How Do You Know a Fixture Plate Is Actually Repeatable Between Jobs Versus Just Close Enough?

You know by measuring the same reference feature at the same point in the process across multiple setups and comparing the numbers, not by trusting that it felt tight going back on the table. "Close enough" is a feeling. Repeatability is a number you can put on paper, and the two are not the same thing until you've actually checked.

What "repeatable" means in practice

A repeatable fixture puts a part in the same position, to within a stated tolerance, every time it's mounted, whether that's the same day, next week, or after the plate's been off the machine for a month. The test isn't whether the part machines correctly once. It's whether a known reference point, touched off with an indicator or probe on setup one, lands within your tolerance band on setup two, three, and ten. If you've never actually measured that, you don't know it's repeatable — you know it worked last time, which is a different claim.

How to actually check it

Pick a feature on the fixture itself, such as a ground locating pin, a precision dowel hole, or a tooling ball, rather than a feature on the part, because you want to isolate fixture repeatability from part-to-part variation. Indicate that feature's position relative to machine zero, unclamp and re-clamp the fixture (or swap it out and back on if it's a removable subplate), and indicate again. Do this several times, not once, because a single repeat that happens to land close doesn't rule out a fixture that drifts under thermal or clamping variation. A probe cycle logging the result automatically is more useful here than a hand indicator, since it removes the operator's touch from the measurement and gives you a number you can track over time.

Track that number across jobs, not just within one setup session. A fixture that repeats to 0.01 mm today and 0.05 mm three months later is telling you something wore, loosened, or got knocked out of true. Catching that from a log is a lot cheaper than catching it from a batch of parts that fail inspection.

What actually causes drift

Locating pins wear from repeated insertion and removal, especially if they're not hardened. Clamping mechanisms lose preload over time, particularly anything relying on a spring or a bolt torqued by hand instead of a specified torque value. Thermal cycling, a fixture that goes from a cold shop floor to a running spindle and back, can shift a plate that was dead flat when it was made. And a fixture that gets bumped, dropped, or re-bolted to a different table location without re-qualifying its zero is the most common cause of all, because it's the one nobody checks for until parts start coming out wrong.

The practical takeaway: build the repeatability check into your process, not into your gut. A fixture earns "repeatable" status by data across multiple setups over time, and a fixture that hasn't been checked that way is running on assumption, not verification.