EMC / RF Rejection

[Sam Jelfs]

Hi all,

 

Possibly somewhat outside of the scope of this forum, but I guess some of the people on here might be able to help Working on a prototype system at work, and due to the way that it will be used it has to pass EMC testing.

 

The important parts are a Foxconn NT535 'nettop' pc, with an M Audio Fasttrack USB sound card, and an Earthworks M23 reference microphone.

 

When testing for susceptibility for RF it is placed inside a test field of a carrier somewhere between 150MHz and 1GHz, modulated at 1kHz, with a strength of 3v/m. At this point we look at the frequency response of the received microphone signal and can clearly see peaks at 1kHz (as analysed on the Foxconn computer running some custom (and tested) analysis scripts). If we don't have the microphone connected but do have the mic cable connected we don't see peaks, connect the mic and the peaks return. We have tried a couple of different combinations of mics and cables, all to no avail, have even tried a different usb soundcard (with external PSU instead of bus powered) and still the 1kHz peaks return...

 

So the question is, does anyone have any suggestions of what to change / test to see if we can remove this problem of RF leaking into the system? It is basically the same problem as a GSM phone breaking into a PA system, and typically I would expect the use of shielded cables to help stop that, but it doesn't seem to in this instance.

 

Cheers

 

Sam J

[paulears]

I don't know the exact Earthworks mic - and the spec describes it as a pre-polarised diaphragm - but I wondered if it's a RF design - these used an oscillator at 9/10MHz or so - so I'm wondering if you've created some intermodulation with the applied RF field you are generating. When you swapped the mics, did you try a dynamic I wondered if you discounted them as they're obviously no use for measurement purposes, but could possibly show you where the RF is getting in.

 

Are you testing the microphones, or are the microphones listening to the devices under test - Not quite sure how your test regime works?

[Sam Jelfs]

the other mics we had to hand were also reference mics, couldn't actually find a dynamic mic in any of our labs... We shall hopefully get some more time in the chamber soon, but as it's a different department we have to pay per hour, so the less time we spend in there the better :D

 

The mic is part of the DUT, so we have computer, usb sound card, mic and cabling all in the field as a single system (the system will eventually be used for noise monitoring and recording). Without the mic connected, no peaks, with the mic connected, peaks at 1kHz + higher harmonics. This seems to suggest either the microphone, or the mic cable is picking up the RF modulation. but if it is the cable why should it matter if the mic is connected or not. We also get the same peaks if the mic is connected but phantom power turned off.

[Brian]

At the frequencies you're testing at, the screen on your mic cable is anything but a screen; it's going to look like an aerial. Let's assume that the EUT end is a good solid earth. The RF will induce a voltage in the screen which will cause the mic end to rise in voltage. This then gets transferred by the mic to the signal conductors which then go down inside the cable to your soundcard were it get demodulated.

[Sam Jelfs]

Thanks for that brian, its kind of what I was thinking, but this is a bit outside my range of experience.

 

I would be concerned about assuming that the DUT end of things is a good earth. The micro-pc and soundcard have no earth connection, both run through class-2 DC power supplies. I am yet to investigate exactly how the shield pin of the XLR sockets, plus ground on the USB sockets on the pc are bonded, thats tomorrow mornings job. Anechoic chambers, mesh floors, and small screws seemed like a bad combination today.

 

So the next question, how to get rid of it. In real terms, given the location of the system, it will never be subjected to these levels of RF field, but to get through the safety testing and get the bits of paper we need we have to reduce the systems susceptibility to RF some how.

[Lamplighter]

Sam

It is probably worth your time to read the words of wisdom on this subject from Jim Brown. Try googleing Jim Brown pin 1.

Brian

[DrV]

I don't know a lot about EMC but a quick fix just might be to put a ferrite bead around the mic cable, as close as possible to where it disappears inside the box. The sort of thing I'm thinking about is the beads you usually find around VGA cables and the power leads on laptops. The principle is that high frequency induced voltages see this as a low pass filter choke so never get inside the box (or out of it in these two examples). It would certainly be a cheap solution if it did happen to work.

 

Good luck - it's a black art as far as I can see!

 

Dave

[jamesperrett]

In the old days, Hifi mags would regularly run articles on how to reduce RF interference as it was a common problem with old gear. Inductors and capacitors in the right places are your friends - but I suspect you need to draft someone suitably experienced in to show you the right places in your kit. You will almost certainly need to open up the audio interface and modify it to achieve decent results.

 

James.