PROJECT DESIGN REVIEW — MINUTES OF MEETING
Project: Rotary assembly station, 5-station index table
Review subject: Geneva mechanism and index drive selection
Date: 14 March, 09:30 to 11:00, meeting room 2
Document ref: DMR-2026-0314-RT04
Attendees
| Name | Role |
|---|---|
| A. Rostov | Lead mechanism designer (chair) |
| M. Klein | Senior machinist, workshop |
| P. Okafor | Controls / servo engineer |
| J. Tanaka | Quality / metrology |
| D. Silva | Procurement |
1. Purpose of the review
The table indexes five fixtures through a screw-driving, pressing, inspection and unload sequence, cycling 1.2 s per index. The draft design uses an 8-slot Geneva mechanism on the drive shaft with a 0.5 kW induction motor and a friction clutch. The review was called because the index repeats at the metrology station drifted past the 0.05 mm target during prototype trials. The meeting was held before the first production batch, so changing the drive family was still a real option.
2. Review comment CP-01: indexing accuracy of an 8-slot Geneva
A. Rostov walked the table through the geometry. An 8-slot Geneva gives 45 degrees of table rotation per input revolution, with the star wheel locking on the curved arcs between indexes. The lock is geometric, not friction, which is why a Geneva holds its position without a brake. But the entering and exiting clearance in the slot and roller matters. On the prototype we measured a dwell drift of roughly 0.06 to 0.09 mm at the 340 mm pitch radius, against a target of 0.05 mm. The drift comes from the roller-to-slot clearance growing from the 0.05 mm drawing value to about 0.12 mm after the first week, because the roller was a plain bearing pin and the slot edges were not hardened.
M. Klein added the workshop view without being asked: the slot was cut with a mild steel blank and left soft, so the roller hammered the flanks, and the wear was visible as a bright line in an hour of running. He had flagged it on the trial report. The chair noted the comment was logged, not shouted down.
3. Review comment CP-02: the drive motor and clutch pair
CP-02 questioned the 0.5 kW motor plus friction clutch combination. P. Okafor measured the torque transient during indexing and saw a peak near 6.5 N-m for about 90 ms as the roller engaged, which is acceptable for the motor, but the clutch was set to slip above 4 N-m to protect the fixture, and the slip appeared on every index, heating the friction face. The clutch disc felt warm at the end of a two-hour run, and the temperature was not quantified because nobody had put a thermocouple on it. The controls engineer asked whether the motor should simply be switched off during the dwell instead of running against a clutch. A. Rostov replied that the Geneva still needs the input to stop at the right angular position, so a clutch or a brake does that job.
4. Review comment CP-03: star wheel locking accuracy and a plain bearing pin
CP-03 came from J. Tanaka, and it was the comment that changed the design. He asked for the measured angular deviation between indexes rather than the radius drift, because the fixture on the table sees position error, not radius error. The team had only recorded the 0.06 to 0.09 mm radius drift, which is not the same quantity. When the drift was converted, the angular error came out near 0.032 degrees, which at the 340 mm pitch radius translates to roughly 0.19 mm of tangential position error at the fixture pad. That is four times the 0.05 mm target, and it explained why the screw-driving head missed its hole centre on every fifth part.
M. Klein pointed out that the entering clearance in the slot, not the lock, governs that error, and that the fix is to move the wear surface out of the mechanism. He proposed hardening the slot flanks and replacing the plain pin with a needle roller stud. P. Okafor noted the needle stud would raise friction during the short engagement, but the engagement lasts less than 100 ms, so the extra drag is irrelevant to the cycle time.
5. Review comment CP-04: the 0.5 kW motor and index torque margin
P. Okafor had prepared a torque balance. The table carries four filled fixtures at 2.1 kg each plus the empty one, the centre disc weighs 9.4 kg, and the worst kinematic case is the first 30 ms of roller entry, where the linear acceleration of the table rim reaches about 3.4 m/s-squared. The resulting peak equivalent load torque was calculated at 3.2 N-m. Chosen motor: 0.5 kW, 4-pole, 1420 rpm, with a 30:1 worm gearbox, giving 47.3 rpm at the Geneva input and a torque capability of about 76 N-m at the same speed. The running torque draw therefore sits near 6 percent of capability. The margin is there; nobody disputed the motor. What the review did dispute was the clutch, and the recommendation was to delete the friction clutch and let the motor drive a cam-indexed Geneva directly through the worm gearbox, using the motor brake at the end of each index.
6. Review comment CP-05: cycle time budget and index timing
J. Tanaka asked for a timing sheet. The proposed cycle breaks down as: table index 1.2 s, screw-driving 1.6 s, pressing 0.9 s, inspection 1.1 s, unload 0.8 s, giving a station cycle of 5.6 s and an output of 642 parts per hour at 100 percent efficiency. Two points stood out. First, the 1.2 s index is faster than the motor is comfortable with at full inertia, so the ramp-up and ramp-down must come from the motor drive, not from ramming the roller into the slot. Second, the inspection station is the bottleneck with 1.1 s, so a 0.15 s saving there lifts the whole line. D. Silva asked whether the saving should come from a faster camera trigger. P. Okafor answered that the camera itself is not the limit, the conveyor between stations is, and that moving the trigger earlier would only buy noise.
7. Review comment CP-06: procurement questions on the bespoke star wheel
D. Silva raised cost and lead time. A hardened and ground 8-slot star wheel in 1.2379 tool steel, with the flanks induction-hardened to 58 to 60 HRC, came back quoted at 18 working days and 420 euros per set from the current milling supplier, versus a hardened and ground off-the-shelf Geneva unit at 640 euros and 9 working days from the catalogue. The workshop could cut its own blank in 1.2379, send it for vacuum hardening, and finish-grind the flanks on the existing cylindrical grinder, at a material cost of roughly 90 euros and a two-week total delay. M. Klein stated the shop could hold the 0.02 mm slot width tolerance on the grinder, and that claim was accepted because the same grinder already holds 0.01 mm on the gauge blocks. The review therefore agreed to make the wheel in-house and to buy only the needle roller studs and the hardened pins.
8. Disagreement that had to be resolved: clutch or no clutch
The one real dispute of the morning was whether to keep a friction element between motor and Geneva. A. Rostov wanted to keep a clutch as belt-and-braces against a jam. P. Okafor argued that a jam in a Geneva locks the mechanism solid, and a slipping clutch then overheats in seconds while the motor keeps pushing, which is worse than a shear pin that breaks once. M. Klein supplied the practical evidence: in the previous machine the clutch slip had to be re-set after every two weeks because the friction face wore and the slip torque drifted between 3.8 and 4.6 N-m. The review took a vote and resolved to remove the clutch, add a shear pin rated for 8.5 N-m in the drive train, and rely on the motor brake for controlled stops.
9. Risks and mitigation before the batch
Three risks were logged. Risk 1: the in-house ground slot flanks could come out with a burr that jams the roller; mitigation is to break the edges with a 0.3 mm chamfer and run a hand inspection on the first fifty cycles. Risk 2: the motor brake engages every 1.2 s and could overheat the brake coil; mitigation is to measure the brake winding temperature after two hours of running and fit a thermal switch if it exceeds 60 degrees C. Risk 3: the needle roller studs are only stocked by one supplier; mitigation is to order two spares per machine and keep a warehouse note. Each risk got an owner in the action list.
10. Actions and owners
| ID | Action | Owner | Deadline |
|---|---|---|---|
| ACT-1 | Machine in-house 8-slot star wheel, grind flanks to 0.02 mm width | M. Klein | 28 March |
| ACT-2 | Replace plain pins with needle roller studs and hardened pins | M. Klein | 28 March |
| ACT-3 | Delete friction clutch, fit 8.5 N-m shear pin, rewire motor brake | P. Okafor | 31 March |
| ACT-4 | Add angular position transducer check on station 3, log every index | J. Tanaka | 1 April |
| ACT-5 | Order two spare needle roller stud sets | D. Silva | 19 March |
11. Decision summary
Adopt the Geneva index table with an in-house hardened star wheel. Delete the friction clutch. Introduce a shear pin and a motor brake. Keep the 0.5 kW motor and the 30:1 worm gearbox. Add an index-by-index position check at the metrology station. The 0.05 mm target stays, but it is now measured as tangential position at the fixture pad, not as radius drift of the disc.
12. Post-meeting addendum, 18 March
The star wheel was finished ahead of schedule and ran on the bench. Slot width measured 0.019 to 0.021 mm, roller stud runout 0.008 mm. Index-by-index position at the fixture pad came out between 0.03 and 0.045 mm over two hundred cycles, inside the 0.05 mm target. The shear pin has not blown, and the brake winding stayed below 48 degrees C over a two-hour soak. The minutes are closed. The machine went into the batch, and the screw-driving head has not missed a hole centre since.
Glossary
Geneva mechanism: an intermittent drive where a driving roller engages radial slots of a star wheel to make a fixed fraction of a revolution. Face width: the axial length of the slot that the roller rides in. Dwell: the phase where the star wheel is locked on the convex arc. Pitch radius: the radius of the fixture-holding circle on the table. Index: one fixed angular step of the table.
Appendix. Dimensions that the review checked
Because the minutes would not be complete without the numbers the debate actually turned on, the key checked dimensions are listed here as they were written on the whiteboard. Star wheel outside diameter 260 mm, eight radial slots, slot width 12 mm nominal with 0.02 mm grinding tolerance, roller stud diameter 12 mm with rolling element, centre distance between driving crank and star wheel 122.7 mm, locking arc radius 104 mm with 0.15 mm dwell engagement, input shaft speed 47.3 rpm, index time 1.2 s, dwell time per station 4.4 s of the 5.6 s cycle. The fixture pad circle sits at 340 mm pitch radius and carries up to 2.1 kg per fixture plus a 9.4 kg centre disc. The shear pin is sized to fail between 8.0 and 8.5 N-m so that it breaks before the star wheel slot flank yields, and the motor brake is set to hold the table against a 12 N-m disturbance, which is the worst push a manual unload operator can apply in practice.
The review closed with the usual note that the minutes are only worth their paper when the actions get done, and with the three dozen cycles run in the workshop before the addendum, the machine earned its sign-off. Any future change to the index timing is to be raised at the next review, and the author of this document reminds the reader that a Geneva that indexes cleanly still needs its oiled roller and its honest clearance, because geometry without maintenance is just a drawing.
