Status¶

Last reviewed 3 September 2026.

✅ Settled¶


The retrofit boundary	Eight connectors on the Housing's front panel. No field wiring disturbed, no field wire re-terminated
Servo control mode	Analog velocity. Proved from the drawing, not the drive panel — the pulse inputs are not wired and there are not enough spare pins or conductors to carry them. The drive settings stay exactly as they are
Homing	The drives are not self-homing; the controller homes them, switch plus encoder index. Reproduces in LinuxCNC with no extra wiring
Mesa hardware	7I76EU + 7I77U + 7I84U, all bought. 96 inputs and 48 outputs against 41 and 41 required
The I/O map	All 82 points: address → module terminal → wire → cabinet terminal → connector pin → device
The Housing connectors	Every one, pin by pin, with part numbers
Locating-stop interlock	Spindle inhibited until the stops are commanded down and a dwell has elapsed. No sensors, by decision


Control logic	Every function mapped to its I/O and to the LinuxCNC mechanism. Next: mine the original PLC source function by function
Simulation config	Written, not yet run
Connector order	Panel-mount D-subs, ribbon headers and cable specified. Buying

These three can be answered in a single session and all need the original controller still in place. They cannot be answered after strip-out.


Ribbon way ↔ I/O address	Which of the 26 ways carries which point. Likely ways 1–24 in order with 25–26 the module supply, but unproven. Test at four ways per ribbon, not all 24
Ribbon logic voltage and polarity	Sets the Mesa field-power domains
Where the I/O modules get logic power	Two spare ways on a 26-way suggests the ribbon carries it


`IF.21`: index sensor or change piston?	The symbolic map and the connector schedule disagree about the same wire. Settle before writing the ATC — if it is really a spindle index, tool-change orientation is achievable; if not, it is not
Drive scaling parameters	The PG divider ratio and speed reference gain still have to be read off the drive panels. Do not assume catalogue defaults
The rope switch identity	A dashed device in the under-table box is the strong candidate. Confirm physically
Should a raised stop also inhibit traverse?	Not just the spindle. One more logic block if wanted


Sensors on the locating stops	Decided against. The interlock proves the command was given and time allowed, not that the pin moved — the same exposure the machine has today
Cable restraint inside the Housing	Not needed for short internal runs
Replacing the safety chain with software	The enable rail, the drive-OK relays, the overtravel path and the Z brake stay hardwired. Mesa only monitors
Touching the drives	They work, they are correctly configured, and their looms are professionally made assemblies