Skip to content

Design review, datapath cards (control-path pass) — 2026-09-21

Two cards of the 2026-09-21 datapath review, reviewed together: the Mem Switch 1.1 and the Mem Register 1.0, findings B1-B2 (the Mem Register's file points here). The other cards, the system-level items S1-S3 and the summary: hardware/reviews/datapath-2026-09-21.md.

Second pass after mechanical-2026-09-21.md (mechanical). Cards: ALU v3.2, Index Register 1.1, Memory v1.3, Mem Switch 1.1, Mem Register 1.0. Method: every select / enable / strobe / clock / reset net was traced pin-to-pin from kicad/<ver>/reports/netlist.net (pin names and types from the Eagle symbols) with chip types from the Eagle .brd, and checked against the way the sequencer microcode actually drives the bus (firmware/microcode/ucode-generator2/), because several hardware choices only work under a microcode convention. Signal polarity: - prefix = active low. Gate delays quoted are LS-TTL typicals from the datasheets in docs/datasheets/.

Severity: HIGH = would stop the card working or damage parts; MED = unreliable / depends on a convention that is not enforced by hardware; LOW = housekeeping or documentation.


Mem Switch 1.1 (hardware/cards/mem-switch/kicad/v1.1/) and Mem Register 1.0 (hardware/cards/mem-register/kicad/v1.0/)

B1 — MED (configuration): both cards, and low RAM, overlap — see M4

Mem Switch decodes $0000-$000F and Mem Register $0010-$001F (SW1 sets ADDR4-7 against IC4 74LS85 with RN1 pull-ups; IC4 A=B_I(3) = V1 (M74HC4078) pin 13 = NOR(ADDR8-15), so only page 0). Neither can coexist with the memory card's undisableable low RAM.

B2 — LOW: Mem Switch design files fit 74LS244, the notes say 74240

Notes.md for 1.1 lists "Change to 74240 (inverted output)" and the README repeats it, but the v1.1 schematic/board show IC5-IC20 = 74LS244N with the switches pulling to GND against RN2-RN17 pull-ups (switch closed = 0). If 74LS240s were ever fitted, every switch bit reads inverted. Document whichever is true.

Checked, no issue (bring-up cards)

  • Read qualification: IC1 (74ALS27) N$19 = NOR(-VMA, -BUS-EN, -MEM-RD), N$82 = NOT N$19 -> IC2/IC3 (74LS138) G2A; IC2 G2B = ADDR3, IC3 G2B = NOT ADDR3 (IC1.6); G1 = comparator match. Exactly one of SW0-15 / RD0-15 low per read; each enables one 74LS244 (switch card) or one 74LS374 OC (register card) onto DATA0-7. DATA8-15 are not driven.
  • Write (Mem Register): IC7 (74ALS27) N$2 = NOR(-VMA, -BUS-EN, -MEM-WR), N$4 = NOT N$2 -> IC5/IC6 138 G2A; WRn -> 374 CLK. The 374 loads on the rising edge of WRn, i.e. at the end of the strobe (trailing edge); the microcode holds the source one line after -MEM-WR rises (main.c:243-247) and the address is stable a line before, so the decoder cannot produce a second rising edge on another WRn during the write.
  • No reset inputs on either card; nothing to mismatch.

Last updated 26 October 2026