IEC60601-1: HiPot Test Method for Internally Powered Devices

[Jokaru]

Hi everyone,

We developed a wearable device with an internally powered battery with type BF applied parts. We recently sent the device for IEC60601-1 certification before proceeding with a clinical trial using the device. The device is powered by a rechargeable battery that can NOT be operated/used on patients during charging.

However, we had the device fail during the HiPot Dry Test (500V DC & AC) the first time. The Dry Test passed when we removed the Protection Circuit (biased to Vcc, & Gnd with diode) the second time, but failed after placed in humid chamber (93% Humidty @ 25 Celcius for 48hrs).

What i don't understand is the testing method the engineer employed during the test. I understand the method for AC devices, where HOT & Neutral are shorted and connected to INPUT end of the HiPOT, and the return is at the Enclosure/PE (as image below).

However, the test engineer conducted the same method for our battery-powered device. He shorted V-Bat and GND to the INPUT end of HiPOT, and then return to the enclosure body and applied parts (all wrapped with aluminum foil).

Pardon my ignorance, but if anyone has knowledge of the test method for HiPOT for Battery Powered devices or a reference on the testing method, that would be very helpful. Any suggestions would be very much appreciated.

Thank you.

 

[Peter Selvey]

The dielectric strength test (hipot) is intended for "solid insulation" only, like plastic insulation on wiring or a plastic enclosure, or a plastic bobbin inside a transformer. It's not intended for testing air spacings or tracking along surfaces, that is covered separately by creepage distance and air clearance evaluations.

It's quite literally an aging test for the insulation, intended to prove the insulation will handle the normal working voltage for the life of the device. The test after humidity is intended to check if the insulation is hygroscopic (some types of insulation like impregnated paper, powder do have the potential to absorb water).

Also it's not meant to be a "black box" test, both the designer and test lab should know exactly where the "solid insulation" is and should have a good idea if it will pass or not based on material specifications. It is important to avoid black box testing (just blindly wrap in foil) as this can easily end up testing parts that are not intended to act as safety insulation and also not test parts that should be tested.

With modern materials, it is very rare to have a genuine failure for dielectric strength, most plastics can handle 20kV/mm. If a failure occurs occurs, it is 99.9% sure to be due to either an incorrect test set up or inadequate creepage and clearance (spacings), which again is a separate issue. Similar to solid insulation, minimum spacings should be known and controlled by the designer to meet the requirements.

A side problem is that 500V test on a battery operated device is way way way overkill, there's probably some kind of crazy worst case logic to justify this. So all of the above, while very important for mains insulation and genuine high voltage parts, is overkill for a battery operated device. However, unfortunately the standard requires it, so it's best to approach it the same way as "real" insulation; that is, plan the insulation barrier design, know exactly which parts form the insulation barrier, control and test the solid insulation, control and test the spacings.

 

[Jokaru]

Hi Peter,

Thank you for your very thoughtful insight. it does help us to understand while communicating with the test lab. I have also read through the other thread on "Internally Powered Medical Device and Dielectric Strength" in which you provide a very thorough explanation of the possible exclusion of the Dielectric test for Battery powered devices. I would like to take a similar approach to communicate with the test lab, however, we still need to use the Type BF marking.

For our device, the insulation diagram is similar to this (EMG electrode instead of SPO2 probe):

Screenshot 2022-10-27 033321.png

• Jokaru
• Oct 26, 2022

insulation diagram for battery-powered devices

The issue is, path "C" is electrically connected to the applied part(electrode). in this case, can I exclude path C in the diagram?

Thank you

 

[Peter Selvey]

The issue here is not whether isolation C can be excluded. The issue is - what is your isolation plan to ensure currents remain below the limits in the standard?

There is a battery with 3V which has the potential to provide 3mA to the patient (patient is assumed to be a 1k resistor). How does your design prevent this?

It's possible for example that between the input op amps and the patient electrodes there are 330kΩ resistors, which guarantee that the maximum current in any normal or fault condition is <10µA. In this case, it could be argued that no isolation is required (except maybe around the 330k, but that is another story ...)

Or B could be designed to pass 500V (1 MOPP), and the input resistors are 100kΩ, in which case the maximum current in single fault condition is 33µA (which is below the 50µA limit).

Or you could have your own design solution. The point is, figure out how your design limits the current to <10µA in normal condition and <50µA in single fault condition, and if that involves MOPP barriers, then test those barriers.

 

[Peter Selvey]

It is a bit of a stretch but the point of view of the standard is that negative pole of the battery could somehow get connected to earth (either in normal or fault condition) which then makes the positive pole "live" and requiring an isolation plan to prevent patient leakage currents (and operator currents as well).

Let's consider a "bad" implementation: A 12V battery operated device "X" is metal enclosed and this metal is directly connected to the negative pole of the battery. The box is sitting on top of ventilator which is mains powered, and has an earthed metal frame. The device also has wiring to a finger probe pulse oximeter. This has thin insulation between the LED circuit and the patient's finger, which would fail at 1000Vdc. The patient is lying on a metal frame hospital bed which is earthed, and their bare foot is sticking out from the sheets and touching the bed frame.

In that example, the zero volt of the battery gets connected to the supply earth in normal condition via the ventilator frame. Then, if the insulation in the finger probe breaks down you could get leakage currents sourced from the 12V flowing into the finger, out of the foot, through the bed frame, supply earth, ventilator frame, X's frame and to the negative pole of the battery.

In reality even this "bad" implementation hazardous currents are highly improbable. Most likely the one or more of the metal frames (X, ventilator, bed) are painted or anodized and this is normally enough to block 12Vdc. Also metal boxes normally have rubber or plastic feet making contact with other earthed metal items less likely. Also the patient's foot contact with the bed is likely high impedance due to skin, in the order of 1MΩ. Finally the insulation in the probe, although unable to handle 1000Vdc, is still extremely unlikely to break at 12Vdc since it has to be also mechanically strong and waterproof to meet the demands of normal use. And even if currents did flow, they are not high severity harm.

All in all it's extremely unlikely even with a "bad" design that uses a metal case directly connected to the battery's negative pole.

Even so, relying on things like patient's skin impedance, paint etc is not good practice and it's better to have a solid plan (intentional isolation) to prevent the 12V from flowing. For this, point I agree with the standard. My main issue with IEC 60601-1 is that the values are not reasonable for battery operated circuits. For example 1MOPP @ 12V requires 1.7mm creepage and 500Vdc dielectric strength (2MOPP 3.4mm/1000Vdc). These values are way too high. Reasonable values might be say 0.5mm / 100Vdc for 1MOPP.