6502 chip

Greg Hermann bearbvd at sni.net
Fri Oct 2 03:13:26 GMT 1998


>It was not really something that would be useable for EFI, especially since
>the chips that made up the cpu board proabably cost $1+ each, and there
>were several hundred, maybe into the low thousands that made up what would
>probably have been an equivalent to the <1Mhz processors that easily fit
>on a single chip now.
>
>                                Roger
>
>On Thu, 1 Oct 1998, Mike Dillon wrote:
>
>>
>> I am not sure but I think the only "chips" commonly around in the
>> late 60's were DTL, I think TTL was early 70's and therefore would
>> not have been avalible for the Apollo program.

I dunno much if anything about the processors involved, but can tell you
that shuttle technology, (dating back about 13 years, may well have changed
since) in order to get the level of reliability deemed suitable, was as
follows:

For any given set of control functions a total of five identical
controllers (sensors and all, I believe) were used (all self-diagnostic, of
course). Three were functional, two were back-ups. The three functional
ones voted on any given control function. If the three were not unanimous
on a control action,  (despite a lack of any self-diagnostic admission of a
problem) the odd system got thrown out, and replaced with one of the two
back-ups, then unanimity was sought again, etc.

This info was extracted (gently) from the brains of a couple of Otis
Elevator guys with whom I had the pleasure of serving on a code-writing
committee for part of a ski-lift and tramway code (ANSI B-77). Their
general reaction to what the early fifties vintage tramway guys thought was
"Fail-Safe" was laughter (actually guffaws), followed up with a description
of how NASA was currently doing things for those of us who wanted to
listen.

Perhaps there are lessons in this kind of approach for grasshoppers,
Princes of Darkness, and other such critters.

Regards, Greg





More information about the Diy_efi mailing list