EBME Forums Forums Maintenance Technical Assistance How should we test stacks?

"Out the door safe" sounds good to me, Richard. But going back to John Sandham’s original question, isn’t 60601-1-1 the most appropriate standard for medical systems (and wouldn't a stack come under that category, as Vaughan mentions)? And, am I not right in thinking that 60601-1-1 expects an isolation transformer? How about it, John?

Here’s a nice list of standards (probably a bit out of date vis-à-vis current issue dates), but it gives an idea of the amount of stuff published.

And finally, why aren’t we all simply following DB2006(05) anyway (isn’t this the current "bible")? For those who may not have been paying attention (?), it is available for download (as a .pdf) here.

Those researching further shouldn’t forget this lovely link, of course (even though it could do with a bit of updating)!

Hi Geoff,

To paraphrase (hence simplfy and give an idea rather than try to give the impression I'm really clever and have access to standards, which I do anyway, by finding a paragraph in a standard and quoting it); a medical system is one that consists of multiple devices, one of which, at least, is a medical device, that are connected together via a common mains connection, e.g. extension socket and/or functional connections, e.g. RGB video cable, within the patient environment or even the same room, particularly.

Non-medical devices can be interconnected by functional means but, ideally (the standards indicate) system mains connections need to be made inaccessible - in my experience this is not always the case - neither are separating transformers absolutely necessary if only medical devices are considered. However if there's the possibility that operators can manufacture their own systems, i.e. plug and unplug or interconnect medical devices and non-medical devices within a system, as they wish, then the electrical safety of the system may be compromised.

A separating transformer is a means of protection against leakage currents that exceed the limits of 60601-1 thus it's used on systems that can generate leakage currents in excess of the limits in the standard. The idea is to remove the electrical connection between N and E (that exist back at the mains substation) so that leakage currents from the mains parts of the system have less affinity with Earth - thus earth leakage is reduced since a circuit path no longer exists (in theory). Leakage currents are much reduced anyhow, earth leakage significantly. Separation can never be perfect but medcial grade separation is sufficient for our needs generally. The more mains-powered devices (sharing a common mains connection and/or functional connections) there are in a medical system then, generally speaking, the greater the combined leakage currents.

Incidentally there appears to be some confusion caused by the 3rd edition standard as to what the earth-leakage limits should be for systems connected to an MPSO, permanently installed or otherwise. Perhaps someone would be willing to give their "interpretation"? As I pointed out to an esteemed colleague of mine earlier on in the week, the limits appear to be 10x greater in the 3rd edition 60601-1 standard, to me. 3 different clauses in the standard, referring to systems particularly, indicate this and the rationales seem to imply a change in reasoning about risk due to earth leakages prevously regarded as "excessive".

A typo appears doubtful to me but perhaps someone's made an error somewhere - we await feedback from the MHRA. Typically a separating transformer, non-conductive shielding and/or additional earthing, is only required when the medical system or device has leakage currents that exceed the 60601-1 limits under NC and/or SFC. The leakages are inherent due to limitations in materials, current construction methods and interconnections, etc, that, even under NC not just SFC, can lead to excessive currents, etc. Leakages vary depending upon what the system consists of and how it's interconnected. In my opinion the "safest" way of manufacturing a system is to interconnect medical devices, using a separating transformer that's rated appropriately, in such a way that the operators cannot adapt it at all, i.e. give no access to mains or functional connections unless they're needed in normal use.

The idea of applying safety tests after a system is "manufactured" in-house is to ensure it's safe. Use of a separating transformer may not be necessary on a system that passes the applicable safety tests (to 60601-1) that's fixed and will never chage configuration but there's a significant and unacceptable risk, in my opinion, if operators can adapt a system in any way, irrespective of means of protection, or the system meerting 60601-1 when it leaves the workshop.

Again I'm only giving a very brief overview of my own opinion (read as: professional opinion - for those who think opinion means "doesn't know what he's talking about" - as a qualifed and experienced individual providing a simplified interpretation and application of the standard for routine maintenance purposes - since there's no alternative guidance - taking into account the significant safety issues rather than dwelling on "academic" issues and for the sake of brevity). I do this "on the fly", "off the top of my head" and don't have time to prepare material for the "classroom", i.e. I'm not lecturing nor do I have vested interests nor am I being paid to do so, just trying to give a simplified answer. Feel free to pick it to pieces.

Richard

R J Ling BEng(Hons) MSc MIEE IIPEM

P.S.

Common sense dictates that if you have medical devices that are connected together and can be adapted by anybody, i.e added/removed, then they need to be tested as individual devices and then as a system - those analysing risk in this situation are making a big assumption that the system will not be modified - no-one can predict what changes can take place and how this will affect safety in practise - risky in my opinion.

If the system connections are non-accessible, as per the safety standards, and the system is fitted with a separating transformer testing the system will only test the integrity of the system overall, primarily the separating transformer, won't it? Not give an indication of degradation or problems with individual devices necessarily.

If the system is made up of medical devices connected via an MPSO that's inaccesible and that does not require a separating transformer then perhaps testing of the system, via the MPSO without individual tests, is satisfactory, i.e. carries less risk of failing to detect issues, than the previous testing scenario with a separating transformer, since any significant issues with individual devices are more liklely to be picked up on?

Irrespective of the scenarios above I'd be inclined to test all of the devices in the system and as a system certainly after repairs/reassembly/re-configuration. This allows safety test figures to be provided that indicate the nature and significance of individual leakages, etc, to and from devices, and as a system. Especially if by doing reassembly/re-configuration I'm considered a manufacturer according to the standards and MDD legislation.

Hi Geoff / Richard. I believe you have taken my comment "pick and mix" the wrong way. In absence of a mandatory standard in the UK (unlike Germany or Australia) each EBME / Med Physics / Clinical Engineering department can apply any safety test routine. I am sure those who post on EBME are highly passionate about their job and do so very well in determining the best practice when it comes down to safety testing however, it seems no unanimous approach exists. Its down to the individual expert to decide whether to follow MDA (some people still use HEI95) or follow the IPEM guidelines, follow manufacturer's guidelines or test to IEC 60601-1-1. When using the latter, several tests are available from manufacturers such as Rigel / Fluke / Metron etc that are not used because individual departments deem them destructive (F-Type or SIP/SOP leakage). Note that IEC 60601-1-1 states that all leakage measurements must be performed from an isolated supply (part safety part to ensure TN system). I have not come across many hospitals in the UK (or Europe) that actually respect the test conditions in IEC 60601. (I appreciate this is a type test standard however it demonstrates that people are willing to take a standard and apply to their own needs (pick and mix). Many departments also choose to do a 1-2 sec earthbond test (not referenced in IEC 60601-1-1). Some do a IEC 60601-1-1 test but start with an isolation test (not referenced in IEC 60601-1-1). Not to forget the a Dielectric test on Double insulated / reinforced parts (test equipment readily available).This is what I mean with pick and mix. Many departments (perhaps not yours) decide themselfs that the Isolation test (MDA or best practice) is best to de done before a variety of Safety tests from IEC 60601-1-1.

I trust this explains the term "pick and mix" It is not a derogative term meant to offend professionals like yourself.

Enjoy the rest of your weekend.
Greetings from sunny Boston.

Don’t worry, John, I wasn’t having a go at you, but just trying to emphasize the nonsense of the situation (which, by the way, has been going on in various guises for at least the last thirty years – hence my comment about not expecting sudden changes anytime soon)!

Personally, I like to keep things simple. We have standards (eg, the IEC 60601 series, etc.), then we have guidelines (…based upon, and referring back to, the standards – eg, DB2006(05)). All I’m suggesting now is that we, as a group representative of the people doing the actual work, form a consensus about how to proceed. We could indeed call it "best practice"!

Or, put another way, why are different hospitals still doing this stuff differently after all this time? Beats me!

Hi John,

I didn't think the term was used in a derogatory manner and I actually agree with you.

For manufacturing and type-testing purposes those doing the testing have to "pick and mix" applicable 60601-1 tests using appropriate methods - so do we in EBME/MP if we're manufacturing medical systems. I actually welcome a "definitive" test standard that applies only tests that are absolutely necessary to ensure the safety of medical devices, that're undertaken routinely and after repairs, but at this moment there isn't a recognised standard that I'm obliged to apply.

As far as I'm aware there are currently no guidelines RE:safety testing from IPEM that relate to medical devices that are not permanently installed, neither are there recognised standards relating to recurrent safety testing from HSE, MHRA, DoH or my employer - that's the point of my comments regarding current, i.e. recognised, standards and guidelines (Rigel supplied guidelines are not "definitive" unfortuately). The MHRA recognises current standards because they're involved with the regulatory aspects of medical devices that come under the 3 medical device directives. However I doubt they will be interested in 62353 until this standard is ratified/harmonised thus recognised in the UK, at least.

From experience I've very rarely seen devices damaged or proof that they've been degraded by the additional mains insulation tests or mains on applied parts tests specified in 60601-1 for example. I would never consider destructive testing such as "flashover" i.e. dielectric withstand tests, specifically. Earth bond may have caused a very few problems perhaps - but usually when undertaken on device and systems where the integrity of internal earth bond are suspect or they're modular systems with external combinations of parallel funtional and protective connections anyhow.

A common sense approach to testing is to apply non-powered tests before applying power to a device that's just been reassembled after repair. If only for the safety of individuals at the bench, forget standards. I'd prefer an insulation test to be performed before powering up a device and damaging it beyond repair.

300Vdc is recommended in DB2000 supplement on acceptance testing but I find 500Vdc causes no problems L/N-E, mains parts to enclosure or applied parts to earth. We routinely reduce it when there is breakdown and calculate equivalent impedances to assess applied parts to earth resistance for example, when there's an issue - and there are occasionally.

On recent example is that it allowed us to detect that a manufacturer of diagnostic ECG recorders is using a single 68MOhm resistor between isolated I/P section and earth. A marginal fail but it illustrates that the test does find problems that leakage under normal conditions does not. Don't know why 68MOhm was chosen and fortunately resistors tend to fail O/C.

Also recently found degradation of a reciprocating motor actuator electrically connected to the BF applied part via a saline-primed circuit to the patient (at just over 2MOhm) during acceptance tests on evaluation/loan on an old type of a 5 year old unit compared to newer ones at 6MOhm. The potential for degradation overtime with this device would have been missed otherwise. Leakage currents didn't pick this up, i.e. a benchmark insulation resistance figure allows us to make more meaningful comparisions when applying voltages externally to applied or non-earthed parts for example.

I think some of these people who get involved with writing standards should come and see people who actually use safety testers as bench instruments, rather than automated "black boxes that churn out figures", every day, to assess and evaluate safety of equipment on acceptance and after reassmbly to ensure that manufactured systems are safe to using EXPERIENCE & SKILLS of REAL-WORLD scenarios rather than quote some manufacturers standard as a solution to all our safety testing problems.

Trading 60601-1 against 62353 does not work with me John since I'm aware that I will likely need to use both (when 62353 becomes current) because we do routine testing, acceptance testing and testing after "manufacture" (according the the current BS EN 60601-1 recognised by the MHRA. They have different applications but achieve the same ends - hopefully they don't conflict in terms of the level of safety that's assured to the patient.

62353 will only be applicable after repair and for recurrent testing and as far as I can tell looking at drafts of the standard it won't help us particularly when we evaluate, acceptance test from new, or manufacture systems since applied parts, applied parts insulation tests, etc are not included. And in my experience these are valuable tests in the correct hands - not every new medical device or system, in particular, is "safe" when it comes through the door.

We don't have a scenario here where medical device and medical safety tester manufacturers are trying to say that technicans and engineers should not be allowed to apply appropriate and equivalent tests to 62353/60601-1 giving justification for such tests do we?

"I have not come across many hospitals in the UK (or Europe) that actually respect the test conditions in IEC 60601. (I appreciate this is a type test standard however it demonstrates that people are willing to take a standard and apply to their own needs (pick and mix)."

Are you saying that we shouldn't bother to use any 60601-1 safety testers - including your products, John, or that we use only safety testers that include an integral transformer to ensure this? In reality all I'm interested in are the limits that are considered to reduce the adverse physiological effects to operator and patient, i.e. those applicable limits in 60601-1 and built into the testers I use, currently, and having the means to apply tests that can diagnose these values.

Sounds as if everyone's expected to throw their 60601-1 medical grade safety testers away and buy a 60601-1/62353 compliant tester tomorrow! On a serious note i will be looking further into the 62353 standard it's application and how testing is implemented - so I'm ready when asked to comply with it.

My word, what can I say. I agree with you Richard but still am somewhat concerned that you did not get my point about IEC 60601-1. I was merely responding to Geoff's initial comments regarding my use of "pick and mix". The only thing I was trying to get across was that you can use IEc 62353 in the UK today just like you can use a part of the 60601-1 tests, or MDA or manufacturers guidelines. Some manufacturers are already rewriting their service manuals and these find its way to a hospital near you. As such you can claim from receiving such manual that the manufacturer recommends the use of IEC 62353.
I am glad that I do receive a lot of useful information in the UK unlike other countries in Europe where most people haven't got the foggiest about safety testing and do listen to the supplier of safety testers (which is wrong in my opinion) We (as manufacturer) are not in a position to make recommendations to the customers. We respond with product designs to the market demand. Therefore, I appreciate your input and that of your colleagues in stating what the UK market is today. This has been taken into account when designing our latest analyser (I won’t mention the name as this would be considered too commercial)

Maybe we should take up Geoff's point and get together to discuss what is required in the UK based on first-hand expertise rather than (I will not use the word biased) applying what other people perceive necessary.

Anyway, I am on my way to present a seminar in Boston on IEC 62353 (I named it " IEC 62353; Clarity or Confusion") which sums up the current state in 'Safety Testing Land'.

If, as you say, most people in Europe don't have the foggiest (idea) about (electrical) safety testing, why should we give a damn what those IEC committees come up with? I think we've all had enough "dumbing down" of our erstwhile good British engineering practices.

"Clarity or Confusion" ... I think we all know the answer to that one, John. Anyway, now we look forward to hearing reports of your success at the seminar (hint - we need clarity). Good luck, Mate!

The way I see it the safety testers that most departments possess will determine the range of tests that are applied to 60601-1 standard. In fact a proportion of these testers are Rigel manufactured - sold on the basis that they're 60601-1 compliant no doubt. These tests are what we "pick and Mix" from generally speaking - a subset of the 60601-1 type-tests (the non-destructive ones). As I said these are applicable to manufactured devices - the definition of a medical system is given in 60601-1 and 60601-1-1. I see a continuing need for 60601-1 compliant testing in NHS departments and for 62353 (when ratified in the UK). We occasionally manufacture medical systems and routinely test medical devices.

IEC62353 can be used by NHS departments in the UK for routine testing and after repair but we need some guidance on its application in the UK from the competent authority, i.e. the MHRA. I would want to determine whether its use is in conflict with 60601-1 or not. For example if a system passes 60601-1 at acceptance then is it likely to pass 12 months down the line using IEC62353, assuming there is no degredation, i.e. are the tests and limits compatible with those that were used for acceptance testing? Or do we "dump" testing to 60601-1 altogether? Should 62353 be used for acceptance checks?

Let's not just jump in with both feet otherwise we end up in a really confusing situation with two standards for safety testing and no guidelines on when or where they're applicable or even a new standard that doesn't fully meet our needs after we dump one that's served our needs fairly well. I can't see NHS departments or the MHRA just "junking" 60601-1 when considering regulatory & safety issues, neglecting NHS acceptance checks, NHS in-house manufacturing checks and tests on evaluation, if NHS departments start using IEC62353. This is because the IEC62353 safety tests aren't likely to be comprehensive enough or applicable for purposes other than routine maintenance, e.g. because of limited tests on applied parts, etc, (even if the limits are aligned with some of the 60601-1 limits).

Commercial manufacturers cannot have the monopoly on the 60601 series of standards because people like us are considered manufacturers as well as purchasers of devices - that we are also obliged to test as fit for purpose and safe at purchase, evaluation, after repair and routinely (since we have a duty of care to operators and patients). I know the regulatory involvement and input on safety issues that in-house NHS departments have can be a pain for manufacturers (and the MHRA) but I have seen so many poorly constructed and relatively "unsafe" or what I consider dubious systems pass through the door, on electrical safety grounds alone, I wouldn't want 60601 "lost" to R&D or maintenance departments.