The 25H7 batch that failed on a micrometer but passed the plug gauge
An inspector checks 40 bores specified as 25H7 (25.000/+0.021) using a 25-50 mm inside micrometer graduated at 0.01 mm. Readings scatter from 25.030 to 25.036 mm, so the whole batch is marked for rework. The same parts pass a go/no-go plug gauge, and the customer's incoming inspector accepts them. Shop temperature is 34 °C; the micrometer was calibrated at 20 °C and its calibration sticker is current. The parts have been on the shop floor for eight hours. The plug gauge has been in use for three years and shows polished, rounded go-pins; there is no wear record. The operator says he feels the micrometer anvils contact the bore wall. Production is stopped pending a decision.
1. Can temperature explain the 0.030-0.036 mm readings? Explain what is really happening.
Read the worked answer
No. Thermal expansion of a 25 mm steel bore over 14 °C is 25 x 11.5 x 10^-6 x 14 = 0.004 mm, and the micrometer and the part are at nearly the same temperature anyway, so expansion largely cancels in a comparison measurement. It cannot produce a 0.030 mm reading.
The real problem is the instrument and the method. A 25H7 bore has a tolerance band of only 0.021 mm, but the inside micrometer is graduated at 0.01 mm - about half the tolerance - and it is a two-point contact instrument read by feel. An operator feeling contact on a curved bore wall can easily be 0.02-0.04 mm out. The micrometer must also be set against a ring gauge or setting master at shop temperature rather than trusted from its 20 °C calibration. These readings are measurement error, not part size. Judgement: withdraw the rework instruction and quarantine the parts.
2. State the go/no-go gauge limits for a 25H7 hole and explain how a worn gauge can give a false pass.
Read the worked answer
For a 25H7 hole the limits are 25.000 mm minimum and 25.021 mm maximum. The go plug is made to 25.000 mm (with a small wear allowance, typically +0.002 mm) and must enter the full bore length under its own weight. The no-go plug is made to 25.021 mm and must not enter; at most two threads or a corner may engage. If go enters and no-go does not, the bore lies within limits.
Wear changes that logic. A go pin worn below 25.000 mm will enter undersize holes, and a no-go worn below 25.021 mm can enter oversize holes. After three years of use with polished, rounded pins, the acceptance is not trustworthy until the gauge is measured on a micrometer or a gauge-setting machine. Check go pin diameter against the wear limit, check no-go for size, taper and ovality, and confirm the gauge's calibration record. Until then the customer's acceptance stands, but the gauge cannot be cited as proof.
3. Recommend the inspection method and records to use from now on, and what you would do with the 40 parts.
Read the worked answer
Use the right instrument for the tolerance: a plug gauge for shop-floor accept/reject, plus a bore gauge comparator set on a 25.000 mm calibrated ring for actual size, or an air gauge where volume justifies it. A gauge should resolve about one-tenth of the tolerance, so 0.002 mm - not 0.01 mm.
Before releasing the batch, run a short gauge study: 10 parts, 3 operators, 3 trials each. The inside micrometer will almost certainly fail the acceptance criterion, while the comparator and the certified plug gauge will pass. Inspect the 40 parts with the certified plug gauge and record go/no-go plus a two-point size for each bore, then release with a proper inspection report.
Corrective action: calibrate the old plug gauge and either re-limit or scrap it, start a gauge wear log with periodic checks, keep the ring gauge and comparator in the same controlled area, train the inspector on feel and on the 20 °C reference, and note the temperature in the inspection record.