TSU Meeting SUNDAY — August 10, 2025

Asa Gordon, David Anderson, Jerome

Sunday's meeting was a special-guest session with Asa Gordon (Robert A. Gordon), a retired NASA Goddard astrodynamicist who, in the mid-1980s, ported his satellite-orbit theory onto a Timex/Sinclair 2068 and demonstrated that an 8-bit home computer could produce ephemerides accurate enough to rival mainframe results. Gordon walked the group through his NASA career — starting at Goddard during Apollo, writing the penumbra/umbra shadow code used on Apollo 8's lunar far-side blackout — and the publications (1973 Celestial Mechanics paper, 1978 GTDS comparison, 1986 Reno conference paper) that led him to look for a microcomputer capable of real scientific math.

The technical heart of the talk was why the TS2068 specifically made his work possible: the upgraded BASIC ROM's expanded floating-point mantissa (4 bytes / ~8 decimal digits) with extra guard digits, the 3.5 MHz Z80, the full 48K RAM for lookup tables, and the 256×192 / 512×192 graphics for plotting ground tracks. He combined high-level BASIC with hand-tuned Z80 assembly for the heavy math, used bank-switching to keep cold data tables out of the working set, and implemented a "pseudo-physical" secular-drag model — averaging the decay of the semi-major axis per period rather than carrying density tables on board. Validation came from fits against simulated four-by-four zonal-harmonic ephemerides and real NORAD tracking data for the Solar Maximum Mission and Hubble; his analytic BG (Brouwer-Gordon) theory matched or beat the full numerical Cowell integration on in-track error over 2/4/16-day predicts.

Gordon also showed his later DOS-era productization — a Microsoft-BASIC menu front-end driving FORTRAN orbit code (orbital elements, Keplerian, Cartesian, spherical inputs; perigee/apogee/node/latitude predicts), copyrighted and sold at $49.95, restricted to 16-bit Windows (XP being the last that ran it). The original TS2068 BASIC+assembly listings appeared in his NASA technical paper. He talked candidly about the institutional resistance he met inside Goddard versus the recognition his published papers earned from outside reviewers (John Vinti at MIT graded his first paper; Samuel Herrick gave him an A at UCLA for ingenuity on a single problem), and about Melba Mouton and Christine Darden as formative influences.

Gordon asked the group for help: he wants someone to re-enter his TS2068 BASIC "Gotcha" game (with user-defined graphics) and his orbit code from the published listings onto modern storage so he can donate working artifacts — TS2068, papers, disks — to the library/archive collecting his civil-rights and scientific papers. David noted modern loaders exist that don't require cassette. Next meeting was set for Labor Day, Monday September 1st at 9 PM.

Highlights

Guest Asa Gordon (NASA Goddard astrodynamicist) presented how he ran professional-grade satellite orbit theory on a TS2068 in the mid-1980s
Talk traced why the 2068 specifically worked: expanded floating-point mantissa + guard digits, 48K RAM for lookup tables, 3.5 MHz Z80, and BASIC + hand-tuned Z80 assembly
Gordon's BG (Brouwer-Gordon) analytic theory with a pseudo-physical secular drag model matched or beat full Cowell numerical integration on Solar Max and Hubble tracking data
He later wrapped the FORTRAN orbit code in a Microsoft-BASIC menu (orbital/Keplerian/Cartesian/spherical inputs) sold at $49.95 — runs only up to Windows XP
Open ask to the group: re-enter his TS2068 BASIC 'Gotcha' game and orbit listings onto modern storage for donation to his archive

Topics: TS2068 / TC2068 · BASIC programming · Machine code, Z80 & assemblers · Utilities & applications · Tape & library archiving (TAP/TZX)

Summary generated from the meeting recording; the full transcript is kept in the group's private archive.