Large industrial systems need a whole different technique for radiated emissions pre-compliance testing.
Last year, I had a chance to perform pre-compliance testing on a couple large systems for radiated emissions. The clients were hoping to get early test data prior to the major work of packing and transporting their large floor-standing systems for formal compliance testing.
Large industrial systems need a whole different technique for assessing radiated emissions. They usually use fixed three-phase power and are comprised of many different subsystems. In fact, most really large systems have their formal compliance testing performed where installed; so-called “in situ” testing, and usually to IEC 61326 for ISM products.
I’ll generally use my “three-step” process for assessing radiated emissions, even for large systems. This process is described in the references below.
Step 1 is to characterize the major emission harmonics using a near field probe and spectrum analyzer. Placing the measurement equipment on a roll-around table is helpful. I’ll measure circuit boards and identify the dominant energy sources, documenting their harmonic profiles and identifying the major 4-5 harmonics. I’ll also keep track of whether the profiles are narrow band, broadband, or a combination.
Once I have an idea of the major energy sources and their characteristic harmonic profiles, Step 2 is to take an RF current probe and measure a sampling of cables (especially I/O and power), since these are normally the major “antenna-like” structures. The harmonic profiles are also documented, identifying the major 4-5 harmonics.
After compiling the spectral characteristics, I now should have a good idea of the dominant harmonics. For this particular example, one of the Ethernet harmonics at 250 MHz was the highest.
At this point, it’s important to determine at what angles the major emission lobes are located. The client was really helpful because he’d laid out a circle 3 m from the faces of the system under test and marked every 30 degrees with blue tape (Figure 1). An average distance, if the EUT is rectangular, is probably sufficient.
Figure 1 Pieces of blue tape were attached to the floor 3 m out from the system under test every 30 degrees.
To determine where the major emission lobes were located, we use Step 3: a simple antenna to identify the actual harmonics being radiated. I took my handheld AIM-TTI model PSA6502T spectrum analyzer, connected to a Kent Electronics 400 to 1000 MHz PCB antenna, and measured the 250 MHz dominant emission at all compass points, recording the approximate amplitude at each point. This gave me a clear picture of the major emission lobes.
Figure 2 Here I’m holding the spectrum analyzer in one hand and the antenna in the other, lined up with the blue tape on the floor, while measuring the dominant 250 MHz emission at each 30-degree compass point.
After identifying the major lobes, I set up the calibrated Chase CBL6111A EMI antenna on its tripod and measured the whole emission spectrum from 30 to 1000 MHz at each major lobe. I used the Tektronix RSA306B spectrum analyzer with its EMCVu pre-compliance software, which quickly plots the emissions versus the test limit and can identify and “subtract out” any non-EUT emissions from ambient transmitters (FM/TV broadcast, cellular, etc.). Knowing the major lobes in advance reduced the number of measurement points required.
While the EMCVu software pre-calculates margins from the test limits and accounts for system gains/losses automatically, if you’re performing manual measurements – that is, taking each major harmonic and comparing to the test limit at that frequency – you’ll need to account for all the system gains/losses in the measurement setup manually, as well (Figure 4).
Figure 4 A manual pre-compliance measurement will require accounting for all the system gains and losses to calculate the E-field captured by the antenna at each harmonic frequency.
To manually calculate each major harmonic in relation to the test limit specified by the standard, you’ll need to add and subtract each factor at the frequency of the harmonic, according to this equation:
E-field(dBμV/m) = SpecAn(dBμV) – PreampGain(dB) + Attn(dB) + CoaxLoss(dB) + AntFactor(dB/m)
Sometimes, we place a 6-dB attenuator at the antenna port to “level out” the port impedance, because it can change a great deal versus frequency. If the harmonic signal can be observed without the use of a preamplifier, then that factor can be ignored.
The coaxial cable loss is usually not significant (1-2 dB for short lengths), but can either be estimated from the manufacturer’s value of dB/100 feet, or better, actually measured with a vector signal analyzer or by using the tracking generator option on the spectrum analyzer (see Reference 1).
Note that if using a log-periodic antenna, similar to the front elements shown in Figure 3, there should be a marking on the antenna boom where the electrical “center” of the antenna is and this is the point that should be lined up with the 3-m marking on the floor.
Notice all those ferrite chokes along the antenna coaxial cable? They help isolate the cable from the measurement. Ideally, the cable should be positioned behind the antenna, which will help even more. Be sure no one stands in front of, or near, the antenna during measurements.
One problem you’ll run into immediately when testing conducted or radiated emissions outside of a shielded room, is the number of ambient signals from sources like FM and TV broadcast transmitters, cellular telephone, and two-way radio. This is especially an issue when using external antennas. I’ll usually run a baseline plot on the analyzer using “max hold” mode to build up a composite ambient plot. Then, I’ll activate additional traces for the actual measurements. For example, I often have at least two plots or traces on the screen: the ambient baseline and the actual pre-compliance measurement.
Fortunately, there are three ways around this:
After all the calibrated measurements were done, I was able to provide a good estimate of pass/fail for radiated emissions for this system. The whole process took about a day.
Assuming you end up with adequate margins from the limit, I’d feel comfortable using this data in the test report to claim self-compliance for those companies wishing to go that route. However, if there are any questions or if the ambient spectrum is so noisy that an “engineering judgement” might be a little questionable, then I’d have a third-party EMC test lab confirm the results for your report.
This article was originally published on EDN.
—Kenneth Wyatt is president and principal consultant of Wyatt Technical Services.