Spencer Morris スペンサー・モリス← the streetLet's talk
SM27

Bench Arm

CASE STUDY · 2 MIN
Execution / case study

Where you put the encoder beats the gearbox

I designed a desk robot for the fiddly parts of board bring-up, and built the whole work cell in simulation first to find out which parts decide whether a probe lands on a pad.

ExecutionSimulated work cell, three arm revisionsNov 2026Design study: CAD, MuJoCo simulation, report
MuJoCo render of the v3 work-cell run (out/full.mp4, 0:41 to 0:53).
median landing error, hobby servo, motor-side vs output-side encoder
tip stiffness, first mock-up vs stiffened arm
RMS position error before and after dot-grid calibration
arm revisions; v3 adds a six-tool dock, a board flip frame, a preheater and feeders
1 Bench Arm

Sensing decides precision

params.py holds every dimension and assumption; build123d generates the arm and the cell, and MuJoCo runs it with lumped spring joints. A 100-trial precision study crossed four drivetrains (hobby servo, planetary, two harmonic drives) with three sensing schemes. A hobby servo read at the motor landed a median 589 µm off and hit a test pad 19 percent of the time. The same servo with an output-side encoder landed at 15 µm and hit every pad, including QFP-pitch targets. Good harmonic drives with motor sensing still sat around 50 µm. The first mock-up deflected 20 mm under a 5 N tip load; the stiffened v2 deflected 90 µm. A wrist camera reading a dot grid cut RMS error from 140 µm to 6 µm.

The v3 cell adds a tool dock with six tools (paste, vacuum, gripper, probe, iron, laser), an upward camera, a board flip frame, a preheater and feeders, and runs a full probe, place, insert, wire and solder sequence. A separate study puts the same job on a wall-rail gantry over a 72 by 24 inch desk. Everything is modelled, nothing is measured, and the docs say which figures are assumed.

Pythonbuild123dMuJoCoNumPySciPy
Sources and method

Written from the project's code, documentation and commit history, reviewed in November 2026. Business figures come from me. Clips and screenshots are labelled with where they came from; none show customer data.

  • bench-arm/README.md
  • bench-arm/GANTRY.md
  • bench-arm/params.py
  • bench-arm/out/precision.json
  • bench-arm/out/stiffness_v1.json
  • bench-arm/out/stiffness_v2.json
  • bench-arm/out/calib.json