CE Mark Surge Testing: Protecting Wireless Products

[dwsexton]

TL;DR

How to pass CE with no ground?

I have a battery powered wireless product in a plastic IP68 enclosure with no exposed metal that attaches to a sensor through a cable that is about 6 feet long that again is completely encapsulated with no exposed metal. I want to get a CE Mark certification and since this for outdoor use my understanding is that the surge test must be applied to each wire in the cable, or the shield of the cable if there is one. Currently there is no specific surge protection on the signals, inputs or outputs. Since there is no conductive path to earth ground I suspect I will only see magnetic coupling. My question is, what is commonly done if anything is done at all to protect inputs and outputs so that no damage is done to the device during the test. The circuits are 3.3V level logic signals. Note, I have currently not run any tests yet but am trying to anticipate issues before I pay for lab time to start testing.

 

[berkeman]

Welcome to PF.

You should ask the lab about this -- there may not be an applicable Surge test for such a device. If there were exposed metal, especially on the cable, you would use the "IO Line Coupler" test of EN 61000-4-5 (Surge), and not the "Powerline Coupler" test. The hits would be in common-mode with respect to Earth ground, and the only return path through your device would be capacitive coupling to the ground plane on the table under your device (with a 10cm non-conductive spacer, IIRC).

If that could disrupt your circuit, you would want to add a metal ground plate under your PCB inside your enclosure, and add appropriate clamping for the IO lines to that plate. That will provide a path for the parasitic Surge transients to return to the test table ground plane without going through your circuitry.

If you have no exposed metal, there is nothing to conduct the transients to from the Surge tester. Kind of like trying to Surge test a plastic walkie-talkie...

 

[berkeman]

BTW, you likely will need to do the EN 61000-4-4 EFT/Burst test, since it injects noise into your cable using a 1 meter clamp. Are you familiar with that test?

And the EN 61000-4-3 RF Immunity test, since that does not rely on any exposed metal. You just get blasted with RF noise in an RF anechoic chamber, and need to keep operating at whatever intensity level you are trying to qualify for.

EDIT: change -6 to -3 for the Radiated RF Immunity test. See my later post for more details.

 

[dwsexton]

Thanks for your replies, I have been in close contact with the Lab on this. Currently I have a shield on my cable but it is not connected at either end. What the lab has asked for is that they have access to the shield at both ends to do the test. My product is effectively like a plastic walkie talkie as you jokingly mentioned and that's what makes this rather confusing.

I have a product from another vendor with a similar configuration and they have CE mark. Yesterday I got their test report and discovered they only tested for indoor applications so the surge test did not apply (even though their product is really an outdoor product). They have the CE stamp and I guess it's up to the customer to dig deeper.

So my plan now is to add ESD diodes to all the I/O lines and maybe add some small ferrites just in case, bring out the shield at both ends of the cable and see if it passes the surge test and just eat the lab fee for a precertification test. I don't anticipate an RF immunity problem and because there is nothing conductive to zap with an ESD gun I am not too worried about that either. We already did the emmissions testing when we tested for part 15. If we don't pass the surge test or can't easily get there then I will just certify for indoor as others have done.

Any thoughts?

 

[berkeman]

What the lab has asked for is that they have access to the shield at both ends to do the test. My product is effectively like a plastic walkie talkie as you jokingly mentioned and that's what makes this rather confusing.

IMO, that makes no sense. You should not have to open up your product and make extra connections for any CE Mark testing. CE Mark Testing is meant to test the product as it is normally used in applications. I'd suggest you get a 2nd opinion from another lab before letting this lab do that test.

And having a floating shield doesn't really help anything. Is there a reason you're using shielded cable without connecting the shield anywhere?

 

[berkeman]

TL;DR Summary: How to pass CE with no ground?

I have a battery powered wireless product in a plastic IP68 enclosure with no exposed metal that attaches to a sensor through a cable that is about 6 feet long

Are there any seams in the product or cable connectors or at the sensor that can provide a creepage path for 15kV ESD hits? If so, then you will want to protect those areas, or do something to lengthen the creepage path so that 15kV arcs will not make it inside. Other than the -4 Burst test on your cable (with the 1m IO cable clamp) and the -3 Radiated RF susceptibility test, I don't see many other CE Mark tests that apply.

EDIT: Both the -3 and -6 RF Immunity tests may apply, since you have the IO cable to your sensor. The -3 test is the one done in the RF anechoic chamber with RF energy beamed at your product from an antenna. The -6 test uses a current transformer toroid to inject RF energy into your IO line to your sensor.