IEC 60601-1 and 60601-1-2 retest after PCBA change

[JoCam]

Hi All,

We currently have the control unit of our class IIa medical device system sub-contract manufactured. A project is currently underway to establish if we can manufacture this control unit in-house to the exact same specification. During investigation into the feasibility, whilst all other components can be reproduced to the exact same specification we are not able to do so for the PCBA. If we were to fit a a PCBA with a slightly different specification, which still operates the control unit in the same manner, would we need to retest the entire system to 60601-1 and 60601-1-2?

Your help will be very much appreciated.

JoCam

 

[Peter Selvey]

Legally, it comes down to a judgement if the reports on record are representative of the final marketed device. It's case by case and depends on the relationship between the object tested and the final device. There are also intermediate options such as spot checks, focused tests on change points etc.

Although it's rarely acknowledged, pretty much all testing is done on prototypes, and there are almost always changes between the prototype and the final device. The decision on the reports being representative is rarely documented, and usually it's just implied by the designers putting the prototype reports in the regulatory file for the final device (usually under huge time pressure!). But legally this issue of tests "being representative" sits there in the background.

What I'm trying saying in a round about way is this is a fairly run of the mill decision

 

[JoCam]

Although it's rarely acknowledged, pretty much all testing is done on prototypes, and there are almost always changes between the prototype and the final device. The decision on the reports being representative is rarely documented, and usually it's just implied by the designers putting the prototype reports in the regulatory file for the final device (usually under huge time pressure!). But legally this issue of tests "being representative" sits there in the background.

What I'm trying saying in a round about way is this is a fairly run of the mill decision

Thank you Peter, but is the decision to retest 'yes' or 'no' because the PCBA doesn't now look exactly like the previous one even though it performs the same, and if 'no' can I justify this by verifying that the change does not affect the performance?

 

[Peter Selvey]

Again that kind of decision is fairly run of the mill so to speak. Secondary PCBs are often updated, changed, re-arranged without triggering re-testing for IEC standards in third party labs. The cost of external re-testing is usually high that it has to be a significant change.

The designer should be able to judge:
- change has no effect on compliance (existing reports are representative)
- change has some possible effect, but the tests can be performed in-house (spot checks, focused tests)
- change is beyond the skills of the designer to check, so engage a third party to assess
(the third party may still do just spot checks, focused tests)
- change is so big it makes sense to start all testing again

To do this the designer should know the content of the standard, in order to assess the impact of any changes. If in doubt, you can contact the test agency that did the original tests.

You mentioned that PCBA has some "slightly different specifications" but the overall functionality remains the same. That sounds like something that could be dealt with using an internal spot check. If the underlying technology is different, it could be more involved. For example the different technology could be more susceptible to EMC, noise, humidity, temp etc. Again, could still be focused tests for any weak point. If there are any safety isolation barriers involved (electric shock, cr/cl dielectric strength), could be good to get third party involved.

Another thought: it's not mentioned if this is under third party certification e.g. NRTL, such as UL, CSA or TUV Mark. If there is a certification mark on the device, you are under obligation to report any significant changes. In that case it's better to err on the reporting side. The agency can still judge that no testing is required.