2
\$\begingroup\$

I want to measure the EMC radiated by a circuit board, not very accurate, but just to give me a hint of how much is radiated.

The frequency of interest is 30MHz to 1GHz.

My idea was to place an antenna in the far-field, and simple terminate the antenna with the correct termination load.

Then measure the voltage across the load (using my 250MHz oscilloscope with spectrum analysator), and from this voltage calculate the dBm, dBuV/m and ect.

(Yes I know my 250MHz oscilloscope, will only give me the spectrum up to 125MHz).

Now for the questions: 1) Is it possible to get an estimate of the radiated EMC in this way? 2) What antenna design to use?

\$\endgroup\$
4
\$\begingroup\$

Note that one wavelength is 10m at 30MHz, so at least the measurements at the low end of your frequency range will likely be in the near field.

In my experience, the only way to really measure what it going on is to use a calibrated test range. However, that doesn't mean you can't get useful, but mostly qualitative, measurements on your own. Even then a spectrum analyzer is very useful. Don't worry about near field and use the spectrum analyzer to identify frequency spikes. You won't know if they are above or below the limit without a proper calibrated test, but you can get a feeling where your likely trouble spots are.

You can then do the usual cheap things (series ferrites, common mode chokes, caps to ground, etc) to reduce the emission spikes you see. Your lab measurements will at least give you a reasonable idea of how much each of your fixes reduce the problems.

Then you go to a calibrated test facility and find what you missed.

\$\endgroup\$
3
\$\begingroup\$

Based on reasonably extensive practical experience, I've found that you can get an amazingly useful feel for radiated energy by using a transistor radio placed in close proximity to the target and tuned over AM and FM bands. This is useful enough to get an excellent feel as to whether your system is likely to be terrible, reasonable or (probably) extremely good.

Obviously this is utterly informal and subjective, but you get a feel for degree of severity and more. You'll quickly find the relative usefulness of AM and FM bands. Sometimes breakthrough and intermodulation on the lower band may be more useful than direct reception in the FM band. Even if your device has no formal radiation outputs in the bands above, there's a fair chance that, placed close enough, you'll still get useful results.

Also useful is tuning a TV across its whole tuning range and observing the screen. Best for this are wholly manually tunable - the ulra cheap portable black and white ones with something like a 6" screen used to be very handy for this.

Of these two methods tha radio is probably the most useful. Best are radios whose whole tuning range can be swept with single twist of the wrist. That is, you don't want some thing geared down that takes many turns to scan the range - although that probably has its place for spuries that are in a known location.

When you can operate a transistor radio on or within a few inches of a device under test with essentially no interference it tells you that you are winning - or that it's not turned on :-).

\$\endgroup\$
2
\$\begingroup\$

Is it possible to get an estimate of the radiated EMC in this way?

It'll be very had to get any useful results without the proper equipment. If you're just curious and want to experiment with the principles, sure, you might get some results on the scope. You should be able to identify the frequency of spikes in emissions so you can target those points. However, don't expect to get useful power levels to determine whether or not you're in compliance.

Your local university may have a test lab or class if you want to learn stuff like this.

What antenna design to use?

It's difficult to span 30 MHz <-> 1 GHz with one antenna; harmonics get in the way quickly. However, there are several broadband antenna designs that could be useful for this.

A standard EMI antenna is the Biconical antenna:

enter image description here

It may be more difficult to precisely construct other wideband antennas like the bilog log-periodic configurations.

\$\endgroup\$
  • \$\begingroup\$ or, if you're just looking for spikes and don't mind apples and oranges comparisons, just use two or three different ones depending on the range you're looking at. \$\endgroup\$ – JustJeff Oct 21 '11 at 0:20
  • \$\begingroup\$ Ok, thanks. I was thinking about using dipole for higher frequencies. And not sure what to use for lower frequencies. \$\endgroup\$ – JakobJ Oct 21 '11 at 10:23
  • \$\begingroup\$ A biconical antenna 1.5m or so across will do a fine job of reading 30 MHz signals. These can be quite large in some test labs. \$\endgroup\$ – Kevin Vermeer Oct 21 '11 at 11:01

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.