Why an assembly cell costs 400 hours — and why it shouldn't
Every module in an assembly cell knows something important about itself. Almost none of it is written anywhere a tool can read. That single gap is where the 400 hours go.
An assembly cell is bespoke because every module in it is opaque.
The screwdriver knows its torque curve. The engineer knows its approach vector. The PLC knows its I/O map. The robot program knows its tool centre point. And none of that is written anywhere a tool can read it.
So a human retypes it into four places. The cell takes 400 hours, and every one of those four copies is a place a mistake can hide until commissioning.
The change we are making
Cellwright says the module ships with a manifest — a machine-readable description of what the module is, what it can do, and what it refuses to do.
The manifest is the single source of truth. Everything downstream is generated from it:
- the collision scene
- the robot program
- the PLC sequence
- the tag list
- the cycle-time estimate
If that holds, the same cell takes 40 hours. If it does not, this is just a nice open frame design — so proving it is job number one, and it is provable on a laptop.
Refusal is a feature
Ask the generator for 6.0 N·m from a tool rated to 5.0 N·m and it refuses to emit a program. Not a warning buried in a commissioning report. A refusal, before anyone cuts steel.
Where we are
Pre-alpha, and we are saying so plainly. The manifest schema validates. The generator is being built. No hardware has been built yet.
This blog is the record of that work as it happens.