Grounded (PE) Cable foil as Means of Patient Protection and solid insulation levels for secondary circuits

FelipeSchneider

Involved In Discussions
Hey everyone, I am working on a second version of a device and the Insulation diagram is attached.

As you can see, the 0V of the secondary circuit is connected to the PE conductor to reduce EMI and we would like to keep it like that unless there is no other choice for the problem that I am about to describe.
I am aware that Y capacitors are a possibility to reduce EMI problems and still provide MOPs, but they aren't as good as one or multiple solid connections.

Other relevant points are:
  • The device has no Applied Parts
  • The End-Effector is made of plastic and we can state/prove that there is solid insulation between its enclosure and the patient
  • There are cables in the patient environment (behind the End-Effector which is not very big, around 15cm).
Then comes the problematic points:
  • We are not able to ground all metal parts of the Moving Parts.
  • In the first version of this device, the lab technician stated that there is a need to provide double insulation between the non-grounded parts and the patient. He considered that double insulation is necessary as it was also very nicely debated at Is any part of earthed (grounded) enclosure required to meet, say, 25A/0.1ohm?
    • The lab technician based his report on the leakage current under an SFC.
  • It would be impossible to find USB 3.0 cables that are certified for 1000V, and if there are any, we wouldn't be able to work with them (price and flexibility).
So I am wondering if:
  • Do I really need 1000V protection there (2 MOPP based on Table 6 of the 60601)?
  • Can I state that the cable's brass and foil connected to the PE is one MOP?
  • Can I state that the patient touching the cables is already a SFC?
    • The operator does not need to touch the cables during Normal Use.
 

Attachments

  • Isolation Diagram.pdf
    10.3 KB · Views: 168

Peter Selvey

Leader
Super Moderator
The root cause here is that IEC 60601-1 treats voltages like 24Vdc as dangerous when in reality they are not dangerous in general. There are special cases where 24Vdc becomes dangerous, in particular direct contact with patients that have skin impedance bypassed (electrodes, catheters, injuries, surgery, liquids, large surface areas), and also requires long term, small contact areas (e.g. <1cm²). IEC 60601-1 does not consider this and hence simplifies it to assume that 24Vdc is dangerous and requires 2MOP as a general case. Also, IEC 60606-1 assumes (simplifies) that the operator can transfer the voltage to the patient through the operator's body via negligible impedance, which again is a huge stretch based on fundamental physics of the situation, in practice this would require the operator's skin impedance to be bypassed in two independent locations, and even then it would be a struggle to have any significant current flow. Even if the operator had cuts in both hands exposing their blood circuit, touches a bare 24Vdc at that point, and then touches a similar skin bypassed entry point to the patient, and the patient also has a skin bypassed exit to complete the return path back to the 0V, even with all of that the probability that anything serious happens (e.g. heart stops) is basically zero. There may be some minor cell damage due to tissue necrosis but again that could take hours to materialise. It's not going to happen. Finally, manufacturers utilizing 24Vdc don't go around with exposed 24V pins or wiring, it will be insulated or enclosed within metal connected to the 0V side (not necessarily earth), using female pins for the energised side, and in general preventing contact in normal use, and insulation/0V grounding appropriate for the voltage in that circuit. Not the crazy stuff in IEC 60601-1 like 1000Vrms (a safety factor of 59!!) or 3.4mm creepage where only 0.2mm is needed, or pumping 25A in the metal enclosure of a circuit that is current limited to a few amps. :bonk:

So, the question is not how to make it safe but how to derive a workaround for a crazy standard.

Option #1 is to utilize 8.4.2 c) which is intended for connectors but by equivalence could also be applied to the whole circuit to eliminate the need for 2MOP.

Option #2 is the consistency rule: it's impractical to provide 2MOP for all secondary circuits, so when a test lab does point out an non-compliance they will be doing it in an inconsistent, erratic way. Consider for example a patient monitor used inside the patient environment (or infusion pump, ventilator, etc etc) that has a touch screen, LEDs, switches and controls, printers etc, there is no possibility to provide 2MOP between the internal secondary circuits and all accessible parts, considering not only in the type test, but also regular production controls, supplier controls, design change controls for all the relevant parts. It's just not practical.

Option #3 is to utilize 8.6.4 b) which in the case of a circuit that is shielded/enclosed in the 0V side of the power supply can demonstrate that the current limit in that power supply (24V, 5V etc) is suitable for the current carrying capacity of the shielding/enclosure. This means you can skip the 25A/0.1Ω test, and just make sure the wiring etc is OK for the current limit e.g. 2A or whatever. This can be by inspection or test if needed (e.g. shorting the 24V to the accessible metal part and verifying no excessive leakage current). I don't like this method as it implies the 2MOP is required, so the test lab may still go looking for the other 1MOP e.g. 500Vrms insulated wiring.
 

FelipeSchneider

Involved In Discussions
The root cause here is that IEC 60601-1 treats voltages like 24Vdc as dangerous when in reality they are not dangerous in general. There are special cases where 24Vdc becomes dangerous, in particular direct contact with patients that have skin impedance bypassed (electrodes, catheters, injuries, surgery, liquids, large surface areas), and also requires long term, small contact areas (e.g. <1cm²). IEC 60601-1 does not consider this and hence simplifies it to assume that 24Vdc is dangerous and requires 2MOP as a general case. Also, IEC 60606-1 assumes (simplifies) that the operator can transfer the voltage to the patient through the operator's body via negligible impedance, which again is a huge stretch based on fundamental physics of the situation, in practice this would require the operator's skin impedance to be bypassed in two independent locations, and even then it would be a struggle to have any significant current flow. Even if the operator had cuts in both hands exposing their blood circuit, touches a bare 24Vdc at that point, and then touches a similar skin bypassed entry point to the patient, and the patient also has a skin bypassed exit to complete the return path back to the 0V, even with all of that the probability that anything serious happens (e.g. heart stops) is basically zero. There may be some minor cell damage due to tissue necrosis but again that could take hours to materialise. It's not going to happen. Finally, manufacturers utilizing 24Vdc don't go around with exposed 24V pins or wiring, it will be insulated or enclosed within metal connected to the 0V side (not necessarily earth), using female pins for the energised side, and in general preventing contact in normal use, and insulation/0V grounding appropriate for the voltage in that circuit. Not the crazy stuff in IEC 60601-1 like 1000Vrms (a safety factor of 59!!) or 3.4mm creepage where only 0.2mm is needed, or pumping 25A in the metal enclosure of a circuit that is current limited to a few amps. :bonk:
I could not agree more with you! Definitely, the odds of such an event occurring are close to zero.

Regarding the alternatives that you pointed out, I think we have a way to proceed when considering the 8.6.4 b) as the worst-case-scenario because:
  • The USB is current-limited and not capable of supplying more than 2A. Definitely, the cable shield is capable of handling those 2 amps.
  • The 24V is also current-limited and the power supply would cut the power if a short-circuit occurs, so I could easily argue that this situation would not happen.
I am wondering if the 4th ed. of the 60601 fixes this issue, and I am hoping to get a reasonable lab-test engineer that really understands the issue.
 

eldercosta

Involved In Discussions
Hello Felipe.

Regarding not protectively earthed metallic parts, we have some issues with the test lab in the past due to the lab technician misinterpretation of the standard because of definition of Class I

3.13 CLASS I
term referring to electrical equipment in which protection against electric shock does not rely on BASIC INSULATION only, but which includes an additional safety precaution in that means are provided for ACCESSIBLE PARTS of metal or internal parts of metal to be PROTECTIVELY EARTHED

We used the standard to demonstrate not every metallic part had to or should be protectively earthed:
CLASS I definition implies accessible parts of metal are protectively earthed. However, along the text of IEC 60601-1 standard, there are several instances where metallic parts not protectively earthed are mentioned, regardless ME classification (I or II), admitting exceptions to the general rule. For example, subclause 8.7.4.6 a) and b) prescribes test methods for the touch current that include accessible metallic parts not protectively earthed. Tables 4 and 5 likewise. Subclause 8.6.3 states that “Any PROTECTIVE EARTH CONNECTION shall not be used for a moving part unless the MANUFACTURER demonstrates that the connection will remain reliable during the EXPECTED SERVICE LIFE of the ME EQUIPMENT.”.

Rationale for subclause 8.5.1 provides some examples of means of protection (MOP) to clarify how they may be implemented. On example 2 one reads “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.”; notice that the last sentence uses “could be”, not “should be”.

Rationale for subclause 8.6 states “Typically, metal ACCESSIBLE PARTS of CLASS I ME EQUIPMENT are PROTECTIVELY EARTHED. However, they could be separated by other MEANS OF PROTECTION, in accordance with 8.5. Also, some metal ACCESSIBLE PARTS could be earthed incidentally, neither by a PROTECTIVE EARTH CONNECTION nor for functional purposes. For example, such a part could be in contact with another part that is PROTECTIVELY EARTHED but does not itself need to be PROTECTIVELY EARTHED.”

Add rationales on why non-protectively earthed parts pose no electrical risk to PATIENT and/or OPERATOR (e.g., clearances are met, no direct path to a dangerous voltage source etc.) For example, the design where the lab tech had issues with had a metallic cover bolted to a protectively earthed chassis and a sliding handle also fixed on the PE chassis; neither would pass the 25A test. We performed a risk analysis to demonstrate there was no added risk due to those parts not being PE.

I hope this helps.
 
Top Bottom