Message from discussion German conference highlights doubts about ESA's manned space plans

Path: gmdzi!unido!mcsun!uunet!cs.utexas.edu!swrinde!zaphod.mps.ohio-state.edu!think.com!linus!philabs!ttidca!quad1!bohica!mcws!p0.f851.n102.z1.fidonet.org!Kieran.A..Carroll
From: Kieran.A..Carr...@p0.f851.n102.z1.fidonet.org (Kieran A. Carroll)
Newsgroups: sci.space
Subject: Re: German conference highlights doubts about ESA's manned space plans
Message-ID: <215.27E4ABEA@mcws.fidonet.org>
Date: 12 Mar 91 00:56:00 GMT
Sender: ufg...@mcws.fidonet.org (newsout1.26)
Organization: FidoNet node 1:102/851.0 - Central Computer Ba, Tujunga CA
Lines: 58

From: kcarr...@zoo.toronto.edu (Kieran A. Carroll)
Path: wciu!elroy.jpl.nasa.gov!usc!rutgers!news-server.csri.toronto.edu!utzoo!kcarroll
Newsgroups: sci.space
Subject: Re: German conference highlights doubts about ESA's manned space plans
Message-ID: <1991Mar11.165626.25181@zoo.toronto.edu>
Date: 11 Mar 91 16:56:26 GMT
Nick Szabo sz...@crg5.UUCP write:
Subject: Re: German conference highlights doubts about ESA's manned space plans
Date: 8 Mar 91 01:53:56 GMT

> In article <1991Mar7.172412.17...@zoo.toronto.edu> 
> he...@zoo.toronto.edu (Henry Spencer) writes:
> 
> >Experience so far says very strongly that if you plan to do in-space repairs,
> >you had better plan to have humans (or the imaginary just-as-flexible
> >teleoperated robots which will be available almost right away now for sure,
> >really, trust us) on hand.
> 
> Or, if you prefer not to pay the $1,000,000++/hour cost of an EVA, you 
> could make sure it works before you launch it. 

Nick, to ``make sure'' that a satellite ``works'' before you launch it,
you might have to spend anywhere from millions to tens of millions of
dollars, depending on the complexity of the spacecraft. What level of
reliability do you want to design into the satellite, and verify via
ground testing? 99%? Or three nines? Or four? Do you want dual-string
redundancy? Or are you going to make your designers hate you forever,
and insist on triple-string?

All those satellites that we've seen fail on orbit, including the ones that
the Shuttle was sent up to repair, were designed by people who tried
to ``make sure that it works'' befoire launch. Many of the people involved 
in design of these are very smart, and very conscientious, probably more so
than you and I. Satellites don't typically fail because of sloppy design or 
construction. They fail because >nobody< can imagine beforehand >all< the things 
that might go wrong with a complicated, custom-built device.

My view is the cost of using spacecraft could be made cheaper if moderately-priced
on-orbit check-out and servicing facilities were available to be rented out.
That way, a great deal of the expense in designing spacecraft could be eliminated
(``the NASA way'' of design wouldn't have toof the expenses of ground-testing the beasts. However, this assumes that
you have >people< in orbit to do the testing and repair --- because the things that
will go wrong with your satellite will be exactly those that you didn't anticipate,
and so you won't have built your teleoperated robot to handle that
unexpected contingency, and so the robot won't be able to effect the repair.


-- 

     Kieran A. Carroll @ U of Toronto Aerospace Institute
     uunet!attcan!utzoo!kcarroll kcarr...@zoo.toronto.edu


--  
: Kieran A. Carroll - via FidoNet node 1:102/851 (818)352-2993
: ARPA/INTERNET: Kieran.A..Carr...@p0.f851.n102.z1.fidonet.org
: UUCP: ...!{elroy!bohica,elroy!wciu,cit-vax!wciu}!mcws!851.0!Kieran.A..Carroll
: Compu$erve: >internet:Kieran.A..Carr...@p0.f851.n102.z1.fidonet.org