A go gauge is supposed to accept every part that is inside the tolerance and reject every part that is outside. It cannot do both perfectly, because the gauge itself has a tolerance. The way the standards resolve this is to divide the part's tolerance into zones and assign the edges to the gauge and to wear.
The three zones
ISO 1938-1 describes the arrangement for plain limit gauges. The part tolerance is divided into a gauge making tolerance and, for the go side, a wear allowance. The usual proportions are:
| Element | Typical share of part tolerance | Effect |
|---|---|---|
| Gauge making tolerance | about 10 % | spread between a new gauge and its nominal size |
| Wear allowance (go side) | a further share, inside the part tolerance | distance the go gauge may wear before it must be replaced |
| No-go side | no wear allowance | worn no-go becomes conservative, so it is used to full size |
The asymmetry is deliberate. A go gauge that wears gets smaller in effective size and starts rejecting good parts — a safe failure that the customer notices. A no-go gauge that wears also gets smaller, which makes it accept parts it should reject — an unsafe failure. So the go side gets a wear allowance and the no-go side does not.
Why "exactly on the limit" fails
If a go plug is made exactly to the low limit of the hole, then roughly half of all parts at that limit will be rejected, because half of the gauge's own tolerance band lies above the limit and half below. The gauge is not wrong; it is being asked to do something impossible.
Placing the go gauge inside the part tolerance, with a wear allowance in front of it, means the gauge accepts everything the part standard accepts and keeps doing so until it has worn through the allowance. That is the whole design.
What this means on the shop floor
The number on the certificate is not the part limit. A go plug marked for a 10.00 mm low limit may measure 9.995 mm when new. Operators who compare the gauge reading to the drawing limit will conclude the gauge is wrong. It is not; it is 5 µm inside, which is the wear allowance doing its job.
Replacement is a wear decision, not a damage decision. A go gauge should be replaced when it has worn to the far edge of its wear allowance — not when it is visibly damaged. Most workshops replace too late, after the gauge has already been accepting marginal parts for months.
No-go gauges need a different check. Since there is no wear allowance, a no-go gauge that has worn by even 2 µm is already non-conforming on the safe side. Its check interval should be shorter than the go gauge's, not the same.
The practical consequence
Two gauges from the same supplier, made to the same nominal size, can differ by several micrometres and both be correct — because they sit at different points inside their tolerance zones. That is why a gauge is only meaningful together with its certificate, and why substituting "an identical gauge" from another bench is not a substitution at all.
The working rule is simple: use the go gauge until it wears out its allowance, check the no-go gauge often, and never compare a gauge measurement directly to a drawing limit without knowing where inside the zone the gauge was made.
