When it involves random quality control inspections (usually conducted supported AQL tables), the inspector should check a particular range of items. Maybe, if there are 5,000 items within the whole batch, he/she should check two hundred items (in traditional severity, level II).
If the check was properly ready, the inspector has checkpoints to follow. However, will it mean that these points ought to be checked on all samples?
This is sometimes not realistic. Let’s take an example: a check of a phone. Doing a full operate check on one piece might take a complete man-day. Thus it’s not realistic to inspect all functions on all samples, during this case.
Very often there are checkpoints that take over a number of seconds per piece. If the inspector checks of these points on the entire sample size, it will simply get the workload from one man-day up to five man-days. This is very common.
To drive my purpose, here are 2 examples for quality control inspections:
- Checking the size of the export carton — it always makes sense to measure one carton, and to have a fast check the others (“are they roughly a similar size?”).
- Abusing a product for five minutes to look if it breaks or gets deformed simply — try this on two hundred samples, and simply this one check can take 1,000 minutes, or seventeen hours.
So, in my mind, the logic ought to be as follows:
- What we are able to check very quickly is checked on the entire sample size.
- What we can’t check very quickly is split in 2 categories:
2.a what’s important to the client (“critical to quality”), doesn’t take a really very long time to inspect, and/or may vary well from one sample to succeeding due to the production process. It may be checked on the entire sample size however it’d increase the workload well.
2.b different checkpoints: to be checked on a smaller sample size.
More remarks for quality control inspections:
- In the 2.a case, any issure is classed as a defect, simply the approach a visible defect is counted. It makes sense since all samples were checked for this time (and the statistics from ISO 2859-1 were respected).
- In the 2.b case, let’s say a point is checked on five items and an issue is found on one of the five samples. The inspector will merely contemplate this issue as vital, which implies the checkpoint is failed. Then the client decides whether it’s critical or not.
- This may be a bit crude however that’s the approach most quality control agencies are doing for a long time. The choice would be to check all points on all samples, however sometimes that’s out of the client’s budget.