Full technical analysis now published of the engineering prototype for an unreleased Coleco peripheral that marketing called the ADAM Universal Interface — an AdamNet device exposing both a standard RS-232 serial port and a Centronics parallel port, with a three-position mode switch (Serial / Parallel / SmartWriter pass-through).
This is NOT a third-party board. The solder side is silkscreened "X00257 REV 0" alongside the boxed COLECO logo — a manufactured Coleco engineering board with a small number of hand-added mod wires, housed in a clear acrylic case. Two units known to exist.
The hardware is fully populated:
• Motorola MC6801 MCU in expanded multiplexed mode (Y1 = 4.000 MHz)
• 2KB 2716 EPROM, socketed, sticker reads "AHSPI REV 1.2 / 00F8"
• Signetics SCN2661A EPCI USART for user RS-232 (Y2 = 4.9152 MHz)
• µA1488 / 1489 RS-232 line drivers with ±12V DC-DC converter
• CD74HC373E octal latch for Centronics 8-bit data (schematic specifies 74LS273)
• Two MM2114N-3 SRAMs (1KB working RAM)
• 74LS04 / 74LS05 logic for wired-AND address decode
• Rear: J1 36-pin Centronics + J2 25-pin D female (RS-232)
• Front: 9-pin connector carrying AdamNet + external power (not the usual RJ12)
The identification rests on four independent lines of evidence that all agree:
1. The prototype itself — PCB silkscreen, EPROM disassembly, firmware vectors.
2. A Coleco Marketing Communications "New Products" memo describing the finished retail product ("ADAM UNIVERSAL INTERFACE," cream enclosure, ON LED, supports up to two units, software-compatible with SmartBASIC and CP/M).
3. The FujiNet project — open-source ESP32 AdamNet emulator that already implements this device (published Dec 2021), reconstructed by its author from the ADAM CP/M BIOS. Places the device at $0E. Hardware list matches the prototype.
4. AdamCalc (Coleco, 1984). The boot device-probe table at $32AB names each AdamNet device the application scans for; two entries explicitly name this one: $0E labelled "Serial" and $0D labelled "Parallel." Period Coleco software already knew about the Universal Interface.
Device identity is split per port: $0E = Serial, $0D = Parallel — each port presents an independent AdamNet device. The third switch position (SmartWriter) is a pass-through.
On the firmware — and this is the part that has advanced significantly: a complete disassembly of the 2KB ROM, originally reconstructed by Chris Braymen and later annotated by Richard F. Drushel, shows REV 1.2 is substantially finished, not abandoned. It drives the SCN2661 USART with buffered, bidirectional RS-232, with both hardware (RTS/CTS) and software (XON/XOFF) flow control; AdamNet runs interrupt-driven on the 6801's built-in SCI while the user serial port is serviced in the main loop. The one piece Coleco left unwritten is a small block of AdamNet STATUS response routines. The full disassembly, the reconstruction of those routines, and the verification work are documented in the analysis rather than recapped here, so the writeup stays the single current source.
A second piece of the story: Coleco shipped the HOST side of this peripheral in retail CP/M 2.2. The ADAM CP/M BIOS contains a complete AUX driver — buffered read with character count, a six-byte init block (baud / bits / parity / stop / flow / control), full status semantics — for a device that never reached stores (disassembled by Drushel, 1992). The host driver was finished and in the field; only the device firmware was left incomplete.
The architecture — AdamNet on the 6801's SCI, user serial via a Signetics EPCI, MC1488/1489 line conversion — is the same template that later appeared in the Orphanware, Hi-Tek, and MicroFox third-party RS-232 cards. The cards that became the de facto standard for ADAM serial were built along the path Coleco's own engineering team had already laid down.
Earliest possible build date: March 1984 or later (SCN2661A date code 8411 = week 11, 1984).
Full document (HTML + DOCX) — nine hi-res photographs, complete schematic walkthrough, the annotated 6801 disassembly, AdamNet device map, I/O memory map, the firmware reconstruction with verification, AdamCalc icon/probe-table extractions, and open investigations — on the site. The document is the living record; it carries the detail and is kept current as the firmware work progresses.
Disassembly credit: 6801 firmware disassembly by Chris Braymen, annotation by Richard F. Drushel; CP/M BIOS host-driver disassembly by Richard F. Drushel.
Open questions for the community:
- Has anyone seen another X00257 board or an AHSPI EPROM in any revision?
- Does anyone have a Coleco "New Products" memo, internal schedule, or engineering document referencing AHSPI or the Universal Interface project name?
- Anyone recall the production unit appearing in late-1984 / early-1985 dealer literature?
ADAM_Universal_Interface_Analysis.html / .docx
ADAM_Universal_Interface_Analysis_v2.0.pdf- (1.43 MiB) Downloaded 12 times