WORK ORDER — Maintenance and Improvement Record
Work order no: WO-A-2026-04 Date issued: 2026-09-11 Category: Pallet transfer station
Machine / cell: Pallet Transfer Station PT-2 serving the machining centre cluster
Status: Closed Engineer: R. Feld (automation) Total labour: 6.5 hours Parts cost: EUR 118.00
Reason for work: machined feature drift on the third pallet position of every transfer batch
1. Fault description as logged by the shift team
Wednesday 08:15, operator record: the machining centre posts a drifting feature on parts loaded from pallet position three, the bored hole comes out 0.09 mm off its nominal position on the flange, while positions one and two hold their tolerance of 0.03 mm across the same batch, and the shift has re-set the machine offset three times in a week to chase a drift that comes back as soon as the part count climbs again. The pallet system is a simple lift-and-shift transfer that carries a 200 by 200 mm fixture pallet on a single rail, and each pallet indexes into the machine on two locating pins, one round and one diamond, with two clamp cylinders holding it down.
The pallet train has run about 40,000 transfers, and the machine operators assumed the drift was the machine, because the hole lands correctly at the start of a batch and drifts only as the batch runs, which smells like thermal growth of the spindle or the casting. But a thermal problem would drift all three pallet positions together, and here only position three moves, which is the first reason I did not touch the machine offsets when I got the call on Thursday. The second reason is the number of cycles, 40,000 pallet indexes on a pair of pins is a lot of wear cycles for pins that are rarely measured, and a locating pin that wears on one side lets the pallet sit a little differently every time it loads, which is exactly the kind of slow repeatability leak that a batch check sees as a drift.
Evening note to the file: the machine does not know which pallet position it is holding, the pin set does, and when only one position drifts the pins are the prime suspect before the spindle earns a minute of the maintenance budget.
2. On-site inspection record (Day 1)
I pulled the pallet for position three off the station and measured the repeatability directly, loading the pallet onto the pins twenty times and clocking the pallet face with a dial indicator at the three datum points, and the readings spread over 0.10 mm in the horizontal plane while the reference pallet, the shop’s master, spread over 0.02 mm. The spread was one-directional, all the scatter lined up along the axis of the round pin, which is the signature of clearance growing in the pin-and-bush pair rather than a bent pallet or a misaligned station. I measured the round locating pin with a micrometer at the bush contact zone and it read 9.93 mm against the drawing 10.00 for the pin end, a 0.07 mm loss, and the bush measured 10.11 mm against its drawing 10.02, so the pair had grown from the design 0.02 mm clearance to a sloppy 0.18 mm, nine times the intended fit. The diamond pin told the same story with a different shape, its locating shoulder was worn flat on the loaded flank by 0.06 mm, taking away the precise lateral register it is supposed to force.
The wear pattern pointed at the transfer motion rather than the machining: the lift-and-shift cylinder lowers the pallet onto the pins with a slight forward drift because the pallet lands while the shift carriage still has residual momentum, and every landing drags the pin into the bush the same way, polishing one flank of the round pin and one flank of the diamond. I checked the landing behaviour with a high-speed recording of five cycles, and the pallet slid on the pin by just under 1 mm while it bedded, so the pins were never sitting free in their bushes, they were being wiped by the landing every single transfer, 40,000 small wipes that took a 0.02 mm fit to a 0.18 mm slop.
Evening note: a pallet that repeats at 0.02 mm on the master and 0.10 mm on the production pallet is not a machine problem, the machine gets what the pins give it, and the pins gave position three a 0.18 mm clearance that the machined hole faithfully copied into the part.
3. Corrective action taken
I replaced the round pin, the diamond pin and both bushes on the position three mount, pressing the new bushes in with an arbor press and checking the press fit, and I lapped the new pin-bush pair to the drawing clearance of 0.02 mm so the pallet registers on the pins the way the designer intended rather than free in a sloppy bore. The pin material is the station spec, alloy steel case-hardened to 60 HRC and ground, so the new set restores the wear resistance the original had on day one. Before I refitted the pallet I did the job the previous line lacked: I slowed the landing sequence in the station controller, trimming the shift cylinder deceleration ramp so the pallet now settles onto the pins with essentially no forward glide, measured at less than 0.1 mm of landing slide instead of the old 1 mm, because replacing pins without fixing the landing would only buy a few thousand cycles before the same one-sided wear came back.
I re-ran the twenty-cycle repeatability test on the refitted position three and the pallet face now spreads over 0.02 mm, matching the master pallet, and I machined a trial part through the station to prove the feature lands where the drawing says. The clamp cylinders got a careful check while I was in the station, one cylinder was down to 3.8 bar against its 5.0 bar spec, so I replaced its seal kit and re-set the pressure, because a weak clamp lets the pallet lift a hair on the pin during machining, which is the same repeatability leak from a different door. The station control change is logged in the programme library with the deceleration values, so a future technician does not re-introduce the old landing while troubleshooting something else.
| Item | Before | After | Spec |
|---|---|---|---|
| Round pin OD at contact | 9.93 mm | 10.00 mm | 10.00 |
| Bush ID | 10.11 mm | 10.02 mm | 10.02 |
| Pin-bush clearance | 0.18 mm | 0.02 mm | 0.02 |
| Pallet repeatability (20 cycles) | 0.10 mm | 0.02 mm | ≤0.03 |
| Landing slide | ≈1 mm | <0.1 mm | <0.2 |
The service interval note: the station manual lists the pins for inspection at major overhaul only, so a 40,000 cycle station ran its pins to 0.18 mm slop with no audit trip along the way, and the new procedure sheet now checks pin clearance at every 8,000 transfers with a simple pin gauge.
4. Acceptance result after the repair
I ran the station through a full production batch of 60 parts across all three pallet positions and measured the bored hole position on every tenth part, and position three now holds the feature within 0.02 mm of nominal right alongside positions one and two, with the batch spread at 0.028 mm total across the whole run against the 0.05 mm drawing tolerance. The machine offset was not touched during the whole repair, and it no longer has to be, because the pins give every pallet the same register and the machine can set its zero once. Quality signed the batch, the drift complaint is closed, and the station has run a further week of production with no re-occurrence, so I closed the work order and moved the pin clearance check onto the 8,000 cycle service sheet where it belongs.
5. Hours and parts list
- Labour — 6.5 h total: repeatability and wear inspection 2.0 h, pin and bush replacement 1.5 h, landing sequence reprogram and cycle test 1.5 h, clamp seal and release batch 1.5 h
- Parts — round locating pin, diamond pin, two bushes EUR 74.00, clamp cylinder seal kit EUR 22.00, pin gauge and shims EUR 12.00, high-speed recording floor stock EUR 10.00; total EUR 118.00
- Scrapped parts in the drift window — 3 flanges re-worked at the machining centre, no part scrapped outright
6. Retrospective notes for the next engineer
A feature that drifts on only one pallet position while the machine repeats its offset check is a locating problem, and the fastest conviction is the twenty-cycle repeatability test because it separates the machine from the pins in half an hour. Nine times the design clearance in the pin-bush pair sounds dramatic, but it grew one 0.001 mm wipe at a time from a landing sequence that dropped the pallet with a 1 mm sideways glide on every cycle, so the real fix was the deceleration ramp as much as the new pins. Check the clamp cylinders while the pallet is out, a 3.8 bar clamp is a repeatability leak waiting to happen, and put the pin clearance on a service interval short enough to catch a 0.02 mm fit before it quadruples, because a pin gauge is twenty seconds on a service sheet and a drifted batch is three re-worked flanges and a week of offset chasing.
Attachments: twenty-cycle repeatability plots for master and production pallets, pin and bush measurement sheet, landing sequence recording stills, release batch report (filed with WO-A-2026-04).
Glossary — locating pins: round and diamond pins registering a pallet in one defined position; diamond pin: flattened pin allowing expansion on one axis while fixing the other; pallet repeatability: how consistently a pallet returns to the same position on re-load; case-hardened: steel with a hard wearing surface and a tough core; press fit: a shaft held in a bore by interference.
Field note — the 3-2-1 locating arithmetic and why one worn pin sat so loudly
The geometry of this station is worth re-stating because it explains why position three drifted while the other two held. A pallet is located by the 3-2-1 rule, three points set the plane, two points set the position and one point sets the rotation, and the pin pair implements the last three: the round pin fixes both X and Y at one corner, the diamond pin fixes only one of them at the far end and frees the other so the pallet does not bind on thermal or tolerance growth. Wear on the round pin grows clearance in both X and Y at that corner, but the machined feature that drifted was the bore position relative to the pallet datum, so the 0.18 mm slop at the round pin showed up almost one-to-one at the bore, while the diamond pin wear hinged the pallet rotation about the round pin, which moved the far corner by a smaller and less noticeable amount. That asymmetry is why one pallet position developed a visible drift and the other two looked innocent, they were wearing at the same rate but the round pin corner is where the datum lives. The lesson for design as much as for maintenance: pin pairs should be specified with their wear-out regime in the manual, the diamond pin needs its lateral register protected and the round pin needs its clearance checked, because both wear one-sided under a landing slide and one-sided wear defeats the whole purpose of a 3-2-1 register. On the repair side, the cheap repeatability insurance is a pin gauge ride in the toolbox and a service interval short enough to catch the pair before the part really moves, and the one pressure check on the clamp cylinders each service, because a pallet that is registered by two pins and held by one tired clamp is only as repeatable as its weakest restraint.
