Subject: Re: VAXen at home
From: "Ethan Dicks" <ethan.dicks at gmail.com>
Date: Tue, 16 Oct 2007 17:08:42 -0400
To: "General Discussion: On-Topic and Off-Topic Posts" <cctalk at
classiccmp.org>
On 10/16/07, William Donzelli <wdonzelli at gmail.com> wrote:
I agree
that the custom gate arrays make the 11/750 unrepairable
(except for raiding parts), but I still think it's a good compromise
between size and performance, for the VAX-11 line.
Has anyone here actually ever come across an 11/750 failure caused by
one of the gate arrays? They are remarkably reliable machines.
They are remarkably reliable. I think I've seen one or two gate array
failures in the 10-15 years that I was around 11/750s on a regular
basis (and none lately, but that's a sample size of effectively zero).
Even so, that's two out of dozens of problems that were in less
esoteric arenas (like bad power supplies and such).
Most common problems were power supplies and perpieral related.
Serial lines drivers and recievers were common failures as well
but in the mill we usualy had long lines and occasional ground
faults to deal with.
If the machine is kept reasonably cool it's a rock.
Within DEC 750s were very common and after they were less in vogue
for actual use they serves lat, mop and routing service as they
required near zero attention if any at all.
I have a bigger sample size to look at from being inside DEC.
the gate arrays are as relaible or more so than any other chip
and the likelyhood of failure is highest for chips that talk
off the board this being due to electrical stress (drivers)
and ESD from handling the board. Theother group of possible
failures is in ram arrays as there are so many it's statisiclly
possible to see a part fail.
Allison
Neither
are _easy_ to find, but I think it's
more likely to find a dusty 11/750 in a corner of a warehouse than an
11/780.
One will find ten /750s for every /780 these days.
I am entirely willing to accept that ratio (and since I have two
11/750s... that means that somewhere in my pile should be 1/5 of a
11/780, right ;-)
-ethan