The assembly face should be finish-machined after polishing, not before. Machining first is cheaper and fails at the seam: the wheel rounds the edge that was just cut, and the two halves close on a rounded edge instead of a flat. The order that holds 0.05 mm is polish first, then finish-machine the mating face from the machined base seat.
Every setup needs a datum. The first is the cast base face, the second is the machined base seat, and the mating face is cut from the second. If polishing is allowed to touch the base seat, the datum moves by the stock removed, and the mating face moves with it.
So the question is not "which order is faster". It is "which datum survives both operations". A datum that polishing can reach is not a datum.
Polishing removes 0.05–0.15 mm from the visible face. The wheel does not know where the mating edge is. It rolls over the edge and leaves a radius of roughly 0.1 mm on each half.
That radius matters because a feeler gauge reads the gap at the surface. Two rounded edges form a V-shaped groove along the joint, about 0.2 mm wide at the face and tapering to zero at the contact line. The halves still touch; the gauge still enters the groove.
Route A machines the mating face first, masks it, polishes the visible face, and sends the part to coating. It is cheaper per part: one setup for the mating face, one polishing pass, no second cut.
The failure is predictable. The wheel rounds the machined edge by roughly a tenth of a millimetre per half. A 0.05 mm seam measured at the surface reads 0.10–0.12 mm, and a 0.05 feeler enters. The faces may still be flat and still be in contact, but the acceptance rule is "gauge does not enter", and the part fails it.
Route B polishes first, then finish-machines the mating face from the base seat, then coats. The final cut removes the rounded edge and establishes the seam after the last operation that can move metal.
The cost is one extra setup and a small amount of cycle time. The benefit is that the seam is cut after polishing, so the reading at the surface is set by a tool, not by a wheel. The mating face also gets a clean edge, which shows less at the joint after coating.
Item | Route A: machine first | Route B: polish first |
|---|---|---|
Datum for mating face | Cast base face, then exposed to polish | Machined base seat, protected |
Edge at the seam | Rounded by the wheel, about 0.1 mm | Cut after polishing, sharp |
Feeler reading | Enters 0.10–0.12 mm | Does not enter |
Setups | One | Two |
Cost | Lower per part, higher rework | One extra setup |
A 0.05 mm feeler does not measure the distance between two flat faces. It measures whether a blade of that thickness can enter the joint. On a rounded edge it enters the V groove until the taper grips the blade, so the number is an insertion depth, not a gap.
That distinction matters when the drawing says "seam 0.05 max". The measurement procedure has to state the blade width, the insertion force, and the depth limit. Without those, two inspectors can read the same joint differently and both be right.
Write the acceptance as three lines: the seam is measured on the assembled bank after coating; a 0.05 mm feeler must not enter at four points; the edge radius at the joint is 0.2 mm or less (ISO 13715). Then add the datum rule: the base seat is masked through polishing and never touched by the wheel.
For the casting side, see zinc die casting; for the coating step, see whether Zamak 5 can take PVD; for the machining sequence itself, see CNC post machining.