparently you missed that part.
I never said that you weren't. I was suggesting to anyone reading the
thread how one can figure it out. You said that you were doing this in
your original email.
> Despite your opinion
> that "it should be obvious", as good engineering practice I wanted to
> understand how the machine was using the input power before making
> assumptions. I'd rather spend some time in due diligence than in
> damage control - or fire fighting!
You seem to have missed the smiley at the end of that sentence.
> Also, since our interest is
> historical restoration, we wanted to make use of the PDU as wired
> rather than try to bypass it with some sort of approach that would
> feed 110V directly to the various elements.
Also, I didn't suggest that. The PSU provides important protection by
shutting down the machine if the fans stop working, or the machine
> > > After all of this, we carefully brought up the machine. We had a
> > > challenge because 120V three phase doesn't seem to be usual
> > > practice in US wiring - we had 240V three phase, but that
> > > obviously wasn't going to do us any good! Carefully looking
> > > through the power distribution unit's engineering drawings, it
> > > became clear that DEC used three-phase simply to balance the
> > > current load among the legs - in fact, everything runs on 120V.
> > > So we used equal care in reviewing the wiring of the warehouse
> > > where we keep these machines and found three outlets that were
> > > (a) on the same side of the 240V mains and (b) not sharing a
> > > breaker and circuit. Those were connected to a three-phase
> > > outlet, the VAXen were plugged in and voila! NOTE: we have a
> > > team member experienced with commercial power circuits. Don't
> > > try this at home - or if you do, be very very careful and be
> > > certain that, from any of the three live blades to another, you
> > > don't have more than 120V.
> > First, I want to point out that all of the power outlets on the PDU
> > are 120V, 20A (NEMA 5-20R) outlets, so it should be somewhat
> > obvious from that, that the machine doesn't need three phase power
> > to run, no engineering drawings required. :)
> > Also, your statements about three phase power aren't quite valid.
> > You may have had three-phase 240V, but that is unlikely, the
> > typical practice in the US for non-motor loads (lighting, general
> > power usage) is 120/208V three phase or 120/240V single phase. In
> > normal use, the three phase connections on the VAX PDU have 208V
> > between them, and 120V to ground, which is a standard 120/208V
> > three phase system.
> > You can run the machine off of single phase power by chosing up to
> > three separate 120V circuits - it's ok if they're on different
> > phases, having 240V phase-to-phase is OK - and running each one to
> > a different phase, and tying all the neutrals together to the
> > neutral in the machine. In fact, it is a good idea to make sure
> > that you have different phases, so that you minimize the neutral
> > current, otherwise you may end up with melted wiring or fire, and
> > selecting opposite phases will do this for you, as the netural
> > current from opposite phases will cancel either other out.
> > Fortunately, an 11/780 doesn't draw nearly the outlet/PDU rating; I
> > think I measured around 24A total draw at 120V from all three
> > phases on mine, and load didn't raise that too much. I've run it
> > from three 15A circuits (in a building with 120/208V power that I'm
> > not allowed to put in my own outlet for it ;), and it was ok, each
> > phase was around 8A max.
> You may be completely
> correct about how buildings are wired, but our building isn't wired
> that way. The EE on our team came up with an approach that's working
> for us.
I said that it's possible that it's 240V three phase, but most other
people aren't going to encounter that when plugging in the machine.
The most important part of what I mentioned, which you skipped over, is
that you should try to put the three legs of the vax PSU on at least
two different phases (two on one, one on the other), so that you
minimize the neutral current. Your electrical guy, having put them all
on the same phase, so that there isn't any delta-V between the hot
phases, may be helping your machine's PDU/power cord start a fire by
overloading the neutral conductor. In the best case, it'll just
nuisance trip the main breaker (I can't remeber off-hand if the neutral
conductor is run through the breaker or not). IME it's unlikely
because of the currents involved, but it is possible, and important to
be careful to avoid.
Also, "working for us" doesn't mean safe or correctly wired.. Not to go
doubting your EE friend whom I don't know, but it's possible to wire
things with in many ways that "work" but are unsafe (eg, it's possible
to wire your house using metal coat hangers, but not suggested :).
> That raises an interesting point I forgot to mention: it's important
> to determine that the various subsystems are properly plugged into
> the PDU. When we first tried to bring up the machine, it would pop
> the main breaker. After some head-scratching, we determined that
> this was because the fan for the memory subsystem wasn't plugged into
> an unswitched outlet. If that fan isn't running, there's a signal
> from an airflow sensor that trips the main breaker!
Yes, I had a similar problem when I had unplugged everything (including
the fans) from the PDU to check out parts without powering up
everything. The maintenance manual made it easy to track down what was
> > > Now that the machines would power up, I scoped all the power
> > > supply voltages to ensure they were really DC, i.e. that I hadn't
> > > missed an important filter cap anywhere. All good, so I tried
> > > booting from the floppies we got with the machines.
> > Fun. I'd suggest putting a scope (or even better a one-shot that
> > triggers when they go off) "DCOK" and "ACOK" outputs from the PSUs
> > as more important; checking voltages with a DMM is probably more
> > useful.
> > > I was able to get the basic
> > > console to boot - hooray! - but was unable to get the
> > > microdiagnostics to run. We had agreed that successful execution
> > > of the low-level diagnostics was a precursor to any attempt to
> > > install the OS, so this was a roadblock.
> > Your OS is probably the best diagnostic that you have. If it
> > doesn't boot, or crashes, there's something wrong, which you can
> > pick a precise diagnostic to examine the problem. Diagnostics
> > sometimes find problems that don't really exist, and miss things
> > that are show-stoppers.
> Yes, if the OS doesn't boot that means you have a problem - but what
> problem? DEC designed these diagnostics to tell you where that
> problem lies.
Right, if the OS doesn't boot, then you have a problem, and it's time to
pull out the diags. If it does boot, and appears to run ok, it's
unlikely that you have a problem. If diagnostics are hard to find (as
it was in your case), I'd much rather have a working machine than not
use a machine because it's hard to find diagnostics.
> I disagree with your opinion on scoping the power supply outputs. In
> my colleagues' work with the PDP-10s we have, they found that 'weird'
> problems were often caused by an unexpectedly unfiltered DC line -
> and DCOK didn't cop to it. I have clean power - under load - to this
VAXes are somewhat newer than PDP-10s and thus probably have smarter
PSUs. Scoping the power lines is a reasonable idea, but DCOK is more
likely to catch issues like what I had than staring at a trace on a
scope. I never said that scoping the power lines wasn't a good idea, I
just suggested that the DCOK/ACOK lines are a good idea to pay
attnetion to, as they'll actually cause the machine to reboot if they
> > > We had CDROM media for OpenVMS 6.2, the latest version certified
> > > for the VAX-11/785
> > 7.x works fine.
> > > Now we were cooking with gas. I had a valid VMS license but not
> > > a UCX license.
> > You should be able to use the hobbyist license, unless you're
> > planning on running the machine commercially.
> We have a valid 6.2 license, which is a 'period' OS for this machine.
> As to your assertion that we would be fine using the hobbyist
> license with 7.2: I'm not a lawyer, are you? I have used the 7.2
> hobbyist release and PAKs for a number of years on several of my own
> VAXen at home. But this is not 'at home' and I know my 6.2 license
> is legal and proper.
Are you planning on using the VAX for anything related to a business or
commercial enterprise? If not, you should be able to use the hobbyist
license. "At home" has nothing to do with it; you could be out of
license by using it for a home-based business, but fine with the
license using it for a computer club at a warehouse somewhere.
Anyways, I'd rather run something period (you can find at least VMS 3.x
by asking people here) or the latest version that'll run, unless you
have some historical interest in 6.2 specifically.
Pat (BSCompE 2004, EIT, need to get off my ass this year and take the PE
Purdue University Research Computing --- http://www.rcac.purdue.edu/
The Computer Refuge --- http://computer-refuge.org
More information about the cctalk