Skip to content

Design review — control paths and I/O (sequencer logic/memory, IO, bus tester, backplane, video) — 2026-09-21

Section 3 of the 2026-09-21 control-paths and I/O review: the IO card v1.1, findings 3.1-3.4. The other cards' sections ("above" and "below" in the original), the cross-card notes and the summary: hardware/reviews/control-io-2026-09-21.md.

Second pass after the mechanical one in mechanical-2026-09-21.md. Read from the KiCad netlists (reports/netlist.net, pin types from the Eagle symbols), chip values from the Eagle boards, the firmware that drives the ATmegas (embedded/bus-tester/bus-driver, embedded/sequencer-card/sequencer4) and the microcode generator (firmware/microcode/ucode-generator2). Gate pin-outs were checked against the symbol pin functions (I0/I1/O for 7400/08/32, O/I/I for 7402; 74175/74/192/244/374 pin functions as printed in the netlist).

Every finding: the trace (ref.pin -> net -> ref.pin, chip types), the failure scenario, severity (HIGH = stops the card / damages parts, MED = unreliable or latent, LOW = housekeeping), and how to confirm it on the bench. Already-known items (BACKLOG.md, video README) are not repeated except where a new detail changes them. Report only; no redesign.

Scripts used (not committed): a netlist dumper (parts + nets with pin function/type), an SV1/SV2 pin-by-pin comparator, a bus-driver-table-vs-tester-wiring check and a microcode-bit-vs-RAM-wiring check; the results of the three checks are in the "checked, no issue" lists below.


3. IO card v1.1 (hardware/cards/io/kicad/v1.1)

3.1 Correction to v1.1-mechanical-2026-09-21.md — XTAL2 floating is correct

Y1 (ECS-2100AX oscillator) p5 OUTPUT -> N$31 -> IC1 XR16C550 XTAL1 p16; XTAL2 p17 (N$61) open, which is the datasheet connection for an external clock. Not a fault. The Eagle value "ECS-2100AX-200" is the package name, not a frequency; firmware/monitor/monitor.asm:74 programs divisor 3 for 38400 baud, which implies a 1.8432 MHz oscillator (20 MHz would give 416 kbaud). Confirm: read the can; put the frequency in the BOM.

3.2 LOW — Strobe polarity is fine, but every strobe passes through an open-collector inverter with a pull-up of unknown value

-IO-RD (B25) -> IC8 74LS06 p1->p2 = IO-RD; -IO-WR (B26) -> p3->p4 = IO-WR; -RESET (C30) -> p5->p6 = RESET -> IC1 p35 (active-high reset, correct) and -> p9->p8 = -B-RESET -> CLR of IC4/IC9/IC10 74LS273; -IO-DATASEL -> p11->p10 = IO-SELDATA -> IC1 CS0 p12. Pull-ups are RN2 (value blank in the Eagle board). The UART's IOR/IOW (p22/p19, active high) are used with -IOR/-IOW (p21/p18) tied to VCC — correct. The latching edges (273 clocks N$40/N$41/N$75 = NAND of IO-WR with the select bits; LCD E = N$30; 16550 IOW trailing edge) all come from the fast falling edge of the OC outputs, so the design is sound; the leading edges rise through RN2 and, if RN2 is 10k, take a few hundred ns through the LS10/LS00 inputs (IC6, IC7, IC11), where a slow edge can produce a double transition on the LCD E line (N$30) = a double write. Confirm: scope IC8 p4 rise time; watch for doubled LCD characters.

3.3 LOW — The port decoder is address-only and the device-select bits have no hardware exclusivity

IC5 74LS138: A/B/C = IO-ADDR0..2, G1/G2A from the IO-ADDR-HL0 jumper (IO-ADDR3 or fixed), G2B = GND; -IO-SEL0..7 go only to the two select headers (DATA-ADDR0, IO-ADDR0) that pick the data port and the control port. Every consumer ANDs the select with IO-RD/IO-WR (IC6/IC7/IC11), so the lack of -VMA/-BUS-EN qualification (the V1.2 idea in Notes.md) has no functional effect; with the pipeline floating, -IO-RD/-IO-WR read high at IC8 and the strobes are inactive. The control latch IC10 74LS273 (Q1 SWITCH-LED, Q2 LCD-ENABLE, Q3 LCD-REGISTER, Q4–Q6 UART-A0..2, Q7 UART-CS, Q8 TIL311) lets software set SWITCH-LED and UART-CS together, in which case IC3 74LS244 (enable IC6 p6) and the UART both drive DATA0–7 on a data-port read. Software rule; note it in firmware/abi/.

3.4 LOW — LCD backlight straight across the rail

X2 HD44780 p15 (A) = VCC, p16 (K) = GND with no series resistor: relies on the module having its own. Confirm on the module.

Checked, no issue (IO)

  • UART register select vs chip select: CS0 = IO-SELDATA (data port decode), CS1 = UART-CS (control latch bit), -CS2 = GND, -AS = GND; A0–A2 from the control latch; every register access is "write control byte, then read/write data port" — reachable, no aliasing. RCLK p9 = -BAUDOUT p15 (ICLK). CTS/DSR/CD/RI tied high.
  • -INT (C13) = IC8 p12 open collector with RN2 pull-up, from the UART INT (p30) through the INT0 jumper — the only -INT source in the machine; the tester keeps -INT as input; the logic card adds an optional 10k.
  • Data bus: no buffer at all (Notes.md V1.2 idea); the only drivers are the UART (CS·IOR), IC3 (read of switches) and nothing else — LCD R/W = GND so the LCD never drives; 273s and TIL311 latches are loads only.
  • 74LS273 latches clock on the trailing edge of IO-WR while the data source is still asserted (generator asserts the source one line before the strobe).
  • TIL311 HI0/LO0: latch strobe p5 = GND (transparent), blanking p8 = GND (not blanked), as Notes.md V1.1 says.
  • OUT (C27) drives only an LED through R1; IN (C26) is driven only by the IN0 switch (hard VCC/GND) through the INPUT0 jumper — fine while nothing else drives IN (tester: input; logic/register/video: unconnected).
  • Spare gates IC7C, IC11B, IC11C: inputs grounded.

Last updated 26 October 2026