@S1-A.ARPA,@MIT-MC.ARPA:jim@TYCHO.ARPA (06/19/85)
From: jim@TYCHO.ARPA (James B. Houser)
A recent posting about shuttle cooling reminded me of an
article I ran across in an old (1980) IBM document. The article
described the onboard computers as System/4Pi model AP-101 processors.
The AP-101 appears to be a 3 KIP machine with an intensely ugly
architecture. Main memory was described as 104KW (36 bit) of core
(really core!) with 400 ns access time. What startled me were the
physical characteristics. Each of the five "processors" appears to
consist of two boxes. The boxes weigh 58.9 pounds and draw 350
Watts!!!!! This would imply a system total of almost 600 pounds and
3500 Watts. Did they actually use this refugee from the Smithsonian
on the Shuttle and if so are there any reasonable plans afoot to
upgrade? In a related question, what are the major contributors of
heat during the orbital portion of a shuttle mission?
-------arnold@ucsfcgl.UUCP (Ken Arnold%CGL) (06/21/85)
In article <2322@mordor.UUCP> @S1-A.ARPA,@MIT-MC.ARPA:jim@TYCHO.ARPA writes: >A recent posting about shuttle cooling reminded me of an article I ran >across in an old (1980) IBM document. The article described the >onboard computers as System/4Pi model AP-101 processors. The AP-101 >appears to be a 3 KIP machine with an intensely ugly architecture. >Main memory was described as 104KW (36 bit) of core (really core!) >with 400 ns access time. ... Did they actually use this refugee from >the Smithsonian on the Shuttle and if so are there any reasonable >plans afoot to upgrade? One of the main problems with space ship design is that, by the time the design sits on the launching pad, its technology is quite out of date. This is true of all technology, but since computer technology evolves considerably faster than other kinds, it is much more critical (as the above description shows). The California Space Resources institute (attached to the University of California at San Diego) is currently coordinating a study which includes this problem, relating the design of the space station. As I understand it, the problem is to define computer (and where possible other techonological) needs by interface, and maximum physical attributes (which, as things get smaller and generate less heat, becomes not much of a problem); basically a form of modular technology design. That's what seemed most promising last time I was informed, but it has been several months. As an aside, in case you've been wondering where your defense dollars are going, the same problem exists in front line equipment, which takes just as long to get to the front. Of course, the degree to which computers are out of date varies among different pieces of equipment. Ken Arnold
ken@hrpd3.UUCP (K.COCHRAN) (06/21/85)
Could the use of core memory have anything to do with data integrity when cosmic rays pass through the shuttle, or if power drops unexpectedly. Ken Cochran vax135!hr1ar!ken
al@aurora.UUCP (Al Globus) (06/27/85)
> > A recent posting about shuttle cooling reminded me of an > article I ran across in an old (1980) IBM document. The article > described the onboard computers as System/4Pi model AP-101 processors. > The AP-101 appears to be a 3 KIP machine with an intensely ugly > architecture. Main memory was described as 104KW (36 bit) of core > (really core!) with 400 ns access time. What startled me were the > physical characteristics. Each of the five "processors" appears to > consist of two boxes. The boxes weigh 58.9 pounds and draw 350 > Watts!!!!! This would imply a system total of almost 600 pounds and > 3500 Watts. Did they actually use this refugee from the Smithsonian > on the Shuttle? I'm almost positive the answer is yes, they do and will continue to do so for the forseeable future. If it's not that particular model, it is a close relative. Sickening isn't it? Software note: software development for launch is supposedly very close to the critical path and definitely must see major improvements to support 24 launches a year. More horrors: there are some 250 microprocessors on the shuttle with a variety of incompatable languages, operating systems, and development systems between them.