Xref: utzoo rec.humor:18837 rec.humor.d:1621 comp.misc:5086
Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!ames!hc!lanl!cmcl2!ccnysci!sukenick
From: sukenick@ccnysci.UUCP (SYG)
Newsgroups: rec.humor,rec.humor.d,comp.misc
Subject: Re: Looking for Computer Folklore
Message-ID: <1255@ccnysci.UUCP>
Date: 12 Feb 89 14:56:37 GMT
References: <7143@pyr.gatech.EDU> <532@geovision.UUCP> <4575@tekgvs.GVS.TEK.COM> <6255@saturn.ucsc.edu> <1232@raspail.UUCP> <9183@bloom-beacon.MIT.EDU>
Reply-To: sukenick@ccnysci.UUCP (SYG)
Organization: City College Of New York
Lines: 39

One of my firstjobs (in college) was to interface  a PDP-8 to  a spectrometer..
except that the PDP  didn't work and there was no service contract, etc.
Turning on the system revealed a strange look to the front panel lights,
which meant a blown fuse, which I changed and it blew again.
I pulled out the drawer with the bus & boards and noticed
a funny cloud above them which turned out to be some sort of fruit flies.
(that should have warned me off :-))  
Pulling the boards and shaking the tray with the bus upside down
got out lots of dust,small pieces of wire,insulation and a few dead
roaches.  It worked after the cleanout........

Soon the application was written and the machine was interfaced, and the
setup was getting lots of use from people doing their experiments and
storing data.  Every once in a while though, the entire sytem would
freeze: the lights would indicate the last  instruction/address and none
of the front panel switchs would work. Powering it down and up would sometimes
work, but not always. It always happened when the computer accessed the
clock with a particular instruction. DEC,usually helpful, said "run the
diagnostic" but the diagnostic froze at that instruction intermittantly
also thus revealing no new information.

The clock was sent out for repair; they could not figure out what was wrong
except "darndess thing; same thing happens here too!".  A different clock
board also did the same thing...  I found that shaking the entire bus
would always get the thing working. It was a lot of fun when they
"call in the computer expert"; I pull out the drawer, and while they expect
some sophisticated fix, I give it a good hard shake, and gently slide it in
and start it up.  This fix would always work.
After one time, the computer started to blow boards.  It seems that someone
saw me shaking the machine to fix, and decided to do it on his own,
except he didnt power down first....(He stopped using the system after
he was caught plugging  the BNC from the photo tube power supply
(a kilovolt or two) into the PDP's counter input.(it wasnt on at the time).

(I finally discovered that the problem with the closk was powersupply:
the traces were thin and drew enough current so that chips
on top of the  board were getting 4.6 volts: (minimum needed 4.75 TTL)
Tweaking the power supply up a little to ~5.1 V solved the problem and
it's been running fine for many years now.