Ferrari Mailing List: Re: old times flying in their warbirds

[Hans E. Hansen (FList]

If you can find that WIRED article, it is a good read.  I think it's
available online.

The computer was sort of committee designed, but MIT had much to do
with it.  A leading computer science professor designed a rather
sophisticated multi-tasking operating system for it, which was
remarkable as it arguably worked as well as multitasking on
mainframes.  It prioritized operations, and knew what was using how
much resources.  Which is where it gets interesting.  The guidance
program that Neil Armstrong was using at the critical time was using
87% of the machine's capacity.  Aldrin had turned on a program that
was using radar to track the command module. That switch had 2
settings, one of which used more of the computer's resources.  It
wasn't needed right then, and he wasn't supposed to be using that
setting.  It used 13% of resources.  So the system was maxed out, but
operating.  However - and  this was the big problem, there was a
process that would run automatically every 30 seconds or so, using 2%
of resources that was also related to the incorrect switch setting.
That's what overloaded the computer.

The brilliant MIT professor had designed into the op system a "gentle
crash", I guess you could call it.  When overloaded, it didn't
completely crash dead, but would store critical data and then stop
with some sort of error message.  I think it displayed "1202", but the
article is not at hand, so that might not have been the code.  In
pre-flight stress testing, the academics frequently forced scenarios
on the computer that it could not handle, just to see what would
happen.  So in testing, the "1202" was seen frequently.  However, few
if anybody at NASA knew what it meant, and Aldrin and Armstrong were
not trained on it.  A simple hard reboot was all that was needed to
get it running again.  The recovery from the failure was very
predictable and reliable.  Armstrong rebooted the computer several
times, and it worked as predicted until that internal process that was
the electron on the camel's back that caused things to come tumbling
down.

Once on the lunar surface, NASA investigated what was wrong.  That's
when they noticed that some telemetry showed a switch setting on a
radar unit was set wrong, feeding too much info into the computer.
Turn the switch, all was good.

I'm curious if the hardware was upgraded for Apollo 12.  A quick
search wasn't revealing.

Hans.

On 8/22/19, Hans E. Hansen <FList [at] hanshansen.org> wrote:
> WIRED had an interesting read a month or so ago about the 3 W's (What
> Went Wong) of the lunar guidance computer.  Turned out it was operator
> error - Aldrin had a switch that hooked up a radar to the computer
> turned the wrong way, overloading it with info.  For the return
> reunion with the command module they turned it back to the correct
> position.  But it was panic and mayhem until they figured out what was
> happening.
>
> 5600 discreet NOR gates manufactured by Fairchild made up the computer.
>
> Hans.
>
> On 8/21/19, Erik Nielsen <judge4re [at] gmail.com> wrote:
>> The average age of the engineers on Apollo was 26...
>>
>
>>> Sponsored by BooyahMedia.com
>>> and F1 Headlines
>>> http://www.F1Headlines.com/
>>
>