Radiated Emissions Measurements Using an EMI Receiver

Radiated Emissions Measurements Using an EMI Receiver

Andrew Ware
Hewlett-Packard Co.,Santa Rosa Systems Division
Santa Rosa, CA

Electronic products designed for global markets must meet both national and international standards for electromagnetic compatibility (EMC). Measurements based on these standards fall into two basic categories of electromagnetic (EM) testing: radiated and conducted emissions testing and radiated and conducted immunity testing. Radiated emissions testing is designed to identify signals emitted into space from a device under test, typically at frequencies from 30 to 1000 MHz. However, measurements can be performed at much higher frequencies since the US Federal Communications Commission (FCC) requires testing to 200 GHz for intentional radiators such as wireless transmitters that operate with carrier frequencies above 30 GHz.

Because they are performed in an open-air environment, radiated compliance emissions tests are somewhat more difficult to perform than conducted emissions tests, which are accomplished in a tightly controlled environment. (However, radiated compliance measurements can be performed in 10 m chambers.) As a result, proven guidelines and test procedures should be followed to optimize a test setup for the most accurate radiated emissions measurements. This article reviews radiated emissions measurements and presents guidelines to optimize measurement performance using an electromagnetic interference (EMI) receiver.

Reviewing Regulations

Any product that moves an electron must pass applicable EMC requirements for the area where that product is to be sold. Several factors determine the type of EMI testing to be performed, including where the product will be sold, the product's classification (industrial, scientific, medical, automotive or information technology (IT)) and how the product will be used. For example, in the US, the FCC and the American National Standards Institute (ANSI) have established extensive regulations for EMI testing. The FCC's Part 15 requirements apply to broadcast receivers, digital devices, switching power supplies and fluorescent lights, while its Part 18 requirements apply to industrial, scientific and medical products.

In Europe, Comite International Special des Perturbations Radioelectriques (CISPR) and European Norm (EN) regulations provide the guidelines for EMI testing. For example, CISPR 11 and EN 55011 regulations apply to scientific and medical products. Typically, class B products are used in domestic establishments while class A products are designed for industrial use. Products are further broken down into two groups: Group 1 defines medical and scientific equipment such as weighing machines, signal generators and spectrum analyzers; group 2 includes industrial induction heating and dielectric heating equipment and domestic microwave ovens. CISPR 14 and EN 55014 apply to electric-motor-based products, while CISPR 22 and EN 55022 apply to IT equipment. Table 1 lists these emissions regulations.

CISPR 16 regulations apply not only to scientific instruments but also to the EMI receiver used to check those instruments for compliance. Such a receiver should provide absolute amplitude accuracy of ±2 dB from 9 kHz to 1000 MHz. Filter bandwidths (6 dB bandwidths) are also specified for this receiver at 200 Hz from 9 to 150 kHz, 9 kHz from 150 kHz to 30 MHz and 120 kHz from 30 to 1000 MHz. The frequency response of these filters is also specified by CISPR regulations.

In addition, the regulations require the use of different detectors: peak, quasipeak and average (see the sidebar on p. 114). The charge and discharge rates and time constant of the quasipeak detector are also specified. The input impedance is specified at 50 W with nominal SWR values. Specific harmonic/intermodulation requirements are also specified. The receiver must pass a CISPR pulse test and product immunity testing in a 3 V/m field.

Current FCC regulations usually call for maximum test frequencies to the fifth harmonic of the highest clock frequency for an unintentional radiator (such as a computer) and to the tenth harmonic of the highest clock frequency for an intentional radiator (such as a cellular telephone). In addition, current FCC regulations and proposed CISPR regulations require a 1 MHz measurement bandwidth for measurements above 1 GHz, and no quasipeak detector is required for measurements above 1 GHz. The CISPR pulse test is not required above 1 GHz, but high sensitivity in the measuring system is important to achieve sufficient dynamic range. The EMI receiver should be equipped with preselection to reduce broadband noise at the front end. The instrument should also be capable of a noise floor low enough to measure signals at low pulse repetition frequencies.

Making Measurements

Performing radiated emissions measurements is not as straightforward as performing conducted EMI measurements. There is the added complexity of the open-air ambient environment, which can provide additional signals that interfere with readings of the equipment under test's (EUT) emissions. Fortunately, methods exist to differentiate between signals in the ambient environment, such as cellular telephone signals and television broadcast signals.

Radiated emissions measurements involve the measurement of EUT emissions in an open-area test site (OATS). ANSI C63.4 and CISPR 16 documents specify the requirements for this test site, including preferred measurement distances of 3, 10 and 30 m and antenna positioning at 1 to 4 m heights. The CISPR guidelines call out a measurement area termed the CISPR ellipse. This ellipse is defined by a major diameter 2X and minor diameter (3X)^1/2 where X is the measurement distance. The ellipse must be free of any reflecting objects, with a metal ground plane for the measurement area.

Any outside area used for radiated emissions testing must pass a normalized site attenuation (NSA) test prior to measurements. The NSA determines the amount of attenuation imposed by the receiving antenna on a wave from the transmitting antenna (EUT) referenced to a signal directly transmitted via cable. The receiving antenna actually picks up a combination of direct and reflected waves. The wave attenuation of the OATS must fall within a specified accuracy band. The test is performed at the same distance at which the compliance test will be performed, as shown in Figure 1 . Details on OATS requirements can be found in CISPR 16 and ANSI C63.4 documents; details on the construction of an OATS are available in ANSI C63.7.

Assuming that a measurement setup is compliant with the applicable standard, radiated emissions tests can be conducted by following some basic guidelines. The radiated test setup shown in Figure 2 incorporates a measurement antenna that is placed 1 to 4 m above the ground plane. The EUT is placed on a nonconductive table that is 80 cm high. The antenna and EUT first should be separated by 3 m (or 10 m if required by the regulations). Both CISPR and ANSI requirements call for the EUT to be in a worst-case operating mode, that is, a mode in which it is likely to radiate an EM field. The EMI receiver should be set up for the correct resolution bandwidth, antenna correction factors and limit line for the applicable requirement.

Antenna correction factors will depend upon the EMI measurement receiver used. For example, the HP 8542E/8546A series EMI receivers have two preset radiated emissions test bands: 30 to 300 MHz and 300 to 1000 MHz. The lower frequency band uses a biconical antenna while the higher frequency band uses a log-periodic antenna (although the range from 30 to 1000 MHz can also be handled with a single broadband antenna). By loading typical antenna factors into an EMI receiver, the unit's display will be corrected for the loss of the antenna. For field-strength measurements, the level will be shown in units of dBmV/m.

In order to evaluate an EUT's radiated emissions, ambient EM levels at the test site first must be determined, as shown in Figure 3 . The ideal situation is for all of the ambient signals to be below the measurement limit line. In many cases, the ambient signals may be above the limit line, so it is a good idea to measure these signals with the EUT turned off and store the results in the receiver's internal memory for comparison to measurements made with the EUT turned on.

When limit lines have been set in the EMI receiver (or a predetermined test mask is used), an automatic measurement can be performed to capture signals above the limit line. Measurements are made with peak, quasipeak or average detection, depending on the test requirements.

Placing the EUT

Radiated emissions from electronic devices are not uniform. The strongest emissions may emanate from the rear or front panel or slots in the shielding. Proper positioning of the EUT can ensure that worst-case emissions are captured by the EMI receiver.

With the EMI receiver adjusted to view the span of interest, the EUT should be moved through a 360° rotation in at least 45° increments. At each 45° step, the amplitude of the largest signals should be noted. With a printer connected to the receiver's I/O port, the screen can be printed at each 45° interval for comparison purposes at the end of the 360° measurements. The position of the EUT should be indicated clearly on each printout.

After all the printouts have been made, they should be compared to locate the position of the worst-case emissions. In some cases, worst-case emissions may appear for different frequencies at different rotational positions. For example, a worst-case emission may be found for 100 MHz emissions at the 90° rotational interval while the worst-case emissions for 200 MHz are found at the 270° position. In this case, the subsequent emissions tests must be performed at both positions. If it is uncertain whether the measured signal is an ambient signal or from the EUT, the EUT can be switched off; the level of the ambient signal will not change. Worst-case emissions must be determined for both horizontal and vertical signal polarizations.

With the EUT turned on and oriented to the worst-case position, automated testing can be performed in a manner similar to that performed for the ambient signal testing. The difference is that the results stored in the receiver's internal memory will contain ambient signals plus signals from the EUT. To remove the ambient signals from the second set of measurements, it is possible to perform a sort on the two sets of signals using automatic functions in the test receiver. The process involves using a receiver function that automatically marks all duplicate signals between the two sets of measurements and then deletes them. What remains are the signals from the EUT minus the ambient signals common to the first and second sets of measurements.

At this point, most of the ambient signals have been deleted from the internal list. Some ambient signals may have been present for one set of measurements but not for the other, in which case they would not be duplicated or deleted. The remaining signals are the peak, quasipeak and average detected values of the EUT's emissions and undeleted ambient signals.

The next step is to locate signals that lie above the specification limit line by sorting the list of quasipeak values with the highest levels listed at the top of the list. Then, the column of signals that indicates the value of the quasipeak measurements vs. the limit line should be switched on.

If all of the values in the right-hand column are negative, then the EUT's emissions are below the limit and the EUT passes those radiated emissions requirements. If some of the values are positive, then the quasipeak measurements are above the limit and the product fails the radiated emissions measurements. To be sure that a signal is not ambient, it should be remeasured and the demodulation function should be used to listen to the signal of interest. The AM and FM demodulators are good tools for determining whether or not a signal is ambient.

The AM/FM demodulation function allows an operator to hear the audio portion of a transmission by dwelling at the marker for a specified length of time (usually 500 ms). If there is any doubt about the signal being ambient or from the EUT, the power to the EUT can be removed and the receiver's display screen observed for changes in the signal. If the signal remains after power has been removed from the EUT, it is ambient. If a signal is determined to be ambient, marker functions can be used to isolate and delete the signal from the receiver's internal memory.

Report Development

After the ambient signals have been deleted from the list, a report should be developed that can be used by design engineers to correct any problems in the EUT. The report should include a list of signals, a graphical representation of the signals and as many as two pages of text that summarize the testing and methods of acquiring the signals.

Commercial software packages are available that automatically control the movements of towers and turntables, capture emissions and locate worst-case emissions. The HP 85876B commercial radiated EMI software performs these functions and also can be used to set up an equipment path and automatically import correction factors. The software can save files in Rich Text format (.RTF extension) for use by most word-processing programs, such as Microsoft Word.® In addition, graphics files can be saved in Windows Metafile format (.WMF extension) for use in several Windows-based applications. The use of such software packages enhances measurement repeatability and improves throughput in radiated emissions tests.