Denver, CO. EMC 2013 convened here last week with a variety of technical sessions and, on the exhibit floor, product highlights. Sessions covered topics ranging from spectrum analysis to antennas, while on the exhibit floor vendors highlighted products ranging from simulation software to EMI test receivers.
On Tuesday, I attended “Special Session: Spectrum Analysis and Measurements in a Congested Electromagnetic Environment” (chaired by Bob Johnk and Sarah Sequin).
Brian Cordill, University of Kansas, discussed the results of simulated interference to radar receivers from in-band OFDM systems. The simulation was very basic, involving just one receiver and one source of interference. A comment from the audience suggested that the simulation should be made more realistic and less ideal. However, analysis soon becomes too difficult and the suggestion was made that a real-time spectrum analyzer was a good tool to capture the actual interactions occurring in the real world.
Chriss Hammerschmidt, Institute for Telecommunications Sciences, described the work she and her colleague Heather Ottke had been doing to quantify spectrum usage at various locations across the US. They measured and analyzed spectrum usage in San Diego, Chicago, Denver, and at Table Mountain. The presentation was very well prepared and delivered, the graphs in particular being significant. Spectrum monitoring involves time: all parts of the spectrum are not used in the same way and neither are the signals identical in different frequency ranges. The measurement techniques have to be tailored to the signal characteristics. After the data has been acquired, its analysis has to account for the usage statistics. Because the graphs represent high levels of data compression, their development is important as well as the actual trends they show.
Sarah Seguin, University of Kansas, discussed work done to optimize radar waveforms according to the actual emitted signals. Unfortunately, this presentation was very general, but the approach actually was tried to good effect with experimental hardware.
Bob Johnk, Institute for Telecommunication Sciences, reviewed work done several years earlier to develop a security capability based on radar. An array of 8’ high poles was erected with a radar transceiver at the top, center, and bottom of each pole. The poles were about 30’ apart and formed a grid that, via software, behaved like one large interconnected radar. Apparently, this approach has been very successful in trials and is intended to be used as an electronic anti-intrusion fence. The resolution is such that a dog can easily be distinguished from a crawling human.
On the exhibit floor Tuesday, Agilent Technologies highlighted EMPro 2013 electromagnetic simulation software, which simulates radiated emissions, correlating the electrical performance to circuit board traces. New FEM hybrid boundary conditions provide higher simulation speed and efficiency. The PXA and MXA signal analyzers were on display with their optional 160-MHz realtime analysis bandwidth being featured. I was given a thorough demo of the N9038A MXE EMI receiver. It features the capability to move easily between EMI-receiver and signal analyzer modes. Up to three separate detectors of different types can be specified to run simultaneously. In addition, the ENA Series E5061B VNA now features an RF network analysis (NA) option as well as an impedance measurement (ZA) option. The company’s existing N/W6141A EMC measurement application was also on display. This application lists signals out of specification, employs realtime detectors to measure selected signals, and supports re-measuring signals for easy verification that failed signals have been repaired. And finally, the remote operation of the FieldFox handheld analyzer was demonstrated.
Rohde&Schwarz demonstrated the new ESRP EMI test receiver that speeds up precompliance measurements. I met with James Young, EMC sales specialist. Although he has an RF background, he was very enthusiastic about the FFT capabilities offered across the R&S scopes and signal analyzers. The approach taken allows instrument control from the viewpoint most comfortable for the user. Time domain instruments, primarily scopes, can be controlled by span, center frequency, and RBW when displaying frequency domain data. Similarly, signal analyzers can be given amplitude and time-base settings when displaying time domain signals.
Here are some other highlights on display:
• The BBA150 amplifier covers two frequency ranges: 0.8 GHz to 3 GHz with power from 30 W to 200 W and 2.5 GHz to 6 GHz with power from 15W to 100 W.
• The ESR family of EMI test receivers now extends to 26.5 GHz. The new receiver’s broadband architecture enables standard-compliant disturbance measurements up to 6,000 times faster than other testers.
• A new EMI option for the FSW high-end signal and spectrum analyzer supports EMI bandwidths in compliance with commercial as well as military standards. The option has a selection of detectors as well as a database of predefined limit lines.
• To answer the need for better testing of the new ETSI EN 300 328 requirements for smart device operation in the 2.4-GHz band, the company has developed the TS8997 test system.
In a meeting with Ozgur Ozturk, senior business manager at TÜV Rheinland, Youngsville, NC, I learned that EN 60601-1-2 will soon present significant challenges to medical device manufacturers. In addition to covering susceptibility to external interference, which the standard already deals with, the changes include risk assessment. In particular, the manufacturer must provide a risk management process (clause 4.2). The risk assessment based on this management process determines which tests must be performed. A risk management process defined in ISO 14971 is necessary to demonstrate compliance.
I had a discussion with Glen Watkins, ETS Lindgren, about publishing and media in general. He wondered just what the motivation might have been for Jeff Bezos’ Washington Post purchase. About then, Ron Brewer walked by. The discussion was wide ranging, with a very detailed recounting of the Don White automated EMC test machine and its demise followed by the rise of the Don White EMC publishing empire—yes, publishing was the common thread.
On Wednesday, I attended the “Measurements General” session. “Investigation of the ESD Induced Clock Disturbances in Portable Electronic Products” was nominated for the best symposium paper award and was presented by Viswa Pilla, Missouri University of Science and Technology. The authors examined the levels of electric or magnetic fields required to cause soft failures in a number of electronic products.
Basically, they injected an increasing amount of noise until failures occurred. It matters how the noise is related to the clock phase. In particular, many products have an internal PLL, and spectrograms of the PLL output clearly showed deviations caused by the injected signals. The short-time FFT (STFFT) was used for analysis.
A paper presented by Edward Savage from Metatech Corp. in session WED-AM-5 (TC5) reviewed the results of network protection device tests. A number of Ethernet protection devices were subjected to a series of pulse tests. The EFT test generator produced a 5/50 ns pulse while the CWG and telecom generators provided higher energy pulses of 1.2/50 µs and 10/700 µs, respectively.
Two aspects of the tests were important in addition to the survival of the protection device. The device had to clamp the input waveform very quickly and to a low level. In addition, the output of the device had to remain below a level that would damage connected circuitry. None of the devices operated faster than about 10 ns and few limited the input amplitude. However, one device with coaxial construction showed very good output performance and survived all the tests. Smaller devices that used diode or IC protective components did not fare nearly as well, arcing being evident in one device and physically broken chips in another. For many of the devices, repeated pulsing lead to destruction even though they appeared operational after one pulse. The authors’ recommendations are for better device descriptions to be provided by the manufacturer. The test conditions must be clearly spelled out and the detailed results presented.
On the exhibit floor I met with Greg Senko of TESEQ. With the addition of IFI and Milmega products as well as the development of products that address new spec requirements, TESEQ is rapidly expanding. And, the company has a new exhibition booth, used for the first time at this show.
Several CDN products have been developed to address specific needs. For example, one product now allows Ethernet immunity testing while monitoring device performance. In the past, the Ethernet device would be disconnected from the network, tested by injecting interference, and then reconnected to determine if it still worked correctly. The new CDN supports continuous monitoring during injection so that the degree of impairment can be determined.
In another case, a new LISN addresses PV inverter testing. Inverters switch at high frequency, which may cause EMI on the input DC lines. There have been LISNs appropriate for inverter output testing for some time, but this one allows input testing.
I also met with Chris Loberg of Tektronix. The company’s 33-GHz active probe was demonstrated connected to one of Tek’s high-end scopes. Although an eye diagram was being displayed, the real power of the setup was the compensation and de-embedding provided by the software. Impairments such as 30” of FR4 PCB trace could be applied to the acquired signal. In addition, you can specify the type of equalization to apply to recover an open eye at the receiver. These functions are simply stacked up by the software to account for the characteristics of the signal path. In particular, the 33-GHz probe is calibrated from the probe tip and the S parameters for that particular probe are loaded into the scope when the probe is connected, so the probe’s effect can be compensated for.
The MDO4000 also was on the stand as well as a 5000 Series RTSA. Both products stress Tek’s time-domain heritage but also integrate RF capabilities. Most important, when analyzing a signal in both domains, a marker in one display is time correlated to markers in the others. This means that in addition to viewing a signal’s spectrum, you can correlate activity in the time domain with a change in the spectrum—a powerful troubleshooting tool.
Ken Javor of Pearson Electronics, a consultant associated with the company, said Pearson manufactures wide-band current probes with frequency response from about 10 Hz to 200 MHz. Recently, Javor developed a power-line ripple detector to aid in power-line ripple injection testing. Basically, the device is a passive signal conditioner that acts as an attenuator, impedance match, and isolator. This type of testing is called for in CS101, RTCA/DO-160 Section 18, and similar standards.
Glen Watkins and Charlie Ferris of ETS-Lindgren highlighted one product recently launched by the company that they said has far exceeded expectations. This product is the flexible RF attenuator pyramid—the FlexSorb absorber. The company has made pyramidal absorbers for many years, but a common complaint, especially for absorbers near the door in a large chamber, was breakage. The FlexSorb product solves that problem and actually results from a process that can be applied to most of the company’s absorbers to make them flexible. By the way, the FlexSorb is just one of the many products made by ETS-Lindgren, in-house, 100% in the USA. This includes antennas, chambers, and positioning devices.
And, on the topic of antennas, I attended the WED-PM-1 TC2 session on antennas, chaired by Zhong Chen from ETS-Lindgren. Harry Gaul from General Dynamics described work he did relating to “Electromagnatic Modeling and Measurements of the 104-cm Rod and Biconical Antenna for Radiated Emissions Testing Below 30 MHz.” His fundamental finding was that grounding makes a lot of difference.
Compared to a FEKO simulation of the rod antenna, very close agreement was observed in an ungrounded antenna that used fiber optics to couple the signal. Grounding of several forms, some with a ferrite bead on a coaxial lead, all lead to results several dB higher than when the rod was not grounded. The problem is that grounding and the use of a ferrite bead is specified in MIL-STD 461F, so a spec change is required that incorporates Gaul’s findings.
The biconical antenna sometimes is specified for use at 20 MHz to avoid uncertainties associated with the rod antenna. However, Gaul showed that a biconical antenna is operating far below its optimum range at 20 MHz and under-reports RF levels by several dB compared to an ungrounded rod antenna.