When Knuth and I were young - Part 3
- From: Charles Crayne <spamtrap@xxxxxxxxxx>
- Date: Mon, 21 May 2007 23:44:16 -0700
The sandstorm in New Mexico may have been a minor factor in my turning
down the job offer from Sandia, but the offer from NCR seemed to me to
be the better choice on other grounds, as well. It was now 1965, and
the state of the art in computer technology had moved from vacuum tubes
to discrete transistors, and the new frontier was integrated circuit
logic gates. As a result, the new computers were smaller and faster,
but still very expensive.
In addition, IBM had recently announced a new line of computers with
an architecture which combined the features of both their commercial and
their scientific machines into a single family. In theory, every
machine in the family, regardless of processing power, was
binary compatible with every other machine in the family. [In
practice, however, not all features were available on all models.]
At the time, this was a revolutionary concept, and NCR felt pressured
to do something similar. Their response was the Century series, and my
primary responsibility was to draw the logic diagrams for the cpu
portion of the entry level version, the Century-100. Of course, I had
no say in the functional specification, which had already been written
by a separate group of system architects.
However, I also helped in the testing of the engineering prototype for
the first model to be released, which had already been designed by
the time I was hired, and which was being constructed in our department
by an ex-navy electronics chief, and a young woman borrowed from the
manufacturing floor.
In those days, the wiring which connected the various ICs together was
attached, not by soldering, but by wrapping the stripped ends of each
wire around pins which stuck out the back side of the circuit boards.
In the production department, this repetitive task was done by robot
arms, which I found fascinating to watch. But, the wires could also be
manually attached by a person with a "wire-wrap gun", which, on our
prototype, was the young lady's task.
After the box passed its power-up tests, but long before there was any
software to run on it. I put together a small set of diagnostic tests
(in machine language, of course) and promptly found a problem, the
cause of which turned out to be that she had wrapped a wire to the
wrong pin.
The problem was quickly corrected, of course, but, in later years,
during my IBM career, whenever some arrogant programmer complained to me
that since there was nothing wrong with their program, the computer
must be broken, and after I had shown them that the error was indeed in
their code, I always thought back to that time when the machine really
was wired wrong.
The next major test was to hook up a line printer to our prototype
machine, and so we rolled our box down the hall to a corner of the
manufacturing floor, where one of the production engineers had set up a
printer which had just come of the production line. We all stood around
while he made the connections and turned on the power switch.
About this time, the electronics chief nudged us in the ribs and
pointed to a spot on the printer frame, where he had grounded one lead
from an electrolytic capacitor, the other end of which he had connected
to the printers AC power supply. As we watched, the capacitor started
to glow, and, after a few seconds, exploded loudly.
The poor production engineer made a flying leap for the master power
switch, and then -- visibly shaken -- started talking to himself about
what might have gone wrong. At this point, the chief took pity on him,
and tried to explain that it was just a joke, but the poor guy was too
shocked to understand. Even when the chief showed him the evidence, all
he could think about doing was to go look at another printer, to see if
it too had a capacitor attached to the frame with an alligator clip.
After that, it wasn't funny anymore, but we did finally manage to get
him calmed down, and nothing more was said about it.
In the next installment, I will explain why I decided to get out of the
logic design business.
-- Chuck
.
- Follow-Ups:
- Re: When Knuth and I were young - Part 3
- From: idle
- Re: When Knuth and I were young - Part 3
- Prev by Date: Re: When Knuth and I were young - Part 1
- Next by Date: Re: generating a.out format file with gcc
- Previous by thread: When Knuth and I were young - Part 2
- Next by thread: Re: When Knuth and I were young - Part 3
- Index(es):
Relevant Pages
|
