IEC TR 60601-4-2 or IEC 60601-2-18, who takes the precedence?

[FelipeSchneider]

Hey everyone, it's widely acknowledged that the FDA has been requiring compliance with IEC TR 60601-4-2 for evaluating EMC performance criteria.

However, while going through IEC 60601-2-18 (which is relevant to the device I'm working on), I encountered this section:

202.6.1.10: The following shall not be considered unacceptable DEGRADATIONS for ENDOSCOPIC EQUIPMENT:
- the short interruption of illumination or image display, or resetting to 'standby' or 'safe' mode, when clearly indicated on the operation panel of the SUPPLY UNIT;
- if the RISK MANAGEMENT process shows that the DEGRADATION does not lead to an unacceptable RISK.

Let's suppose my risk management concludes that an ESD event causing one disposable endoscope tip to burn is not an unacceptable risk because an operator can easily replace this "inexpensive" part. However, this would contradict the IEC TR.
So, my question is, which takes precedence? Can I disregard the IEC TR on this specific topic?

 

[Tidge]

I can't speak for a certification body. My suspicion is that 2-18 "takes precendence" to the extent that in normal use is is expected that there are very likely to be large EM fields present in the vicinity of an endoscope. If this type of damage is occurring outside of normal use (for example: I am imagining a electrosurgical device like a monopolar hook) then I suspect there might be an issue with the device.

 

[FelipeSchneider]

Thanks, Tidge, we haven't tested the device yet, I am trying to come up with a plan to prevent problems if any arise during the EMC test.

 

[Peter Selvey]

IEC TR 60601-4-2 is a guide only so doesn't take precedence because it's not a type test standard (and the FDA should not be using it for compliance as well). Most likely the FDA has a guide that gives this standard as a reference (which is fair enough) and it's persons without suitable understanding of the difference between a guidance document vs type test standard.

Regardless of the guides, FDA or anyone, you should be looking to eliminate issues in EMC as a first step, especially if it is easy to do. Only if it is technically difficult to address any issues that arise in testing, you might invoke special justification that allows degradation. That kind of analysis would be highly specific to the case, and take into consideration how often an EM event occurs, as well as the type harm arising from degradation (i.e. basic risk management). For example, if every tenth operation an ESD related failure occurs, it would not be realistic from a risk acceptability or even an economic standpoint, even if the degradation is in the "annoyance" level. But some of the tests in IEC 60601-1-2 are getting a bit over the top, for example I think ESD is up to 15kV these days. If 15kV is actually rare in the real world, then annoyance type of degradation might be OK. However - and this is very important - if you say 15kV ESD some degradation is allowed, there should still be a reasonable point, which is tested to show no degradation occurs. For example, if it is expected that 2kV is more representative of real world ESD in an operating theater, then the unit should be tested to show no degradation at 2kV.

Obviously, the best way is just to comply without getting into justifications and multi-level testing.

 

[FelipeSchneider]

Thank you again, Peter, you are always brilliant with your answers.

I definitely agree with all your points, but as I mentioned earlier, I am trying to get arguments for a future problem that might exist.
In the meantime, we are trying to address any EMC issues that may exist, but as an engineer, I do recognize the challenges, especially considering the design requirements (small size, construction methods, and so on) and the huge ESD levels to which the device is exposed to and may not make sense due to the usage sscenario.

 

[Peter Selvey]

Yes, I think the EMC standard hasn't quite figured out a good game plan yet. In safety, we often have two limits, e.g. for leakage currents there is 0.1mA in normal condition and 0.5mA in single fault condition. For a Class A transformer, it's 105°C in normal condition, and 150°C in fault condition. The same thinking, if applied to EMC would create two levels of testing, Level N that represents the typical real world + small margin, and Level R that represents rare but possible conditions. For Level N, the device should in principle function normally, while at Level R we don't expect normal function to continue but no medium or high severity harm either (e.g. loss of control of a high risk device).

The problem, or ... problems ... are (a) when EMC was first introduced for medical devices, they used Level N testing, which is not really enough for probing issues of medium/high severity harm (b) they now shifting to Level R testing but allowing messy risk management as a get out of jail free card (c) there's no clear concept of level N and R testing, (d) EMC testing is expensive, so everyone would like just one level, and (e) it's pretty weak verification tool, regardless of the level.

As a general rule, good design means making a device robust, and then testing it to confirm. Testing is really just a spot check so we shouldn't skip the first part and start relying only on test results to prove something is safe. This applies to all testing, not just EMC.

The thing with standards is they often go into overkill in the name of "better to be safe than sorry", or just making it simple or repeatable test, and as a result the actual test becomes detached from what is important in the real world. In those cases manufacturers do often switch to just looking for a positive test result, to put in the regulatory file and don't care about the robust design. Which is reasonable in my opinion, provided that there is still a robust design for the underlying point that the standard is addressing. Which in this case is normal function in the normal EM environment, and no medium/serious harm in harsh or rare EM environments.

 

[FelipeSchneider]

I completely see from where you are coming from Peter, and I also agree with you.

The IEC TR 60601-4-2 tries to solve some of those issues regarding poor/messy Risk Management and get rid of the poorly designed devices that often rely on the "it can fail, it is not damaging anything" excuse to get away with burned devices for a 1 kV ESD discharge while creating different acceptance criteria. ESD is an example, different performance criteria are required for 8 kV and 15 kV.

But what we have to keep in mind, and I think this is the case here as well, is that sometimes it is better to have a device that is able to safely (at least in the user perspective) address medical issues than no device at all. Maybe the engineers who made the IEC 60601-2-18 had that in mind and loosened the requirements for the lighting systems of endoscopes where it does not pose unacceptable risks to the patient.

 

[Peter Selvey]

I think even for annoyance type of events, there is still risk. If it takes time to address the issue, then that could be additional time the patient is under anesthesia, or additional costs for disposables, interrupts the flow of the procedure, or even just a WTF reaction to something breaking for no apparent reason.

A common issue in risk management is to get the right answer but using the wrong reason. A part could get hot, so stick a label on it and declare the risk acceptable. But in reality, the warning label is fairly useless and the true story is the temperatures weren't that dangerous to begin with. Happens all the time. I suspect that a device that is sensitive to ESD damage could be a similar story. Maybe, in the actual environment of use, the kind of things that could build up ESD charges just don't exist. I just asked the "fount of all wisdom" (ChatGPT ) and I'm reliably told that operating theatres a lot of measures are taken to reduce ESD, such as humidity control, antistatic flooring, equipotential grounding, handling procedures for sensitive equipment ... anyone of these could be the real risk control measure, rather than just saying oops, if it's broken you can replace it, nothing to see here.

 

[FelipeSchneider]

Peter, there is a great variable in this game: the inspector/evaluator, who often has their own view on the topic and doesn't agree with yours. So we tend to play safe and restrict our arguments to whatever is explicit in the standard.

It feels a bit like the revisor in the peer review, often we just agree and add in the paper whatever he/she asks so we get it published... :/

 

[Peter Selvey]

I know where you are coming from, having been on both sides of the coin.

For me the best policy is to do the work properly, use common sense and present it with confidence. Genuinely experienced auditors/reviewers can tell it's good work and won't create any unreasonable issues. Inexperienced auditors/reviewers (unfortunately the majority) react to the confidence and become afraid to challenge anything, since it could expose their inability to understand the technical aspects of the issue at hand.

If however the work is prepared with a view of trying guess what the auditor will agree with, using false justifications to get the right answer, and presented without any confidence (what do think? Is it OK?), then both experienced and inexperienced auditors can sense something is wrong. That's when we end up with idiotic non-conformities (which I'm very much guilty of!), which arise from the auditor sensing something is wrong but being unable to get to the bottom of things in the short space of the audit or doc review.

I guess a compromise method might be to formally accept the ESD degradation on the basis of just being annoyance factor, but add a note that in actual clinical use, high ESD levels are not expected due to the normal ESD precautions such as humidity control, antistatic flooring, EP grounding, antistatic chairs and so on, and therefore in actual use, no ESD related degradation is expected.