Skip to content

Timex Sinclair User Group Meeting — January 3, 2022

Watch on YouTube

David, Carl Miles, Jeff Burrell, Stewart Newfeld, Adam Trionfo, Tim Swenson, Keith

The bulk of the meeting was a deep dive into TS2068 cartridge internals, kicked off by Adam getting his test EPROM cartridge working. The group walked through the .dck file format (the nine extra bytes are Warajevo-emulator header, not part of the ROM), the eight cartridge header bytes, and the LROS vs AROS distinction — including which DOCK chunks each occupies. That led into a long thread on whether the TS2068 can auto-boot a Spectrum ROM: Jeff Burrell explained that on his expansion-bus interface he pulls the /BE (bank-enable) signal low via a 74LS138 decoding only the lowest two banks, plus a small Z80 stub that writes to port FF to set the SCLD video mode — which gives him a reliable Spectrum boot. The plan to move that approach onto a cartridge (and bypass the 2068 boot ROM entirely) was sketched out, with the Portuguese TC2068's emulator cartridge cited as proof it can be done. Adam also got a primer on chunk layout (the Zebra 64-column ROM and Spectrum ROM live in chunks 0–1; HotZ and Android in 4–7), and why BASIC-on-cartridge needs a 28K-to-32K relocation dance.

Stewart Newfeld joined and showed off the microswitch ZX81 replacement keyboard he sells (made in Germany), along with a recent eBay Spectrum he bought out of Cairo to refurbish. That spun off into a discussion of Aztec modulators (channel 2/3 vs UHF variants, the punched-out case differences between the British ZX81 and the US-market TS1000), and the lingering supply of those parts. Multi-image cartridge designs came up too — using a 74LS138 plus DIP/rotary switches to select among several ROM images on one larger EPROM, with the Argentine slide-switch multicart and an Astrocade multicart shown as references. Counterfeit chips from AliExpress were a cautionary aside.

Keith presented a genuine real-world TS1000/TS2068 application: in the early 80s he wrote a BASIC program that plotted horsepower curves for the operators at a natural-gas compression station, replacing inaccurate vendor curves. A run of 40 curves took four days of dot-matrix printing — over 170,000 loop iterations — and the operators were happy with the result. David said he'd post the PDF Keith sent. Tim Swenson described still using Quill on a QL emulator (SMSQ/E) on a netbook for word processing because he wants a black background, with Jeff Wicks' Quill-to-RTF converter to get files out. The group reminisced about Jack Dellaheny's keyword-entry program for the auric keyboard, and traced his later life as the "Merry Toy Maker" in Georgia.

The meeting closed with show-and-tell — Tim's TS1500, PC8300, Sinclair programmable calculator, Sinclair TV, and the X-1 button radio — plus a recap of David and Carl's recent in-person session comparing the RGB mod against composite on a TS2068. The RGB output is dramatically cleaner and fixes the "bright black vs black" greyness of composite, but because the TS2068 SCLD never brings the intensity bit out externally (unlike the Portuguese TC2068), RGB can only reproduce 8 of the 16 colours. Carl mentioned the small YPbPr→RGB converter boards (intended for the TI-99/4A) he had built up for trying to recover the bright bit directly from the SCLD's component output. Next meeting was set for two weeks out on a Sunday at 2pm to accommodate European attendees, with the maintainer of the load.zx-museum joining.

Highlights

  • Walkthrough of .dck/LROS/AROS cartridge headers and DOCK chunk layout, prompted by Adam getting his test EPROM cartridge running
  • Jeff Burrell's recipe for auto-booting a Spectrum ROM on a TS2068 cartridge: pull /BE low via a 74LS138 on the lowest two banks plus a Z80 stub poking port FF for SCLD video mode
  • Stewart Newfeld showed his microswitch ZX81 replacement keyboard and discussed the Aztec channel 2/3 vs UHF modulators he still has in stock
  • Keith's real-world TS1000/2068 application: BASIC horsepower-curve plots for a natural-gas compression station, 40 curves over four days of dot-matrix printing
  • David and Carl's RGB-mod comparison: dramatically cleaner than composite but limited to 8 colours since the TS2068 SCLD doesn't expose the intensity bit (unlike the Portuguese TC2068)

Topics: Cartridges, EPROM & dumping · TS2068 / TC2068 · Spectrum emulation & software · Video upgrades (composite, RGB, HDMI) · TS1000 / ZX81 / TS1500 · Newsletters & documentation

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