Every gauging cycle removes a trace of material from the gauge, mostly at the points of highest contact stress. Wear is not linear in time; it is roughly linear in the number of insertions, with an initial bedding-in period where the rate is higher. A gauge is out of service when its dimensions leave the tolerance on its own sheet, not when it looks worn.
Start with a twelve month interval, measure the critical dimensions at each calibration, and record the change. After two or three cycles the wear rate per calibration period is known, and the interval can be set so that the gauge is withdrawn before it reaches the limit. For a gauge in continuous use on a high volume line, that often works out at six months or less; for a gauge used weekly, several years is realistic.
Measure the features that define the limit, not every dimension on the sheet. On a thread go member that is the pitch diameter at two positions along the thread; on a plate gauge it is the hole diameter and the spacing. Record the values in micrometres; a note that says "within tolerance" tells the next auditor nothing.
A worn go gauge fails good parts, which costs yield and credibility. A worn no-go gauge passes bad parts, which costs a recall. The second failure is silent, which is why recalibration should never be treated as optional when a line is under delivery pressure.
