When it comes to checking garments, or precision mechanical pieces, inspectors can spend most of their time checking measurements. And, in many cases, they need to write down their findings one by one.

There are few common pitfalls to look out for:

• Use of measurement instruments that are not calibrated (poor accuracy) or that lack precision
• Inadvertent human errors (misreading the instrument, or miswriting the finding)
• Rounding: the inspector doesn’t write the exact finding (loss of accuracy)
• Flinching: the inspector finds a sample slightly outside a tolerance limit, yet inputs it as just inside the limit
• Cheating: the inspector doesn’t check as many pieces as he/she should, and makes up the numbers that are reported

So many things can be wrong, it makes sense to question that process, as a reader who wrote this to me a few days ago:

I wanted to know what your thoughts are on recording actual measurements for dimensional inspection. We currently are required to record actuals but once the product goes one way and the records go another… there is literally no tracibility to the product.

Why must a actual be recorded if acceptable to drawing specification? We do record actuals if out of tolerance on a nonconformance report… which I agree with.

We do record actuals for first article and internally produced parts.. and parts with serial numbers… which I agree as well.

This is a very valid question. I think writing down the actuals only makes sense in the following cases:

• In case the product has a unique ID number, like a MAC ADDRESS for example, there is traceability. An alternative (especially for garments) is to have the inspector paste a small sticker in a place that is not very visible (for example on the care/composition label).
• If there is very little trust in the inspectors. Some records, even without traceability to the samples that were measured, are better than not records at all.
• Whether a finding is on target, or close to a tolerance limit, is not really the same. Some capability indexes, such as Cpm, take it into account. A company that runs this type of analyses needs to know exactly what values were found. Not to mention measurement system analyses such as gage R&R.

Unfortunately, as a service provider, most of our clients request us to write down the exact findings. It is time consuming, but sometimes we can devise a few tricks to go faster while losing only a bit of precision. Here are couple of examples:

• For certain outer measurements of mechanical pieces, a gauge that makes it obvious when a piece is outside tolerance can be designed and made inexpensively.
• For certain measurement points of garments (for example the waistband), it is possible to pile up 10 pieces and check if they are all roughly the same length.

At the end of the day, the question is: how much time can be saved by not writing all the details? And would that time be more wisely spent on other tasks?