Changing in device configuration implying in need to retest


Involved In Discussions
This question seems a bit silly to me, but I am trying to find a clear answer throughout the IEC 60601-1 and IEC 60601-1-2 standards, and I just can't. So, I am posting it here to get some different views on the topic.

For a given medical device that communicates with a host computer through a 2.0 USB cable, do I really have to duplicate all safety and EMC tests if I have two options for the cable length, let's say 1 meter and 2 meters, given that the cables are from the same manufacturer and have the same characteristics? (I know that above 3 meters, other tests need to be done.)

Based on my engineering perspective and being very strict, it is very likely that the worst-case scenario is the 2-meter cable, but it is impossible to state this with 100% certainty (maybe the 1-meter is a better antenna for some frequency, I don't know...).

At the same time, it seems like a huge overkill to repeat all tests, especially if we consider what some devices and manufacturers have been doing.

This is similar to stating that the worst-case scenario is with the device at full power, which is likely, but not always the case. Nevertheless, manufacturers still test at both standby (when they do much) and full power to show compliance.

So, what are your thoughts on the matter?

Peter Selvey

Super Moderator
The basic rule comes from IEC 60601-1 Clauses 5.1 and 5.2 which state that a test can be skipped (not performed) if the condition can be evaluated by other methods (5.1), and the selection of samples should be representative (5.2), which includes, for example, selecting a particular type of accessory as representative of a larger range of accessories. Both of these need to be considered on a test by test basis. In practice often >95% of tests can be done with a single accessory as representative, but there may be one or two tests where using the range extremes or all items in a range could be reasonable.

For a USB cable, excluding EMC, it's hard to imagine any role in compliance, so one cable as representative should be fine. If the cable is used to power the device, there is a potential effect on the voltage reaching the device but this depends not only on the cable length but also the cable quality (e.g. conductor size, terminations, connector contact impedance).

For EMC, there is an obvious potential to affect test results. But in truth, EMC is not that critical an area, despite what many will say. The uncertainties in EMC are very high, I've seen ±5dB in test reports, which equates to around ±40%. Also, the layout of the cables can affect the test, but generally the tests will be done with a fixed layout. This makes sense from trying to achieve some repeatability, but it makes no sense from the broader objective of the test. And back in the real world, for a USB cable, the potential for a user to pick up any old cable lying around also exists. All of this means that worrying excessively about testing 1m and 2m length of cables would be inconsistent relative to the myriad of other issues.

The real story is that for EMC emissions, the levels are far far below the level that could cause interference with another device. The low levels are designed to minimize the impact on radio and TV transmissions, and it assumes people tuning in a fair distance from any radio tower. Thus, the limits are specified in in the order of 40dBµV/m which 0.0001V/m, which is tiny, and almost impossible to cause interference that would adversely affect another device. So, worrying too much about the impact of 1m or 2m USB cable on emissions isn't really important.

In contrast, immunity test levels are in the order of 3V/m, some 30,000 times higher. This assumes the device can be located near a source. In this case, depending on the risk of the medical device, the USB cable length and type could be a point of concern. However, given the uncertainties in testing, a practical engineering solution would good design in the USB port, use data check sums and so on, essentially defensive design so that no matter what crappy old USB cable is used, how long it is, at least nothing serious will happen. I'm not an expert in USB design, but I guess there would there is plenty of advice available online or in books, and specialist ICs, ferrites etc for this purpose, or even USB isolation modules.

Of course, regulations (at least now) put a high emphasis on "test reports" so in practice you will still need a test report that shows compliance for EMC, despite all the underlying uncertainties. So, I recommend to pick a cable e.g. 2m and just test so as to have a report and tick the box.


Involved In Discussions
I fully agree with Peter, but also, theoretically speaking, change of length of USB cable could change resistance between your host PC and ground contact (which should be 0,1 or 0,2 Ohm depending on specific design of your device). I've got this in my devices after changing HDMI cable to DP cable from PC system box to display.
Top Bottom