Most dimensional inspection asks a question with two answers. A combined go/no-go gauge asks the same question twice, at two sizes, and the interesting outcome is not pass or fail but which of the two the part reached. Reading that correctly turns a binary inspection into a trend, and the trend is what tells you the tool is wearing out.
The go end is the smallest acceptable feature: if the part will not accept it, the feature is undersized. The no-go end is the largest acceptable feature: if the part accepts it, the feature is oversized. A conforming part accepts the go end and refuses the no-go end. That is the whole rule, and it is the reason a gauge with only one end is not a gauge but a limit.
The distance between the two is the tolerance. On a lamp cap gauge to IEC 60061-3 the two ends can be 0.05 mm apart, which is smaller than the thermal expansion of the gauge over a 10 K shift. That is not a design flaw; it is why the gauges are used at a controlled temperature and why they are checked against a setting ring at intervals.
| Observation | Meaning | Action |
| Accepts go, refuses no-go | Within tolerance | Pass |
| Refuses go | Feature undersized | Reject, check the tool for wear first |
| Accepts go and no-go | Feature oversized | Reject, check the part process |
| Accepts go with a loose fit | Near the upper limit | Pass, but watch the trend |
| Refuses go on some parts only | Process spread, not tool | Chart the parts, not the gauge |
| Gauge feels different after cleaning | Resin or oil film removed | Re-baseline, then re-check the setting ring |
A gauge is checked against a setting ring or a master at the start of each shift, not each month. The check takes under a minute and catches the two failures that matter: wear that has moved the go end toward the no-go, and damage that has moved either end at all. A gauge that fails the ring check is removed from service immediately, because every part it has approved since the last good check is suspect.
The direction of wear is predictable: the go end wears first because it enters every part, and it wears toward the rejection side, which means the gauge slowly starts accepting undersized parts. That is the dangerous direction, because it produces escapes rather than false rejects, and escapes are only discovered downstream.
A inspection sheet with a single pass column throws away the most useful information in the process. Two columns, one per end, costs nothing and gives a distribution: parts that fail the go end are undersized, parts that fail the no-go end are oversized, and a mix of both means the process spread is wider than the tolerance rather than shifted.
Standards referenced: IEC 60061-1 (lamp caps), IEC 60061-3 (gauges), ISO 14253-1 for the decision rules that turn a reading into a verdict.
一只钢质量规在正常使用下的磨损速率大约是每 1,000 次插入 0.5 到 2 微米。若每天判 200 件,一个月 4,000 次,一年就是 2 到 10 微米——正好是一只 0.05 mm 台阶规的关键区间。这不是危言耸听,而是为什么"每季度用设定环复核一次"比"每年送校一次"更实用:年校只能告诉你已经超了多少,季度复核能在超之前拦住。
复核用的设定环(setting ring)本身也要有溯源。环的直径公差通常控制在 2 微米以内,且每两年送校一次;如果环本身磨损了 3 微米,所有被它"校准"的量规都会系统性地偏松,而且偏得一致,看不出异常。