Last updated Aug 3 2019, minor update Aug 6 2022, Aug 6 2024. Edited by Herb Johnson, (c) 2024 Herb Johnson, except for content written by others. Contact Herb at www.retrotechnology.com, an email address is on that page..
Note: this section was moved from the Membership Card "home" Web page on July 29 2011. The Membership Card is an 1802-based handheld microcontroller designed and sold by Lee Hart. - Herb Johnson
For some years up to Dec 2009, it was a common error on the Web, to attribute the COSMAC 1802 as in use in the Viking or Voyager planetary spacecrafts of the 1970's. In the cosmacelf discussion group at that time, there was interest in recovering the "Viking 1802 code". At that time, I researched primary information sources for Viking, Voyager and Galileo hardware features. I found at that time that among those crafts, only the Galileo used the 1802 microprocessor. Galileo was among the first spacecraft (apparently, but see below) to use a microprocessor, in fact Galileo used multiple 1802 processors. Concurrently, others in cosmacelf found and reported similar information about Viking and Voyager. I discussed this error in cosmacelf and provided references there and on this Web page as below. Subsequently, many Web sites (including Wikipedia) have corrected their information.
Further discussion led to a review of amateur satellites and their use of the 1802. My findings on amateur radio satellites of the period and 1802's, are in the linked Web document. I concluded that the COSMAC 1802 was "the first microprocessor in space", details described on this Web page. However, in mid-2019, I was informed about another microprocessor used in instrumentation on Pioneer Venus in 1978; that's likely a prior use, but not as a spacecraft controller. Read this Web page and those linked, for all the details.
- Herb Johnson, Aug 2019 updated Aug 2022
the NASA Mars Viking did not use an 1802 processor! check these references
..and neither did the NASA JPL Voyager spacecrafts. More references.
Here's abstracts and links to NASA reports on the Galileo spacecraft and its multiple 1802's
Steve Gemeny, an active member of Yahoo's cosmacelf group, worked on a number of NASA spacecraft projects at the Johns Hopkins Applied Physics Laboratory (JHU/APL). In 2002 and 2005, Steve posted in cosmacelf, remarks about his experiences with the 1802 as a teen in the 1970's, and again as a parent showing his 1802 kit to his son. He also describes 1802 history from his experience in aerospace, and examines early use in medicine. Read his remarks, made available on this site in July 2011 with his permission.
Steve Gemeny co-authored a paper in 2002, published by JHU/APL, on a similar theme. The title is "Ground System Planning for Long Duration Space Missions - Helped by Lessons Learned Resurrecting Obsolete Computers", part of the RCSGSO conference in 2003. The paper discussed reviving or maintaining decades-old space mission hardware. He cited reviving his 1802 kit as an example revival effort, the same processor as in APL's design of an 1802-based command processor for the MAGSAT spacecraft for the US Geological Survey. (An additional reference to the multi-1802 design of the Galileo spacecraft, was misnamed as "Voyager" in publication. Steve has provided a copy of his paper to this site; at his request the copy includes a correction accordingly.)
Was Galileo "first" for the 1802? The Galileo spacecraft was launched on October 18th, 1989 via the Space Shuttle and ran for 14 years. It was also an enormous, decades-long program of research, planning, revision and development before launch. Little of the abundant literature on the mission is about design and construction of its computers; it's like looking in the weeds to find those details. The Galileo program was officially funded in 1977; construction contract was awarded sometime in 1978 to JPL, for a 1982 intended launch, but that date slipped forward for numerous reasons. Initial assembly was completed in 1983 for testing; tests, redesigns and replacements of various parts occurred numerous times, up through the days before launch.
NASA's "Computers in Spaceflight: the NASA Experience" has a pretty informative chapter about the "Unified Data System" design CONCEPT as used in Galileo, noting events and dates for Galileo's orbiter development. It references documents in the academic and trade literature about design, dating back to 1976. It also notes that UDS was previously implemented in Voyager - on CMOS but discrete hardware, not on microprocessors. The confusion between Voyager's and Galileo's hardware years later, may in part be due to mix of methodologies, experiences and people inherited from the older program to the newer. In part, it may be due to years of design consideration, review, and development both before and after hardware was built.
The "Computers in Spaceflight" document refers to a "[JPL's Edward H.] Kopf interview" when it notes "NASA chose RCA's 1802 microprocessor for the Galileo implementation of the UDS...at the time (c. 1977)..." but it's likely the processor was discussed in publications perhaps a year earlier, as suggested by other references in that document. It's likely only some obscure document or reference can date the time the first 1802 processor (of several) was built into either test or flight hardware. And software was developed before that, in simulation and was based on Voyager's UDS and HAL software.
Lee Hart built a model of the Galileo spacecraft as part of a 2014 exhibit of COSMAC technology. See his construction project on the linked Web page.
Meanwhile: In follow-up private discussions during July 2011, Steve noted the prior use of the 1802 by the MAGSAT earth-orbiting satellite. Steve sent me a Feb 1980 report of "The Microprocessor-based MAGSAT Command System", a JUH/APL report written by Ark L. Lew (see references). It describes the MAGSAT command system as designed in the summer of 1977, when it was decided to use the CMOS 1802 over the NMOS 6800 or 8080 processors, due to the RCA 1802's CMOS features of radiation hardness, low power, and static clocking. The paper describes a dual processor configuration supporting some redundancy. The architecture and command structure is described with diagrams and some photos of hardware.
MAGSAT was launched Oct. 30, 1979, according to numerous WEB-accessable accounts of the MAGSAT program. There's a detailed account of the MAGSAT program in Johns Hopkins APL's journal "Technical Digest" which devoted their July-Sept 1990 issue (Vol 1 No. 3) to the program; the report confirms the dual-1802 design was in use. The satellite completed its mission and reentered the atmosphere in June 11 1980, into the Atlantic Ocean.
Based on findings I read to 2019, OSCAR-9, the original UoSAT, orbited in 1981, OSCAR 10 / Phase-3B orbited in 1983, and OSCAR 11 / UoSAT2 orbited in 1984, each definitely had an 1802 processor and made orbit. A Phase-3A AMSAT satellite, launched in 1980 and built and designed earlier, also had an 1802, but failed to achieve orbit. The 1802 processor was the primary microprocessor in use on UoSAT-1 and -2; other processors were secondary (backup). But the AMSAT Phase-3 competes with the JHU/APL MAGSAT and Galileo satellites for the earliest designs of a satellite which used an 1802. However MAGSAT was first to achieve orbit. Various of the OSCARs and other AMSATs, and later non-AMSAT satellites, were developed for and ran under a software system called "IPS"; a FORTH-like programming environment. My findings on amateur radio satellites of the period and 1802's, and references to the IPS software, are described in the linked Web document.
Until 2019, I concluded that the COSMAC 1802 was the first microprocessor to operate in space aboard an operational spacecraft, in the late 1970's. But in July 2019, it was suggested to me that there was an earlier microprocessor in space. Read my Web page on the Pioneer Venus and the launch of a spaceprobe with an Intel 4004, in Aug 1978.
So it's debatable as to which spacecraft was first to use the COSMAC 1802. Either way, Galileo or MAGSAT or AMSAT Phase III, the 1802 appears to be the first microprocessor to operate a launched spacecraft. But the Pioneer Venus "large probe", launched from Earth Aug 8 1978, and landed into Venus' atmosphere on Nov 19 1978, contained an Intel 4004 operating a mass spectrometer. So the first "microprocessor in space" seems to be an Intel 4004.
The earliest COSMAC into space was the MAGSAT in Oct 1979, for certain. The hardware choices for MAGSAT and Galileo, and for AMSAT Phase III amateur satellites, occurred during the same 1976-78 time frame.
As the Galileo spacecraft was far more complex, it took longer to design and construct; so it drew upon older resources. The 1802 was chosen for the Galileo, in part because CMOS is low power and radiation hard - necessary in the Jovian environment. Additionally, the Galileo mission requirements, funding, retesting, repairs, launch windows and weather - all caused delays before the launch in 1989. [At the time I wrote this in Aug 2011], it was difficult to Web research these events for the Galileo mission, or for that matter the earlier AMSAT and MAGSAT missions.
- Herb Johnson
"The Microprocessor-based MAGSAT Command System", written by Ark L. Lew, was a Feb 1080 report published by JHU/APL as document CP 077, and was previously published in the Proceedings of the Technical Sessions, International Microcomputers Minicomputers Microprocessors, 1979, Geneva, Switzerland; Pg 235-243
In late 2015 I was contacted by Gary Camp, an electrical engineer from the 1970's who started with a college senior project for an 1802 temperature controller; and who ended up as a mission controller on several of the NASA/JPL missions discussed here. Read the linked Web pages for details.
This page and edited content is copyright Herb Johnson (c) 2024 Contact Herb at www.retrotechnology.com, an email address is available on that page..