COSMAC 1802 on AMSATs running IPS

Last updated aug 6 2022, IPS and OSCAR references updated. Edited by Herb Johnson, (c) 2022 Herb Johnson, except for content written by others. Contact Herb at, an email address is on that page..


The intent of this Web document, is to document the earliest use of the COSMAC 1802 in amateur radio satellites, designed, built and/or launched in the late 1970's and into the 1980s. It compliments another Web document on commercial and research use of the 1802 in spacecraft. Later, I've documented notes and references to IPS, the Forth-inspired operating system used on AMSAT Phase 3 spacecraft. This Web page is part of this Web site's support of of Lee Hart's 1802 Membership Card project.

Amateur radio satellites

Satellites have built by radio amateurs since the 1960's, to be piggy-backed into orbit as ballast during launch of commercial or scientific satellites. AMSAT (AMateur SATellite Corp) was organized in 1972 to design and promote these activities. In the late 1970's, the 1802 processor was incorporated into some of these satellites. Some of those satellites were built by the University of Surray, by what became their Space Science Centre, and called UoSATs. In the same period, AMSAT designed and built a series of "Phase III" or "Phase 3" satellites. Once in orbit, all these satellites were designated by AMSAT as "OSCAR" (Orbital Satellite Carrying Amateur Radio) and enumerated accordingly.

Since the radio amateur community was primarily interested in the communication features of these satellites, most articles about them are not about the details on their internal construction. This material is hard to find on the Web today. Construction details for the early UoSATs are also hard to find. Primary sources for information are mostly professional astronomical or engineering publications, or articles in QST and other magazines. Some of these are online but for purchase only.

In Aug 2011, I gathered references and reviewed on-line material. I also got materials from Steve Gemeny and notes and references from Lee Hart and Dennis Boone. This information was updated in 2015 and in 2019; in 2022 I updated and some information was lost. My references are linked inside this document, follow them and or use Web search.

Some AMSAT satellies with COSMACs

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. So OSCAR-9 / UoSAT-1 was the first amateur satellite to orbit a microprocessor (an 1802 and a F100L). A Phase-3A AMSAT satellite, launched in 1980 and built and designed earlier, likely had an 1802, but failed to achieve orbit. But the Phase-3A likely competes with the JHU/APL MAGSAT satellite and the Galileo planetary spacecraft for the earliest design of an orbited satellite which used an 1802. However MAGSAT was first to orbit, with an 1802. As described on this Web page, other Phase 3 satellites were designed, most were lanuched and some served in orbit. One was built (P3E) but not launched as of 2019; however it does not have a COSMAC processor.

However, as far as "firsts" for microprocessors, read the linked Web page about Pioneer Venus and the Intel 4004. (Herb 2022)

Why was the 1802 considered and used?

From Lee Hart: "The original 1802 and subsequent 1802A are plain old metal-gate CMOS, just like the generic 4000-series CMOS. The later SOS (silicon-on-sapphire) version came out much later. They were very rare (and very expensive), and ran at some amazing clock speed like 18 MHz.

I believe the radiation hardening was in part due to the intrinsically high noise immunity of CMOS. It was a consequence of the low speed, large transistor sizes, and high supply voltages allowed. Then, they used special packaging to provide better shielding. The later CMOS SOS chips were better yet, but not available to us mere mortals." - Lee Hart

The SOS process, is where sapphire wafer (Aluminum oxide crystal) is the material substrate upon which the silicon is deposited. Check around the Web for more information.

In Aug 2022, John D'Ausilio posted in cosmacelf at, that while working at Goddard he spent some time at the AMSAT lab "up the hill". Also "We flew the SOS version on AMSAT birds. I remember seeing one in the parts room :)" David Schultz posted in response to a question about the COSMAC radiation hardening, a link to a government report on the RCA/Sandia Labs rad-hard process for the CDP-1800 series processor, a pure-silicon process not SOS. - Herb

Details of AMSAT satellites

Chronology of amateur satellites of the 70's and 80's

This is by no means a compreshensive description of amateur radio satellites. It's simply notes on am-sats just before and after the introduction of microprocessors into them. had - repeat, had - a chronology of OSCAR satellites on their Web site. Perhaps other WEb sites have this early history. Several OSCARs were launched before the late 1970's but built before "radiation-hard" microprocessors like the 1802 became available. The last of these include these "Phase 2" satellites:

AMSAT-OSCAR 7 (OSCAR 7) - launched in 1974 and semi-operational into the 21st century
AMSAT-OSCAR 8 (OSCAR 8) - lanuched in 1978, non-operational today.

Early AMSATs which used microprocessors include:

Phase-3A was destroyed  during launch failure May 23, 1980.
UoSAT1-OSCAR 9 was launched October 6, 1981,
AMSAT-OSCAR 10 (Phase 3B) was launched     16 June, 1983
UoSAT2-OSCAR 11 was launched 1 March, 1984.
AMSAT-OSCAR 13 (Phase 3C) was launched on 15 June, 1988
UoSAT3-OSCAR 14  was launched 22 jan 1990.

[Note: there were also Soviet amsats in the same period. I've not looked at their design. - Herb, Nov 2012]

According to "The Satellite Experiment's Handbook", Martin Davidoff, pub 1990 by ARRL: "The OSCARs 5-8 were hard-wired logic". And Oscar 8 was built and launched quickly as a Phase II design to replace Oscar 7, even though Phase 3 amsats were in development at the time. If anyone has specific primary references about the specifics of onboard controllers for these satellies, I'd like to know.

I'll describe below the UoSAT and AMSAT Phase III programs. But based on findings I've read to date, OSCAR 9 / UoSAT-1 and OSCAR 10 / Phase3B each definitely had an 1802 processor and made orbit. Phase-3A had an 1802, but it failed to achieve orbit, so it technically can't be counted as a "first microprocessor in orbit".


I can find, with effort, info on UoSAT 2 and 3 on the Web; the UoSAT-1 (OSCAR 9 after launch) is largely described in trade publications not freely available. The original UoSAT was developed by the University of Surrey as an initial effort. It led to what became their Space Science Centre where they developed commerical and scientific spacecraft, to the present day. But detailed information about these earlier spacecraft is not described on their current Web site. At best online, I can find articles published years later, which appear to contradict each other about the processor use on UoSAT-2 and later - and little about the hardware details of the original UoSAT.

But a 1987 IEEE paper, "The primary UoSAT-2 spacecraft computer" by Peel (see references below) describes the UoSat-1 and -2 craft. The UoSat-1, launched Oct 1981, was a product of a "short 4-month" process from design to construction and test. The 1802 is described as the "primary computer", while the "secondary computer" is called "F100L" (a Ferranti 16-bit processor). Memory features (16K of DRAM) and issues are discussed in the paper, obtained for me by Steve Gemeny.

Experiences from the UoSAT-1 led to enhancements in the UoSAT-2 (OSCAR-11), which also used the 1802 as primary processor, but the National Semiconductor's NSC-800 was the secondary computer. 1802 memory was upgraded with an additional two banks of 16K each. The architecture of the UoSAT-2 is described in great detail.

A 2000 article from the U of Surray Space Centre (see references) shows only a chart for processors in various UoSATs. The UoSAT-2 and UoSAT-3 are shown as containing a 1802. UoSAT-3, UoSAT-5 and a few later craft are shown as also containing a "Z80" which I've identified above as a National NSC800, which works like a Z80. The chart also shows UoSAT-3 and later craft as including a 80C186. UoSAT-1 is simply not mentioned!

A Wired magazine article for March 2004 (references), has a retrospective on the 20-year lifetime on the OSCAR-11 / UoSAT-2. It describes it as "[with] only 36K of RAM, and its processor is a 1-Mhz, 8-bit RCA COSMAC 1802. For its time, that was the state of the art for any spacecraft."

A 2003 article from a Space Center member (references) describes UoSAT-3 and -5 with an 80C186 processor; and UoSAT-3 with a secondary Z80 processor operating the craft at that time. UoSAT-2 is only described as containing "SRAM technologies".

UoSAT 3 / Oscar 14, was launched in Jan 1990, and operated until Nov 2003 when shut down due to progressive battery failure. says it's still in orbit!

Phase 3 AMSAT

Phase 3 represents a series of AMSAT designs of the late 1970's into the 1980's. It was a combination of hardware and software designs. But most articles about Phase III in AMSAT newsletter publications of the time, Such as AMSAT Newsletter and AMSAT Orbit, (archived at until 2019, no longer available in 2022) are about radio and observational issues, not the internal hardware of the spacecraft.

The AMSAT Technical Journal for Winter 1987 has an article on the "Integrated Housekeeping Unit" or IHU design for the Phase IIIB spacecraft, written by Gordon Hardman KE3D. It describes the IIIB craft as "a good example of the series" and containing "the RCA COSMAC 1802" as part of the original design "planning back to 1975". A chart at the end of the AMSAT article lists a "genealogy" including "Phase 3-A...lost (launch failure)" Other information says P3B was launched in 1983 and operated with engine damage until 2003 as OSCAR 10.

A description of the AMSAT P3C-OSCAR 13 launched in 1988, was described in an AMSAT document as commanded "around an RCA 1802 central processing unit supported by 32 kilobytes of RAM." It ran IPS software which "was used to control other 1802-based satellites, such as OSCAR-10 and OSCAR-11..". It operated until 1996. (In 2022 I had to replace an reference with the current Web link which does not mention processors.)

The last launched Phase 3 satellite apparently was P3D aka OSCAR 40, launched in Dec 2000. The "IPS" notes below confirm it had a COSMAC processor. After many years of service, it malfunctioned in 2004 and failed in orbit; see Wikipedia or AMSAT for further information. A Phase 3E (P3E) satellite was built in the 2000's, but to date (2019) has not flown. Descriptions of its IHU describe a non-COSMAC processor.


As mentioned above, IPS is a threaded language and develoment enviroment, similar to Forth. It was used to develop the COSMAC programming on the Oscar Phase 3 satellites in discussion on this Web page; it found use on non-COSMAC satellite platforms. This Web document found in 2012 and found again in 2022, titled "IPS Programming" and dated October 28, 2002, says: ""The OSCAR-40 satellite, like its predecessors P3A, OSCAR-10, and OSCAR-13, has as its "brains" a computer system based on the RCA COSMAC 1802 microprocessor..." The AMSAT Web page also discussed IPS and offers MS-DOS and Windows support software. Thanks to Mark Saundby for directing me to the original information in 2012. The ftp link appears to have some IPS software in the ips folder.

Lee Hart found the following information and posted in cosmacelf: "Anyway, the Jan 1979 issue of Byte Magazine had an article "IPS: An Unorthodox High Level Language" by Dr. Karl Meinzer. It described a version of FORTH that was used on the 1802 in the ham's OSCAR satellites. I haven't found dates for all the OSCAR launches, but AMSAT OSCAR-10, launched in 6/16/1983, definitely had an 1802 on board. This definitely precedes Galileo's launch in 1989." Lee also researched this information in QST magazine, a radio amateur publication.

Dennis Boone responded in cosmacelf: "IPS isn't a version of FORTH, though it does have key similarities. The point of developing IPS in the first place was that AMSAT couldn't afford to buy a FORTH license at the time, and in any event they wanted multiprocessing support that I believe they felt was missing from FORTH. Since the original development of the language was by a German, the language keywords are in German. There was an effort to "translate" the language for English-speaking programmers, but my understanding is that the translation never really caught on, and the key US contributors to the satellites ended up writing German IPS."

(Dennis also referred me to other Web resources about IPS, which since have gone away or are redundant to more recent resources as below. - Herb)

In 2015, Paul Willmott (AMSAT P3 Command Team) contacted me with this additional information: "'IHU' means Integrated Housekeeping Unit. AMSAT-OSCAR-40 as you say used the 1802 as well (IHU-1), but it also had a secondary experimental computer t'at was ARM based. The ARM computer (IHU-2) was used to control the camera. Compressing JPEGs was way beyond the capabilities of the 1802. AMSAT-OSCAR-40 was the last AMSAT spacecraft to use the 1802, from then on off-the-shelf ARM chips were purchased. These chips were radiation tested and found to be suitable for flight."

Some years ago, a original, reference book for IPS programming was offered for sale as a printed document, proceeds to support AMSAT. The book is "IPS - High Level Programming of Small Systems" by Dr. Karl Meinzer, ISBN 0-9530507-0-X. In the book, there are IPL assemblers for the COSMAC 1801/1802, the 8080, the 6800 and the 6502; and deep discussion of IPS and some examples. The book was out of print since 2010 or so; then a PDF of the 3rd revision was available. Then in May 2019 a new print edition was published via Amazon, by editor/publisher Juergen Pintaske, titled "IPS - A Forth-like language for Space"; Meinzer is again the author. (PDFs of some version of the book have appeared later, at sites with IPS source codes as described below.)

The AMSAT-BDA organization has work about IPS on this linked Web page from the Berumda AMSAT organization. It also described the Meinzer book, and links to Miller's Web page. Also it links to some IPS software and guides as I'll describe below. Note that a lot of the software, is in the IPS Forth-like language.

The IPS software seems to include a Windows IPS emulator (IPS and Windows executables only, some IPS code fragments). There's a Linux IPS interpreter in C source with some IPS code (implemented in ARM assembly). The ASMSAT-BDA Web page has similar software, plus "pure PC-DOS" IPS-M software in English and German, implemented in Turbo Pascal. Even with the book's content, and these software resources, it's not clear to me that a complete IPS embedded native system (1802 in particular) could be reconstructed from these resources.

However, after I heard of the re-publication of the Meinzer text in May 2019, I made inquiries to James Miller and Dennis Boone. Dennis agreed to make available, an IPS-C code implementation in 1802 code. Dennis requested this attribution: "This work was originally carried out in the early 1980s by members of Amsat-DL, the radio amateur satellite organisation of Germany, then based in the University of Marburg. As the 1802 operating system 'IPS', it performed flawlessly in several amateur satellites for many years in space." My description of some of the OSCAR satellies confirms this description.

In 2019, there was some process going on, to make the old IPS software avaialable. But for reasons unknonw there was both interest and reluctance to do so; I lost touch with the process. In 2022, I learned that sometime later in Aug 2019, a github project was produced, which released various IPS implementations and some AMSAT applications. Here's the github project page for IPS. It's apparently produced/published by "AMSAT-DL", apparently the AMSAT affiliate in Germany. Thanks to John D'Ausilio for posting that reference in cosmacelf at in Aug 2022.- Herb Johnson

an 1802 embedded IPS system?

A plausible way to re-implement a native 1802 IPS, with the Meinzer book and IPS coding examples as previously described, may be as follows.
1) use an established 1802 emulator or hardware system with terminal I/O and known monitor code.
2) hand-assemble the IPS-C code's assembly language, to produce a working IPS interpreter/compiler, which accesses that monitor/emulator code.
3) "feed it" the IPS-C IPS code to compile it, and produce a compiled image that runs "from the terminal".
4) work on that image and backwards through this process, to generate a stable IPS-C embedded system, from a stable set of these programs.
5) replicating the cassette file system in some fashion, so IPS programs and other codes can be saved and loaded.

Please inform me if you attempt and/or complete this task.

32-bit IPS

Based on AMSAT-BDA information (see references above), the Phase 3E and Phase 5A spacecraft will/had 32-bit processors and use a 32-bit implementation of IPS. Here's a coding reference I found on that Web site.

Notes, references

OSCAR 7 in the 21st century

Please keep in mind, there are no microprocessors in OSCAR 7. I describe it purely out of respect and admiration for vintage technology. - Herb Johnson

Courtesy of Lee Hart, Aug 2019:Amsat Oscar 7 was launched in Nov 15, 1974, and is *still* operational [in 2019]... but only when it is in sunlight, as its batteries failed in 1981. [It's run by solar power.]

Its batteries began to fail in mid-1981. The on-time got worse and worse, as its batteries apparently developed shorts. But in 2001, the last short apparently cleared itself, so the spacecraft has resumed full operation whenever it is in sunlight. As a result, it may be the oldest communications satellite still in operation.

In 1982, when Oscar 7 was fading in and out of operation, ham radio operators in the Polish "Solidarity" movement fighting against the communist government used Oscar 7 for secret communications between their chapters. They used high-gain VHF antennas pointed straight up; so the Russian government wasn't able to eavesdrop on their conversations. :-) - Lee Hart

In 2022, there's still references to operating OSCAR 7 or AO-7; do a Web search and start with wikipedia, or this Web page .

additional references

"The Satellite Experiment's Handbook", Martin Davidoff, pub 1990 by ARRL. It only mentions Phase 2 amsats briefly, and puts more attention on Phase 3 amsats. But p 15-20 on "satellite systems" states: "On OSCARs 5-8, hard-wired logic was use to interface the various spacecraft modules to both the telemetry system and the command system." Other information in the book notes that "...With the fear that AO-7 might not last until the the first Phase 3 satellite was launched", Oscar 8 was produced as an "interim Phase 2 satellite..." (page 3-8, chapter titled "the 1970's").


The Experimental IHU-2 Aboard P3D. Thanks to Mark Saundby for directing me to this information.

quote - "The IHU flown on previous P3 missions, and which will fly on P3D, is based on the COSMAC CDP-1802 processor. UoSAT-1 and UoSAT-2 also used the 1802. Originally developed by RCA in the mid 1970s, this processor is currently made by Harris. Sandia National Laboratories (US) spun a radiation-hardened version circa 1980, and AMSAT has had the good fortune to obtain a small number of them for the P3 program..."

P3 is AMSAT produced Phase 3 missions and satellites. UoSAT are early satellites of the University of Surrey. The above Amsat document doesn't mention Phase 2 missions or satellies.


"The primary UoSAT-2 spacecraft computer" by Peel, R.M.A.

This paper appears in: Electronic and Radio Engineers, Journal of the Institution of Electronic and Radio Engineers Issue Date: September-October 1987 Volume: 57 Issue: 5 On page(s): S132 - S142 ISSN: 0267-1689

Abstract from the Web site index: [quote]

The UoSAT-2 spacecraft incorporates two ground-programmable computers. The primary computer is based around an RCA 1802 microprocessor and is intended for general-purpose management and data collection purposes within the spacecraft, whilst the secondary computer, based on the National Semiconductor NSC-800 microprocessor, controls the Digital Communications Experiment and has a more limited interface to the rest of the spacecraft. This paper describes how the experience gained from operating the UoSAT-1 spacecraft led to an enhanced primary computer design on UoSAT-2. [end quote]

This paper is only available by IEEE subscription or payment. Thanks to Steve Gemeny for providing a copy to me, which of course is copywrited. - Herb


"A low-power 16-bit computer for space applications on UOSAT" by Haynes, C.L.F

This paper appears in: Electronic and Radio Engineers, Journal of the Institution of Electronic and Radio Engineers Issue Date: August/September 1982 Volume: 52 Issue: 8.9 On page(s): 391 - 397 ISSN: 0033-7722

The abstract describes "UOSAT secondary computer provides a system for the evaluation of higher risk computing technologies, including a 16-bit bipolar microprocessor and large static o.s. memories." Other references (see above) reference the F100L 16-bit microprocessor by Ferranti as the secondary computer.


from Harvard abstracts

available to members only.

Title: 18 Years of Flight Experience with the UoSAT Microsatellites Authors: Underwood, C. I. Journal: European Space Components Conference (ESCCON) (2000 : Noordwijk, The Netherlands). Proceedings of the European Space Components Conferences ESCCON 2000, 21-23 March 2000, ESTEC, Noordwijk, The Netherlands Edited by B. Schürmann. Noordwijk, the Netherlands: European Space Agency, 2000. ESA-SP, Vol. 439, p.5 Bibliographic Code: 2000ESASP.439....5U

A graphic chart in the paper, shows only UoSAT-2 and UoSAT-3 as containing a 1802. UoSAT-3, UoSAT-5 and a few later craft are shown as containing a "Z80" (acutally an NSC800, National's processor with a Z80 instruction set) and an 80C186. I question this chart's accuracy.


The Space Science Centre for the University of Sussex in Surrey continues to produce spacecraft and instrumentation. AMSAT and UoSAT-2 are only mentioned but not described there.


Wired magazine has an article on the TWENTY YEARS of operation of UoSAT-2 OSCAR-11, published March 2004. "UO-11 has only 36K of RAM, and its processor is a 1-Mhz, 8-bit RCA COSMAC 1802. For its time, that was the state of the art for any spacecraft. "


"Observations of radiation in the space radiation environment and its effect on commercial off-the-shelf electronics in low-Earth Orbit". Craig I Underwood, (Surrey Space Centre, University of Surrey) Phil. Trans. R. Soc. Lond. A 2003 361, 193-197 doi: 10.1098/rsta.2002.1122

The paper was published in 2003. The paper notes UoSAT-3 and -5 have an 80C186 processor; UoSAT-3 has a secondary Z80 processor operating the craft at that time. UoSAT-2 is only described as containing SRAM technologies.



25 Years of space at Surrey - Pioneering modern microsatellites

M. N. Sweeting , 
Surrey Space Centre, University of Surrey
Published in 
Acta Astronautica
Volume 49, Issue 12, December 2001, Pages 681-691

An abstract is at the Web link above. The article is available for purchase online for $31.50. The online outline mentions UoSAT-1 through -5 among other UoSats. A copy is available from for members only.


"Integrated Housekeeping Unit - A Method of Telemetry, Command and Control for Small Spacecraft" by Gordon Hardman KE3D published in The AMSAT Technical Journal, Winter 1987, presented in Oct 1987 at 1st Utah State Univ. conferencs on small satellites.

The design of Phase III satellites is dated as "planning back to 1975", and the IIIB is "the second of the three spacecraft so far constructed, and is a good example of the series". The processor for AMSAT Phase IIIB is described as "The RCA COSMAC 1802....primarily because it was the only suitable CMOS device available at the time of the initial design". A chart at the end of the article lists a "genealogy" with "conceptual design, 2K NMOS; Phase 3-A, 16K NMOS, lost (launch failure); Phase 3-B, 16K NMOS, 1983-1987" and two Phase 3-C craft awaiting lanunch in 1988.


AMSAT history of OSCAR satellites on the AMSAT Web site.

OSCAR 9, 10, 11, 12, 13 and others earlier and later are referenced.

Contact information:
Herb Johnson
New Jersey, USA
To email @ me, see
see my home Web page.

This page and edited content is copyright Herb Johnson (c) 2022. Contact Herb at, an email address is available on that page..