This article explains what incoming quality control is, and how best to use it to gain quality.
What is the purpose of the acceptance test?
The incoming inspection allows the acceptance or rejection of a batch of parts presented by a supplier (internal or external).
What should be defined for the acceptance test?
1/ Initially, it is determined that :
• a level of control of the batch (a severity)
• an AQL (Acceptable Quality Level) per characteristic
2/ Depending on the level of control, the AQL and the size of the batch, the standard indicates a number of parts to be sampled according to the normal, reduced and reinforced modes, as well as an “acceptance” and a “rejection” criteria.
3/ Characteristic by characteristic :
• the batch is accepted if there are less than x bad parts in the sample (acceptance criterion)
• the batch is rejected if there are more than x bad coins in the sample (rejection criteria).
Here is an example:
I receive a batch of 100 pieces.
Based on a defined inspection level, I am asked to take 20 pieces from this batch.
For a monitored characteristic with an AQL of 4, I get an acceptance criterion of 2 and a rejection criterion of 3 :
- I refuse the lot as soon as I have 3 bad coins (or more) for this rating
- I can accept it if I have no, 1 or 2 bad pieces only
For a more critical characteristic monitored with an AQL of 1, I get an acceptance criterion of 0 and a rejection criterion of 1
- I refuse the lot as soon as I have a bad piece for this rating
- I can only accept it if the 20 pieces checked are in conformity
The basic principle is therefore relatively simple.
I only measure a sample of the batch (not 100%), and I have more or less severe criteria depending on the criticality levels of my characteristics.
The advanced functions of the incoming quality control
We will see that this basic principle can be quickly completed.
Depending on the history of the batches of parts already received (for a given supplier), we will also manage types of control :
The batches received were accepted, so the number of parts to be checked is reduced and the acceptance and rejection criteria are adjusted.
The last batches received had problems, so the number of parts to be checked is increased.
If the quality level received is really excellent and is maintained (i.e. we have been in reduced inspection for a long time), we can even switch to “skip” mode, i.e. we no longer inspect the characteristic concerned in all batches, but only one batch out of 3, 5 or even less.
This transition between the different modes (from reinforced to normal and from reduced to skip) and the conditions that allow it is called dynamisation. It is very important that this dynamisation is characteristic! On the same batch of parts, we can control characteristics in normal mode, others in reduced mode, and others still in reinforced mode.
In addition to the simple sampling plan presented above (when the number of pieces requested has been checked, one can only accept or refuse the batch), there are also double, multiple or progressive sampling plans (one can not accept or refuse the “first time”, but ask to withdraw a sample to complete the study).
On a practical level, progressive control is certainly the most promising type of control: one can stop the control (by accepting or rejecting the batch) at each new part controlled; one does not have to wait for the end of the control of the defined sample.
Configured with the same statistical risks as a single or double sampling plan, it is the plan that limits the number of parts measured to a minimum, while ensuring the same results!
There is also a particular way of expressing the acceptance and rejection criteria for the measured characteristics. In this case, we will speak of a K coefficient (Ki and Ks).
The acceptance will be the result of a calculation involving the average and the dispersion of the measurements of the parts in the batch.
What is the advantage of an incoming quality control software?
Depending on the batch size and acceptable quality levels (AQL), the software configures a table of measurements to be checked, characteristic by characteristic.
Depending on the history of the last measurements for each dimension checked, the software can dynamically change the level of control :
- reinforced (more parts are checked and the acceptance criterion is increased),
- or reduced (fewer parts are checked).
Quality Assurance Manager at Infodream
Discover Qualaxy Control, the incoming quality control module of the Qualaxy Suite
Discover the advanced functions of Qualaxy Control, the incoming quality control module of the Qualaxy Suite
To find out more, read the article on our blog entitled Some principles of reception control