Skip to content

TSU Meeting MONDAY — August 7, 2023

Watch on YouTube

David, Adam, Paul, Willie, Carl Miles, Ryan Gray, John, Ingo, Tim, Gustavo, Jeff Burrell, Joe

The headline demo was Ingo's: he showed a hand-built Z80 backplane and breakout board (with a homemade diagnostic board using seven-segment displays for the address and data buses and bar-graph LEDs for the control lines) attached to a ZX81 motherboard. He'd desoldered TR2 to break out the clock line, added a 555 timer with a pot for an adjustable clock and a pushbutton for manual single-stepping, and walked the group through the opening bytes of the ZX81 ROM — the OUT to port FD that sets up NMI, then the LD BC,7FFF for the RAM-top check. He outlined next steps: a hardware breakpoint built from DIP switches that asserts /WAIT once the address bus matches, and proper buffering to stop the bus loading down when he flips between manual and free-running clocks. Gustavo and others noted that only the CMOS Z80 spec allows fully static clocking; original NMOS parts have a 200 kHz minimum.

Ryan reported on a fresh round of fixes to z-makebas (the C-based BASIC-to-TAP tool). The latest bug was in the Spectrum-Next number-parsing path: after the parser was extended to start numbers at a leading + or -, the test for a preceding $ (inline hex) or @ (inline binary) misfired on expressions like CODE "X"-32, causing the floating-point bytes to be skipped. He's pushing fixes upstream to Chris's repo (which Tim's fork tracks), wants to add reference TAP binaries and a regression test for each escape mode, and noted the ZX81 path had clearly been rotting from lack of users. A short side-discussion covered command-line switches to force a target mode (ZX81 / Spectrum / Next) rather than relying on syntax inference.

David explained his Larkin-disk imaging pipeline for someone wanting to move Larkin floppies to an emulator: a Greaseweasel board (USB on one side, 34-pin floppy connector on the other, with external drive power) reads the raw flux, and a Python program David wrote parses the Larkin format — using the original Larkin disk-utility documentation — and emits per-file TAP images. He offered to image disks for members who can't run it themselves. The group discussed the usual cleanup of RAND USR 100 / PRINT #4 Larkin-isms when porting back to stock BASIC, and a TAP-to-BASIC-listing utility to make that editing less tedious. Gustavo gave a status update on his CP/M-for-the-2068 port: the system runs on real hardware with an FDD3 for storage, but he's stuck without an emulator that models a RAM cartridge in the dock bank, so he can't step-debug the CBIOS; 80-column software rendering is the other bottleneck, and he confirmed CP/M really needs a full terminal-style keyboard (the reason Timex built the TS3000 terminal).

Shorter items: Paul showed a TS1000 he's mod'd with twin 2764 EPROMs replacing both ROMs, and is bundling a recased TS1000 + extras to pass on to someone introducing a child to computing. Joe asked for help recovering a dead ROM from a Micronix/CompuSA "FIZ" floppy-controller card for the ZX81 — two pins read as floating on the programmer; suggestions included grounding the bad pins and reading the rest in four combinations to reconstruct by disassembly, and an Irish lab that will delid and probe-read a die for ~$300. Adam is waiting on updated programmer software from Jeff Burrell for the FPGA-2068 board so the Albuquerque members can flash it in person. Carl showed a Bytedelight-built Spectrum diagnostic cartridge (kit + 3D-printed case) and noted his twister board should let it run on a 2068. Other passing items: the ZXpand replaces the ZX81 ROM with its own and remaps LLIST/LPRINT for SD-card commands; the Tandy Model 100 "backpack" (a TPDD-protocol SD adapter on a single AAA) works on Model 100/102/200 and Epson HX-20, and BBC BASIC programs exist that should let a Z88 talk to the same TPDD protocol. Steve North will speak at the September meeting; next meeting is Sunday August 20 at 2pm Eastern.

Highlights

  • Ingo demoed a homebuilt Z80 diagnostic rig single-stepping the ZX81 ROM via a desoldered TR2 and a 555 clock
  • Ryan walked through z-makebas fixes for Spectrum-Next inline-hex/binary number parsing and floating-point byte emission
  • David described his Greaseweasel + Python pipeline that turns raw Larkin disks into per-file TAP images
  • Gustavo: CP/M-for-2068 runs on real hardware but is blocked by no emulator that models a dock-bank RAM cartridge
  • Joe is hunting a ROM image for the Micronix/CompuSA FIZ floppy-controller card — two pins on his are dead

Projects discussed: FPGA "Advanced 2068" Topics: TS1000 / ZX81 / TS1500 · Machine code, Z80 & assemblers · Hardware projects & new boards · Disk & floppy systems (Larken, AERCO, Oliger) · CP/M · Utilities & applications

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