Check the lab for false radiated emissions

-August 26, 2016

Successfully troubleshooting EMI/EMC problems isn't easy. You need time, you need a strategy, and you need to remain calm. Sometimes, you need a special "sixth sense" that only comes from experience because in some cases, your problems aren't even related with the EUT (equipment under test).

Let me explain using two examples of unexpected "problems" while troubleshooting radiated emissions that weren't caused by my EUT. One instance occurred in my lab and the other at a small company close to my University.

While in my lab one day, I was trying to solve a radiated emissions problem from a device. Usually, I have some kind of emissions results from a laboratory with a shielded chamber—my reference data—because I don’t have a chamber. Then, I try to find the frequencies with problems in my lab, identifying the background noise and then debugging as the EUT goes through its operations.

In this case, you can see in Figure 1 (left) how the background noise was before measuring the DUT. Emissions from FM broadcast transmitters in 88-108MHz are clearly identified. That measure was saved in the screen as a reference.

Figure 1 The left plot shows background noise while the right plot shows emissions (red) and background noise (blue). But, the background noise is higher in the 135 MHz-235 MHz range. Why?

Then, we measured the EUT's emissions, shown in the right plot in Figure 1. Emissions from a digital clock were found, but note the additional broadband emissions in the range 135 MHz-235 MHz. We tried to find those emissions in our DUT for some time without success. Note the higher background noise was not present in the original (left) plot's blue trace.

To further complicate the problem, we found the emissions were not there all the time (Figure 2).

Figure 2 The higher background noise masked some emissions from the EUT.

We tried to find the "intermittent" emissions around the PCB of the EUT without success. What was wrong?

I then noticed that my soldering stations (Figure 3) were sometimes on and sometimes off, as needed.

Figure 3 The instrumentation in my lab includes the soldering stations, as many do.

Figure 4 shows the background noise as measured in both situations. You can see the difference.

Figure 4 Background noise with and without soldering stations (EUT off).

So, we found that a "harmless" soldering station was confusing us and hindering our ability to debug the problem for some time. EMI is wonderful, isn’t it?

The second example is similar to the first. We were testing a system with a power electronics converter at a company close to my lab.

While measuring radiated emissions at 3 m, we noticed a considerable level of emissions in the low frequency area as shown in Figure 5.

Figure 5 These plots show background noise (yellow) and emissions from DUT (purple).

With a near-field probe, we found that signals weren't radiating from the EUT. It looked like the noise was not from the background (yellow color STOPPED in screen). What was wrong?

Armed with the experience of the soldering station, I began looking around the room. That's when I found that my portable spectrum analyzer, used with a set of near field probes, was using the charger as shown in Figure 6.

Figure 6 My portable spectrum analyzer was connected to AC mains through its charger.

I removed the charger from the wall. The emissions disappeared from the instrument screen (Figure 7).

Figure 7 Unplugging the spectrum analyzer's charge reduced background emissions.

These experiences prove that you must be careful when troubleshooting radiated or conducted emissions in your lab. Check your tools during your troubleshooting process. Check no chargers, soldering stations, lamps, etc. are emitting in your frequency band. Check background noise continuously. Check, and check again.

Keep in mind that these types of problems can also appear in a shielded chamber. While debugging an emissions problem, check to see if you're using soldering stations or chargers inside the room.

Have you some similar experiences? I would like to hear your stories.

Also see:

Loading comments...

Write a Comment

To comment please Log In