Creepages/Clearances for secondary circuits connected to metal enclosure

Benjamin Weber

Trusted Information Resource
I have a question, which seems to be very easy - but I'm just not finding the right answer by myself:

Very often the (-)-pole of secondary circuits is connectoed to the metal enclosure which is connected to PE and the separation from mains is given by a 2MOOP/MOPP power supply. My first idea would be to require a separation of the secondary circuits to accessible parts and applied parts, e.g. the enclosure.

And here I am struggling:
- If in SFC the (+)-pole of the secondary circuit contacts the enclosure, there is a short circuit at the secondary output. If I touch the enclosure in this situation, I see no hazard regarding electric shock (let's ignore other aspects like fire protection etc.).
- Another SFC would be, that the connection between (-)-secondary and enclosure is interrupted, while the (+)-pole does not contact the enclosure. Again I don't see a risk of electric shock.

Let's now even consider the interruption between (-)-pole and enclosure and at the same time the contact between (+)-pole and enclosure as NC. (Because both are not compliant with any standard requirements regarding PE connections or separation of parts). The secondary circuit is still separated from mains by 2 MOOP/MOPP and the enclosure is PE connected. In this situation the enclosure would be connected to the secondary output, but due to the separation from mains the should be no risk of electric shock. And in the SFC of PE interrupted I also don't see any risk from the (+) secondary voltage at the enclosure.

Actually the only hazardous situation for the operator/patient would be to contact either both poles at the same time (left and right hand) or one pole of the secondary and the other on connected to earth. But if the secodary circuit is completey covered by the metal enclosure and there are 2 MOOP/MOPP between secondary and mains this could not happen and I wouldeventually not require any separation between the secondary (+) and the enclosure.

I hope the text is not to confusion and you get the point. I would be very happy to get any responses with your opinions!

mr9000

Involved In Discussions
Figure J.1 says 1 MOOP ;-)

Benjamin Weber

Trusted Information Resource
OK, I get Fig. J.1. ANd I just read Annex A on cl. 8.5.1 saying:
PATIENT CONNECTIONS and other ACCESSIBLE PARTS are separated from parts different from earth potential by BASIC INSULATION and an intermediate PROTECTIVELY EARTHED metal part, which could be a fully enclosing metal screen.

But actually I don't see the additional safety if the secondary (-)-pole is connected to the enclosure (regardless of being PE connected or not). If the (+) pole contacts the enclosure there is short-circuit and as long as the separation via the power supply is sufficient, there shouldn't be e risk.

What am I missing?

mr9000

Involved In Discussions
Hm, what if the (earthed) patient touches the short-circuited enclosure, the leakage current through the patient could get too big?
One could argue that the risk of a short-circuit is significantly reduced by the MOP barrier.
Maybe it just is, what it is? Arguing with common sense against normative requirements can lead to madness and despair ;-) If the PE counts as one MOP, where is the second MOP in the secondary circuit?

Peter Selvey

Super Moderator
In the background here is that 24Vdc is not considered dangerous by most standards, and to some extent IEC 60601-1. According to 8.4.2c, the operator is allowed to access voltages up to 60Vdc without any MOOP as long as the risk associated with transferring that voltage to the patient is sufficiently low. This second part is debatable and questionable (long story), but if we ignore the patient aspect, the standard is clearly allowing direct access to 24Vdc for operators, which implies excessive effort to prove 2MOOP is a bit of overkill.

Also to note, that for MOOP, there is no dielectric strength test or insulation thickness for 24V circuit. So, basically we are looking at spacings and possibly protective earthing.

Finally, with respect to protective earthing, if the source is 24Vdc and reasonably current limited (e.g. a few amps) then it is not required to pass the 0.1Ω/25A test. It is only required to show that the resistance is low enough to trip any current limit device. Also this test is not required to back to the mains PE terminal, it is in principle a localized test only for parts where current could flow sourced from the 24V circuit. However, for practical purpose, it may be simpler to demonstrate low resistance to a common point like the PE terminal, but again the currents do not need to be 25A.

Overall when using the metal enclosure for a secondary circuit, we can find 2MOOP if we look closely, but often it's not worth the effort because it's not really dangerous in the first place. For example, there may be the following constructions:

Method 1: MOOP1 spacings between 24V circuit and enclosure + MOOP2 low impedance of the enclosure (as needed for the 24V circuit)
Method 2: MOOP1 current limiting device itself + MOOP2 low impedance
Method 3: MOOP1 24V circuit is inherently current limited (unable to provide much current) + MOOP2 low impedance

In the case of Method2, the current limit device should have spacings around it to avoid the device is shorted in normal use, and a reliable device used e.g. certified fuse, PTC device. In theory, the use of electronic over current protection would need to be analyzed for suitable reliability, but again in many cases for secondary circuits, the electronic overcurrent protection is not massively greater than the inherent current limit, it's not like a 24V circuit with 1A fuse can provide 100A, unlike a mains circuit. So it's kind of a waste of time to go into too much detail.

Benjamin Weber

Trusted Information Resource
Hey Peter,
thanks for your detailed response!

Regarding the "save 24V for operators":
In my understanding cl. 8.4.2 c) does not allow all operator accessible only parts to be up to 60 Vc / 42,4 Vp. It says: "...the limits (leakage currents) do not apply to the following parts (...):
- accessible parts of connectors
- contacts of fuseholders (...)
- contacts of lampholders (...)
- parts inside an access cover (...)"

As far as I understand, this is a straight forward list of very specific parts - and not all operaotr only accessible parts. BUT: I don't want to raise this discussion here because.....

Unfortunatley in my particular test case the situation is as follows:
- Applied part type B with PE connection
- secondary high voltage 960 Vdc (photomultiplier tube)
- HV power supply as short circuit safe according to the manufacturer

The minimum air clearance required is 9,6mm, I found 4,7 mm so far

Would it be appropriate to measure the patient leakage current with the HV ouput shorted to the enclosure instead of requiring the minimum distances?

Peter Selvey

Super Moderator
According to 8.1/8.5, MOP is only required if the limits in 8.4 can be exceeded. So, when 8.4.2c provides an exclusion for leakage currents limits for certain parts, what it really means is that those parts are excluded from all other requirements in Clause 8. Of course, this only applies to connectors. My point is, if it's OK for a connector, why is it not OK for other parts? But anyway, as you point out your case is 960Vdc.

A similar case could apply to the 960Vdc. If there is no plausible path for measuring leakage currents (e.g. due to the 960V being totally enclosed in a solid metal enclosure) then MOP is not required. It's not automatic that MOP is required just because there is high voltage there.

Also, in practice, we really only assess the "weak points".

Consider if there is a red hot radiant heater in lounge room, and you want to protect young kids from going near it. It makes sense to put a child-proof protective cage around heater. And we could then assess the quality and effectiveness of this barrier, measure if the gap is large enough, how much force to push it over, and so on. But it doesn't make any sense to worry about access to the same heater from the kids room on the second floor, even though it sits directly above the heater. Yes, we could identify that the flooring in between acts as a barrier. But it's not really plausible that the kids could access the heater though the flooring.

In the same way, when assessing prevention of electric shock, there will be weak points which need to be formally assessed as MOP, but equally there will be many paths that exist in theory but are unrealistic (essentially impossible) in practice. So with the 960Vdc, you need to look at what paths could realistically reach the operator (or patient). For these paths, identify the weaker points which should be assessed as MOP. In this context, a "weak point" is not necessarily bad but it just means an area that is worth checking closely.

So OK, we short the HV to the enclosure, and measure leakage, and result is pass. But what makes it pass? Usually it is a combination of the current limit in the HV supply and the relatively low resistance in the enclosure that provides a return path with negligible voltage appearing on the enclosure itself. If for example the HV is limited to 2mA, then it only needs the enclosure to be <50Ω to stay below 0.1V, which might be so obvious that it's not really a weak point or worth testing.

But there might be a wire connecting to the enclosure that is in a arm that is flexed. This could be a weak point: over the life of the equipment, the wire could break due to flexing, and then shorting the HV to the enclosure will put the 960Vdc on the enclosure with no return path.

Also, how reliable is the 2mA? Maybe, the HV is generated in a high impedance winding in a tiny dc/dc converter so it can't really output much more than 2mA, so it's not really worth looking at, especially noting that the enclosure resistance will be surely be <<50Ω in practice.

But, maybe the designer added a 700Vac winding to the power transformer that can pump out 500mA before the current limit. In this case, we might either look closer at the current limiting device, or make sure the return path is say <0.2Ω @ 1A (for a safety factor), or maybe consider 1MOP clearance which would allow 0.5V limit (0.5mA into 1kΩ) and then use 0.5Ω resistance, or 2MOP ... messy calculations but only if the designer is lazy in using a excessively high current HV supply when a low current supply would suffice.