Design Verification - Managing Changes

J

jamesmiller

Hello Folks,

Wanted to know your opinion on the following scenario

I am in the middle of a design verification and found our that there are two different issues

1. A test failed because the requirement was written wrongly, instead of raising an alarm at 10 seconds, it raised at 20 seconds - but the correct requirement is indeed 20 seconds.

Now in my root cause analysis and resolution, i have to change my requirement to say 20 seconds AND amend the protocol - do I need to rerun my test again ? If the answer is No, then my dates of protocol will be "later" than my dates of execution. Will this be viewed as "fix the spec" than fix the design ?

If I do this, is there a potential design control violation ?

2. Similar as above case, (for a completely different requirement) but this time my protocol was written wrongly and test HAS completed execution and failed, due to a incorrect protocol. Do I just amend the protocol and refer to old results ?

Common sense says, yes - but is there any violation of quality requirements from FDA/CE perspective ?

thank you in advance!
 

John Broomfield

Leader
Super Moderator
Hello Folks,

Wanted to know your opinion on the following scenario

I am in the middle of a design verification and found our that there are two different issues

1. A test failed because the requirement was written wrongly, instead of raising an alarm at 10 seconds, it raised at 20 seconds - but the correct requirement is indeed 20 seconds.

Now in my root cause analysis and resolution, i have to change my requirement to say 20 seconds AND amend the protocol - do I need to rerun my test again ? If the answer is No, then my dates of protocol will be "later" than my dates of execution. Will this be viewed as "fix the spec" than fix the design ?

If I do this, is there a potential design control violation ?

2. Similar as above case, (for a completely different requirement) but this time my protocol was written wrongly and test HAS completed execution and failed, due to a incorrect protocol. Do I just amend the protocol and refer to old results ?

Common sense says, yes - but is there any violation of quality requirements from FDA/CE perspective ?

thank you in advance!

James,

It seems to me that your design verifications have discovered a nonconformity in the spec writing process for which you need to take corrective action.

But the product design itself may fulfill the performance requirements, is that right?

Your design team's risk assessment should determine whether to proceed to design validation or to replan the design verification to start again.

John
 

yodon

Leader
Super Moderator
To add on to John's post, ensure you establish the linkage across everything done. The issues should be tied to your NC / problem report and the problem report tied to the actions / corrections / rationale / justification.

Not being able to connect the dots is when things get uncomfortable.
 

Ninja

Looking for Reality
Trusted Information Resource
Big factor in this is "How much does it cost?"

If it's a 20sec test...just rerun it and take the questions away.
If it means a line shutdown to trigger it, costing tens of thousands...do the paperwork again...
This is a cost/benefit decision...:2cents:
(assuming that a customer hasn't said otherwise)
 

Candi1024

Quite Involved in Discussions
For both of these situations, you create a deviation and explain what the change is and why it is acceptable. After you approve the report, you should go back and change the protocol if you ever plan to use it again. Do not modify the protocol, then attach old results.
 
Top Bottom