Remember that process validation only applies to those processes that cannot be 100% verified. If you can inspect or measure all the CTQ's resulting from a process, no validation is required. If not, such as soldering, welding, bonding, or other such processes that would require destroying the product in order to measure the process output, a validation is needed.
Also do not forget that part of the process validation, is the attribute agreement analysis or Gage R&R study to show that the measurements and inspections in the process are repeatable and reproducible. This applies to all the processes whether they require validation or not. If all the CTQs can be inspected then the inspection process needs to have evidence of its repeatability and reproducibility in detecting defects.
As stated, a statistician should be consulted to determine the number of samples needed per process. In some cases, one sample may be used to evaluate multiple processes (if they are totally unrelated to each other (no dependencies). Other times the destruction and measurement of a process may require independent samples.
The sample plan should be documented with justification for the methods and quantities of samples selected. You can expect auditors to want to see this in the future.
This is not a simple process -- best wishes for a successful product launch.
------------------------------
Andrew Andreasen
Richmond Heights OH
United States
------------------------------
Original Message:
Sent: 23-Jul-2019 20:40
From: Dan O'Leary
Subject: # of units to validate - process validation
Process validation doesn't apply to a unit, it applies to a set of process steps. The first step is to determine which set of process steps require process validation. The easiest way is to create a process map showing the process steps and the verification steps. It can be at a high level.
For each verification step, determine how you verify the output. Typical choices are:
Verify every unit produced for the CTQ characteristics
Z1.4 single sampling, Level II, AQL is 1.0%
c=0 with an index of 0.65
Z1.9 single sampling Level II, two-sided AQL of 1.0%
No verification
You could have other parameters for the sampling plans. The important question is whether or not there is a sampling plan.
If you verify every unit produced, you do not need to perform process validation for that set of steps.
If you do need process validation, then the sample size question can get complicated. Some of the considerations are whether the CQT characteristics are attributes or variables. For variables you need to understand the specifications and the magnitude of the change you want to detect.
Many people develop an attribute sample size method that typically creates sample sizes of about 50 units. This is not the only way to determine sample sizes. A person with a good understanding the statistics involved is valuable. Start with any ASQ certified Quality Engineers in your organization.
------------------------------
Dan O'Leary CQA, CQE
Swanzey NH
United States
Original Message:
Sent: 23-Jul-2019 14:59
From: Anonymous Member
Subject: # of units to validate - process validation
This message was posted by a user wishing to remain anonymous
Hello, I am wondering how many units we need to process validate. We make electrical medical equipment consisting of hardware and software. Because of the complexity of this device, it's unlikely that there will be more than a few units made for validation. Can anyone shed some light on this?
Thank you.