We are looking into setting up a very basic shielded chamber inside a small room at my company to be able to perform some basic relative radiated EMC measurements.

We will buy a somewhat cheap spectrum analyzer together with a set of near field probes and a biconical antenna, and want to have a room where we can perform measurements with as little interference as possible.

So I've been looking into doing a very cheap DIY shielded chamber inside the room, by building a wooden frame and wrapping chicken wire around the frame with thick aluminum foil to seal all joints.

The only thing I still feel unsure about is how to do the filtering for the I/O panel. What we need to get into the chamber is the following signals:

  • 230V/50 Hz power
  • USB 2.0
  • Ethernet
  • Coax antenna connector for biconical antenna
  • Two BNC connectors for DC power supply
  • SMA connector for various RF stuff

So my question is how I would filter all of these signals in suitable ways? Our plan right now is to slap ferrites on all signals, but I don't think that will accomplish very much. What would be a better way?

  • \$\begingroup\$ Which standards will you be using this room for? I'll tell you now that the I/O panel in your room will probably be the least of your concerns. If you can't find your UUT's signature in the noise floor then you're fine. We've tried doing this a few times but it is tough on a budget. We've just learned which spikes are the radio stations nearby, etc. \$\endgroup\$ – Stiddily Jul 6 at 12:49
  • \$\begingroup\$ "Chicken wire" is likely much too coarse for higher frequencies. I'd suggest steel screen similar to window screening, soldered or brazed, or steel sheet. The Earth connections will have to be "overkill" - at least one ground per side, perhaps two or three, and from thick wire, heavily bonded. Because ground wire inductance at high frequencies becomes a problem. \$\endgroup\$ – rdtsc Jul 6 at 13:10
  • \$\begingroup\$ We will mainly be testing towards EN 61326 & EN 61010, but sometimes other standards as well when our products contain WiFi & Bluetooth functionality. \$\endgroup\$ – Lars Petersson Jul 6 at 13:29
  • \$\begingroup\$ @rdtsc Our focus we be around 10 - 500 MHz since that is where we most often have issues with emissions. I'm pretty sure that chicken wire will be insufficient for frequencies as low as 10 MHz though. \$\endgroup\$ – Lars Petersson Jul 6 at 13:31
  • 1
    \$\begingroup\$ Feed-through capacitors and/or filters. \$\endgroup\$ – Cristobol Polychronopolis Jul 6 at 13:59

You can read about "bulkhead" connectors for low/moderate DC power, or logic level control signals.

Otherwise you need a 6" by 6" by 20" power_line filter, bonded to the steel/copper of the walls.

By the way, aluminum oxide forms INSTANTLY upon exposure to air, thus the aluminum will not make ohmic contact between the various sheets.

There will be high capacitance as you crumple the sheet_edges together, but not ohmic.

C = Eo * Er * Area/Distance

C = 9-12 farad/meter * (Area = 2cm * 50cm)/100 micron

C = 9e-12 farad/meter * (100 cm * cm) / 0.0001millimeter

C = 9e-12 farad * 100/0.0001 * cm/meter * cm/millimeter

C = 9e-12 * 1Million * 0.01 * 10

C = 9e(-12 + 6 -2 +1) = 9e-7 = 0.9 uF

which may be adequate even at 10MHz.

| improve this answer | |
  • \$\begingroup\$ Good point on the aluminum oxide. Another option that we have been thinking about is just bonding the sheets of chicken wire together with metal wire which we twist between the sheets, and then solder at regular intervals. That might work better. Do you have any tips on filtering for Coax/SMA cables? For the \$\endgroup\$ – Lars Petersson Jul 7 at 9:43

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.