Two numbers for the same diameter
A drawing specifies a bore as 12.50 ± 0.05 mm. The setter in the machining cell has been told to aim for 12.48. A new quality engineer sees the process average at 12.48 and asks why the process is “off-centre”. The setter answers that it is exactly where it should be.
Both are right, because they are talking about two different values.
The nominal value
The nominal value, often shortened to Nom, is the value from which the tolerance is expressed. If the tolerance is written as 3.0 ± 1.0, the nominal value is 3.0 and the tolerance limits are 2.0 and 4.0. The nominal value belongs to the design: it is written on the drawing and changes only when the drawing changes.
The nominal value does not have to sit in the middle of the tolerance. A dimension written as 20.0 +0.2/-0.0 has a nominal value of 20.0, but the tolerance runs from 20.0 to 20.2, so the nominal value is at the lower limit.
The target value
The target value is the value the process is meant to produce on average. In a well-run process, the mean sits at the target value and the spread around it is small.
Most of the time the target is simply the nominal value, or the middle of the tolerance when the tolerance is asymmetric. But there are situations where you deliberately choose something else:
- Tool wear. A dimension that grows as the tool wears can be started below nominal, so that the process takes longer to reach the upper limit.
- Asymmetric consequences. If a bore that is slightly too large means scrap, but one that is slightly too small can be reworked, it makes sense to aim towards the side that can be rescued.
- Assembly fit. Two mating parts may be aimed at values that give a better fit than both nominals would.
- Material cost. A coating or fill weight may be aimed slightly above the lower limit rather than at the middle, if the process spread allows it.
The figure shows the tool-wear case. The tolerance is symmetric around the nominal value, but the target has been set slightly below it. As the tool wears the values rise, and the process stays inside the tolerance for longer before the tool has to be changed.
Worked example
Take the bore at 12.50 ± 0.05 mm, so LSL = 12.45 and USL = 12.55. Experience shows that the diameter grows by about 0.01 mm per hundred parts as the tool wears.
- Started at the nominal value of 12.50, the average reaches 12.53 after about 300 parts, leaving a margin of 0.02 mm to the USL for the normal spread.
- Started at the target value of 12.48, the average reaches 12.53 only after about 500 parts.
The numbers are illustrative, but the principle is general: a target value is a way of using the tolerance more cleverly. It is not a way of avoiding a process problem.
Why it matters to say which one you mean
When the target differs from the nominal value and nobody has written it down, the same data leads to different conclusions. A quality engineer judging centring against nominal sees a shift. An operator who knows the target sees a perfectly centred process. An auditor sees neither, unless the target is documented in the control plan or the work instruction.
The same applies to analysis. Indices such as Cpk measure how close the mean is to the nearest tolerance limit, not to the nominal value or the target. A deliberately offset target will lower Cpk, and that is a fair price to pay as long as the process capability remains sufficient. Some indices, such as Cpm, measure the distance from the target value directly.
Common mistakes
- Treating the nominal value as the target without thinking, even when the tolerance is asymmetric.
- Moving the target to hide a process that is drifting, instead of finding the cause.
- Changing the target on the shop floor without recording it, so that later analysis compares against the wrong value.
In short
The nominal value is the designer’s number; the target value is the production team’s decision. Keep them apart, document the target when it differs, and judge the process against what it was actually meant to do.
