Re: [Fwd: starship-design: HIGHLY OPTIMIZED TOLERANCE]

[KellySt@aol.com]

In a message dated 3/13/00 11:48:52 PM, stevev@efn.org writes:

>KellySt@aol.com writes:
> > Hate to tell you, but the Star Trek like tech would be to crude for
>a couple 
> > decades from now, and a IC chip is a IC chip.  So using and old CPU
>is not 
> > going to bye you anything.  
>
>When it comes to space applications, not all ICs are equivalent.
>There's a reason spacecraft don't use off-the-shelf Pentiums or
>whatever; most processors designed for terrestrial use aren't
>radiation-resistant and aren't capable of enduring thermal extremes, as
>well as being overkill for many control applications.  Galileo, launched
>in the 80s, uses 1802 microprocessors from the early 70s, and so do many
>other probes of its generation.  Even with the 1802's CMOS design,
>Galileo is experiencing frequent computer resets whenever it makes close
>passes to Jupiter and its radiation belts.  The Hubble Space Telescope's
>recent computer upgrade got it up to using space-certified 486s.  You
>should note that the Mars Pathfinder lander, which used a relatively
>modern IBM RS/6000 processor, failed after 90 days, most likely due to
>daily thermal cycling between cold and really damn cold.  The rover used
>a CMOS 8085 that probably lasted quite a bit longer.

Its not that the newer chips can't be space rated, just no one fliped the 
bill to do it.

Oh, last I heard the pathfinder died when the bats went.  It was the '90's 
and long survival nuclear power cells were politically incorrect.  So when 
the photi voltaics could deliver enough charge to get it through the night....

I'm asuming yuour talking about the lander from a few years ago, not the one 
that crashed a few months back?  No CPU issues there.


Kelly