Index Home About Blog
Newsgroups: sci.space.history
From: Henry Spencer <henry@zoo.toronto.edu>
Subject: Re: 'real' spin-offs from space
Date: Mon, 30 Mar 1998 02:48:54 GMT

In article <351EE756.5955@ozemail.com.au>, IanD  <iand@ozemail.com.au> wrote:
>We've heard tell that the space program gave us teflon frypans and
>microelectronics, and I'm led to believe neither of these is actually
>true (teflon existed beforehand, and solid state electronics was
>following its own path anyway, and had other things propelling it
>other than space).

Actually, the microelectronics case does have some limited validity,
contrary to recent folklore.  NASA -- or more precisely, the MIT
Instrumentation Lab (now Draper Labs), NASA's contractor for the Apollo
navigation system -- was the first big *customer* for ICs.

At the time, various microelectronics technologies were being pursued --
ICs being only one of them -- and there was no clear winner and nobody
with a commitment to buying any of them in bulk.  When everybody else was
still just experimenting, MIT-IL was placing large orders for standardized
ICs, and working with the suppliers to get performance up and failure rate
down.  At the time they were finished, the first few prototypes of the
Apollo computers contained about 2/3 of all the ICs in the world.

The total volume of military IC orders did not exceed MIL-IL's orders for
several years.  Civilian orders were minimal until considerably later; the
early logic ICs had poorer performance than transistor circuits, and they
required a wrenching change of design philosophy, so their early adopters
were all people who had a desperate need for miniaturization.

>But I was wondering, can anyone suggest things that were developed only
>for the space program to solve specific problems within the space
>program, but which have now found deployment in everyday life...

There've been a great many little odds and ends; possibly the biggest
single area was medical instrumentation.

But most people consider the question a rather pointless one.  Spinoff
doesn't win the space program any brownie points with the average
non-techie.  (It's been tried; it doesn't work.)
--
Being the last man on the Moon                  |     Henry Spencer
is a very dubious honor. -- Gene Cernan         | henry@zoo.toronto.edu



From: Paul Dietz <dietz@interaccess.com>
Newsgroups: sci.space.history
Subject: Re: 'real' spin-offs from space
Date: Mon, 30 Mar 1998 07:02:01 -0600

Henry Spencer wrote:

> At the time, various microelectronics technologies were being pursued --
> ICs being only one of them -- and there was no clear winner and nobody
> with a commitment to buying any of them in bulk.  When everybody else was
> still just experimenting, MIT-IL was placing large orders for standardized
> ICs, and working with the suppliers to get performance up and failure rate
> down.  At the time they were finished, the first few prototypes of the
> Apollo computers contained about 2/3 of all the ICs in the world.
>
> The total volume of military IC orders did not exceed MIL-IL's orders for
> several years.

However, the investments in ICs weren't being made because of a few
prototypes.  They were being made because of the perceived large future
market, mostly military.  Remember, this was just after the 1950s, when
military spending was a much larger part of the GNP than it is today.

At best, the Apollo computers helped support some of the companies.
But, IMO, TI would have gone ahead anyway (not clear about Fairchild).

	Paul



From: Paul Dietz <dietz@interaccess.com>
Newsgroups: sci.space.history
Subject: Re: 'real' spin-offs from space
Date: Mon, 30 Mar 1998 19:54:46 -0600

Henry Spencer wrote:

> Quite true.  But NASA and MIT-IL accelerated the process significantly by
> a bold decision to make immediate extensive use of what was still a very
> new technology (as witness the fact that several of their suppliers took
> quite a while to get useful production started).

But did this really accelerate the process significantly
(and what is "significantly" here, a year or two?).

As an alternative, perhaps the problem with getting useful
producton started just showed that Apollo was jumping
in before things were quite ready, and that had demand
increased more slowly the suppliers would not have been
as stressed.

	Paul




From: "Paul F. Dietz" <dietz@interaccess.com>
Newsgroups: sci.space.history
Subject: Re: Help re History of Flight Computers
Date: Sat, 31 Jul 1999 11:56:11 -0500

Jack Crenshaw wrote:

> He made the comment that the first IC's were Fairchild's Micrologic, ca.
> 1964.  I believe that's quite wrong.

1964 is much too late.

> As I say, the first spaceborne computer _I_ was exposed to was the
> Apollo flight computer.  I'm sure that early ICBM's had analog, rather
> than digital, controllers.  What I'm fuzzy about is how many flight
> computers, for military missions, preceded the Apollo mission.

An important one was the Minuteman II guidance computer,
in 1962.  It used ICs from TI (who, with the Air Force,
demonstrated the first IC-based computer in 1961.)  The MM II
was, I think, the first large scale use of ICs.

Apollo bought a substantial fraction of the US IC production
over some period in the early days.

> The entire issue, for this gentleman, becomes:  What was the
> impetus for the efforts to miniaturize digital electronics?

Reliability.  It required a heroic and extremely expensive
testing effort to make the Minuteman I computer (which used
discrete transistors) sufficiently reliable.  Had the same
level of testing been applied to all military electronics,
it is estimated that the cost would have exceeded the US
GNP of the time.  Moving to ICs promised to reduce the number
of soldered connections.  Size, weight, and power reductions
were also welcome, of course, but reliability was a problem
even in non-space applications where these were less
important.

This problem with reliability would have bitten even in
the absence of a space program, so I think it is reasonable
to say that ICs would have developed without NASA.  Whether
they would have developed without the *cold war* at anywhere
near the rate they did is more debateable, but then we spent
trillions of (current) dollars on that war, so the spinoff
justification alone would be a hard sell (not to say that
the cold war spending could not be justified on other grounds.)

	Paul


Newsgroups: sci.space.history
From: henry@spsystems.net (Henry Spencer)
Subject: Re: Help re History of Flight Computers
Date: Sat, 31 Jul 1999 20:03:26 GMT

In article <37A32AAB.9D1BDBDD@interaccess.com>,
Paul F. Dietz <dietz@interaccess.com> wrote:
>This problem with reliability would have bitten even in
>the absence of a space program, so I think it is reasonable
>to say that ICs would have developed without NASA.

Correct.  Self-serving press releases notwithstanding, NASA was not
involved in the invention of ICs.  What NASA, more specifically Project
Apollo, did do was to hasten their widespread commercial availability, by
being the first large-scale customer for them.  At a time when military
customers etc. were ordering them by the hundreds for experimental use,
Doc Draper's lab at MIT was ordering them by the tens of thousands for the
Apollo guidance computer.
--
The good old days                   |  Henry Spencer   henry@spsystems.net
weren't.                            |      (aka henry@zoo.toronto.edu)


Newsgroups: sci.space.history
From: henry@spsystems.net (Henry Spencer)
Subject: Re: Help re History of Flight Computers
Date: Sat, 31 Jul 1999 20:27:33 GMT

In article <37A314BC.39328E93@earthlink.net>,
Jack Crenshaw  <jcrens@earthlink.net> wrote:
>This conflicts with my own memories, but I must confess that, after 40
>years, those memories are becoming suspect.  My recollection is that the
>Apollo flight computer was among the first digital computers in space,
>as well as the first miniaturized computers...

"Among the first" is reasonable.  The computer in the Polaris missile came
a bit earlier (and the Apollo computer evolved from the Polaris computer
to some extent, both being done by the same MIT group).  Using digital
computers for Apollo was a fairly bold step, and there was some debate
about whether it was really a smart move.

In the end, of course, it worked out pretty well.  When the Apollo CSM
design was rethought for Block II, a fair bit of Block I analog circuitry
was scrapped in favor of adding more software in the computer.

>He made the comment that the first IC's were Fairchild's Micrologic, ca.
>1964.  I believe that's quite wrong.  I recall going to work for IBM,
>ca. 1967, and learning about the System 360, which was already in
>production for some years before that.  I'm quite sure that the 360 was
>among the earliest computers to have IC's in them, though I readily
>admit that the things IBM called IC's were nowhere near what we think of
>today, and we might not even call them IC's at all...

IBM may have called them ICs, but nobody else did.  They were hybrid
circuits, assemblies of discrete components on ceramic substrates in
sealed packages.  An IC is an *integrated* circuit, multiple components
and interconnecting wiring formed in and on a single piece of silicon,
with no assembly required except for packaging.

IBM considered ICs for the 360, but rejected them as immature.  That was
arguably true at the time -- the usual oversimplified histories neglect to
explain that a wide variety of miniaturization technologies were being
explored around then, none was really out of the lab yet, and it was not
clear which would win.  IBM's hybrids ended up being one of the losers,
although IBM made them work quite well within their limits.  MIT made a
considerable gamble in choosing ICs for Apollo, and won big on it.

>...I'm sure that early ICBM's had analog, rather
>than digital, controllers.  What I'm fuzzy about is how many flight
>computers, for military missions, preceded the Apollo mission.

See above.  I think Polaris was the only operational one preceding the
start of Apollo development, although one or two more may have flown
before Apollo got off the ground.

>The entire issue, for this gentleman, becomes:  What was the impetus for
>the efforts to miniaturize digital electronics?  I and others  believe
>that NASA and DoD were at least strong drivers of these efforts.  He's
>basically of the opinion that it would all have happened anyway.

He is probably right, except that it would have happened much more slowly.
The military funded a lot of the early miniaturization work, while NASA
(via MIT) was the first large-scale customer for ICs in particular, and
ended up working with the major suppliers to improve reliability and
performance of the early types.

>...Again, this friend is saying that the space
>program had nothing to do with the development of computers, that it all
>would have happened anyway.

This is probably mostly true.  Space was never a big customer for the
computer industry, although it may have hurried things along in small ways
here and there.
--
The good old days                   |  Henry Spencer   henry@spsystems.net
weren't.                            |      (aka henry@zoo.toronto.edu)


Newsgroups: sci.space.history
From: henry@spsystems.net (Henry Spencer)
Subject: Re: Help re History of Flight Computers
Date: Sun, 1 Aug 1999 21:25:50 GMT

In article <37A46592.9D2894B@earthlink.net>,
Jack Crenshaw  <jcrens@earthlink.net> wrote:
>> Correct.  Self-serving press releases notwithstanding, NASA was not
>> involved in the invention of ICs.  What NASA, more specifically Project
>> Apollo, did do was to hasten their widespread commercial availability, by
>> being the first large-scale customer...
>
>Henry, do you know what kind of circuitry was in the Apollo flight computer?

I'm not sure, offhand, of the exact cirucit technology used; the Hall book
("Journey To The Moon") may have this information, but I don't remember
the details.  The digital circuitry used a single IC type, a NAND (NOR?)
gate, in the interests of getting maximum production volume and dealing
with the smallest number of parts.  A later upgrade, which went into the
Block II computer I think, used a two-gate chip for denser packing.  One
or two other ICs ended up being used out on the fringes, e.g. in the
core-memory interface circuitry.

>How about Mercury?

If memory serves -- I don't have Mercury technical details on hand -- the
Mercury capsule had no computer at all, just an assortment of specialized
circuits and timers.
--
The good old days                   |  Henry Spencer   henry@spsystems.net
weren't.                            |      (aka henry@zoo.toronto.edu)


Index Home About Blog