When does a checking fixture or functional gauge pay off over CMM for a repeat part?

What makes CMM-only receiving inspection slow or inconsistent?

In rough order of how often they cause the trouble: a shared CMM queue, alignment and fixturing that change from one setup to the next, programs that do not evaluate the feature the way the drawing or the mating part does, and hand checks that cannot see geometric tolerances at all.

1. Queue, not measurement. A CMM is a shared, general-purpose machine. First articles, engineering samples and disputes take priority over a routine lot, so the routine lot waits. The measurement may be quick; the wait is not.
2. Setup-to-setup variation. Every time a part is clamped differently, the datum simulation changes. On thin, flexible or cast parts, a different clamp location or clamp force can move a profile or a position result more than the tolerance allows. Two good CMMs with two different setups can then disagree on the same part.
3. The program evaluates something other than the requirement. A position callout at maximum material condition earns bonus tolerance as the hole departs from its smallest size, and the datum features may earn shift. A program that reports position without those, or that fits a datum with a least-squares plane where the drawing intends a high-point contact, rejects parts that fit and can accept parts that do not. The MMC bonus tolerance guide shows the arithmetic.
4. Hand checks where geometry matters. Calipers, pins and a height gauge measure size and distance well, but they cannot simulate a datum reference frame. When the dock substitutes them for a geometric check, inspectors improvise, and improvisation is where two people get two answers.
5. No agreed method between supplier and receiver. If the supplier measured one way and your dock measures another, a part can pass at the source and fail at the dock with nobody wrong. A shared gauge, or a shared CMM program and fixture, removes that argument.

Before buying any tool, confirm which cause you have. These checks take a few lots, not a project:

• Log two times per lot. Record when the lot arrived, when measuring started and when the verdict was issued. If the wait is most of the total, the fix is scheduling or a dedicated method for that part, not a better program.
• Re-measure the same parts in two setups. Measure three to five marked parts, unclamp, re-fixture and measure again. If position or profile moves by a large share of the tolerance between setups, the problem is fixturing and datum simulation; a dedicated CMM fixture or checking fixture on the drawing datums is the fix.
• Audit the program against the callout. Check that the program applies the modifiers on the drawing (MMC on the feature and on the datums), fits datums the way the drawing intends, and uses the datum order on the feature control frame. Fixing the program is cheaper than any gauge.
• Run a quick gauge study on the hand method. If the dock uses calipers and a height gauge for a geometric feature, a short R&R or attribute agreement study usually shows the spread between inspectors in a day.
• Compare methods with the supplier. Ask for the supplier’s inspection method and raw readings for the same serial-marked parts you measured. Different datum setups or different evaluation of the modifiers explain most source-versus-dock disagreements.

If the cause is a program or fixturing error, fix it and keep the CMM. If the cause is the queue or inspector-to-inspector spread on a repeat part, a dedicated gauge or fixture is the fix, and the payback section below tells you whether it is worth building.

Measurement-system studies make the last two causes visible. Industry guidance in the AIAG Measurement Systems Analysis manual, as summarized by Minitab, treats a gauge R&R below 10% of study variation (%StudyVar) as acceptable, 10% to 30% as acceptable depending on the application and the cost of the device, and above 30% as needing improvement, and expects the system to separate parts into at least five distinct categories. If the method the dock uses today fails that test, a fixture with fixed datums is often the cheapest way to pass it. The gauge R&R acceptance guide explains how to run and read the study.

When does a functional gauge or checking fixture pay for itself?

It pays off when the yearly saving, which is parts inspected per year multiplied by the inspection and waiting cost saved per part, is larger than the yearly cost of the tool, which is its build cost spread over the life of the design plus calibration and upkeep. If that only works out over a design life you do not trust, keep the CMM.

Write the rule as two lines on a spreadsheet. The cost side is the design and build of the gauge or fixture, its first inspection and certificate of conformance, and its periodic calibration and repair, divided over the number of years you expect the drawing to stay unchanged. The saving side is the number of parts you actually inspect each year, which depends on your sampling plan and not only on the order volume, multiplied by the difference in inspection cost between the CMM method and the gauge method. Count operator time and machine time, and count the cost of a lot waiting on hold if the line or the dock schedule pays for it.

Three adjustments change the answer more than the arithmetic does. First, a fixture that cuts disputes with the supplier saves engineering time that never shows up as inspection minutes; if lots are regularly argued over, give that time a value. Second, sampling matters: if the dock checks only a few parts per lot, the saving per year is small, and a fixture pays off only when it lets you inspect more parts, for example full inspection of a safety feature, or when the supplier uses an identical gauge at the source. Third, design stability: a single engineering change to a gauged feature can scrap or force rework of the gauge, so the planned life should be the time until the next likely change, not the life of the program.

A practical rule of thumb: build the gauge after the first article has passed on the CMM and the design is released for production, never before. Prove the gauge against CMM results on a set of parts that includes good parts and parts near or beyond the limits, and run an attribute agreement or gauge R&R study, per AIAG MSA guidance, before the gauge replaces the CMM on routine lots. Keep the CMM for first articles, audits and disputes.

Who does what on a Saola order when a part needs a checking fixture or functional gauge?

Saola Manufacturing is the contracting party, so the gauge requirement goes on one RFQ and one quote, and the quote states how each critical feature will be checked and who supplies any dedicated gauge. CNC turning and milling are done at our lead CNC plant in Vietnam; casting, stamping and surface finishing go to partner plants we manage, and each quote names the production site and country of origin.

For turned and milled parts, inspection is done by the machining and inspection team at the lead plant, which runs CNC turning and CNC milling and has a 2D vision measuring machine, a height gauge and a hardness tester on site. If your plan calls for a customer-designed functional gauge, a supplied duplicate of your own gauge, or a CMM report from an outside lab, say so in the RFQ, and the quote states how it will be handled rather than assuming. For a cast or stamped part, a checking fixture is often the only practical way to check profile on the as-cast or as-stamped surface; the managed partner plant that makes the part is included in the gauge plan, and the quote says which plant does which step. See die casting and stamping for how those processes are arranged.

ISO 9001 certification is in progress, not held. To start, send the drawing, the features you want gauged and your inspection plan for a quote.