Vacuum chamber plate: 0.008 mm flatness that survives anodising

The problem

A semiconductor equipment OEM needed a vacuum chamber plate in 6061-T6: 420 × 380 × 12 mm, with a flatness call-out of 0.008 mm across the full face and a sealing groove on the reverse. The first supplier delivered parts that measured 0.006 mm off the machine and 0.03 mm after anodising — outside the call-out, and enough to break the seal.

The obvious answer, lapping after anodising, was rejected: it would cut through the coating on the sealing land. The part had to leave the anodiser flat, which meant the residual stress had to be gone before the coating went on.

What we changed

The fix was sequence, not equipment. Three things:

• Rough, then relax, then finish. The plate is roughed to 0.5 mm stock, stress-relieved at 180 °C for four hours, then finished in a second set-up. Removing 80% of the material in one pass is what caused the original supplier's movement.
• Symmetric material removal. Both faces are machined in equal depths so the stress gradient through the thickness stays balanced. The original process machined one face only.
• A vacuum fixture for finishing. Clamping force on a 12 mm plate distorts it while it is being cut. A dedicated vacuum chuck holds it flat without point loads, so the geometry that comes off the machine is the geometry that was measured.

A light stress relief after the sealing groove — before anodising — removed the last of the movement. The anodiser was given a maximum coating thickness of 20 µm and told not to rack the plate on the sealing face.

The result

First article was signed off at 12 days, including the stress-relief cycle and the second anodising trial. The process has since run 200 pieces across four lots with no flatness rejections and a Cpk of 1.4 on the critical face.

The plate now ships with a CMM flatness map — a colour plot of the full face rather than three spot readings — because a single number does not show a customer where the high point is.