Parallel Testing Reduces EMC Prescan

Although a pre-scan reduces overall test time significantly, a radiated EMI test still remains a lengthy process. Why it takes such a long time is because standards dictate that you have to make measurements on a product in different physical orientations and configurations. You have to mount the product on a turntable and rotate it through different angles. You have to raise and lower the antenna to different elevations. You have to change the antenna polarisation from vertical to horizontal. You have to do all these tests in addition to testing over the range of frequencies. As a result, you have four variables to manipulate, three of which involve mechanical movement. In practice, a pre-scan accounts for 80% to 95% of the total time for a radiated EMI test.

With these problems in mind, a newly-opened EMC testing facility in Spain employs a measurement technique that halves pre-scan test time. The technique involves the parallel and simultaneous testing of emissions in the vertical and horizontal planes, and is now in use at the Laborati General D’Assaig Investigacions (LGAI) in Barcelona.

Parallel Test Poses Challenge
To make multiple measurements in parallel in the vertical and horizontal planes you need two antennas connected to two separate receivers. Although conceptually simple, in practice the process poses some challenges.

Firstly, EMC standards dictate that when you measure a product within a test chamber the only metallic parts allowed to be present are those of the product and measuring antenna. To add a second antenna within the chamber means that it must appear transparent to the first. Two antennas with the same polarisation will always interact. Within an EMC test chamber, a distance of up to 10 metres could physically separate the antennas and even with that much separation they can still interfere. Experiments reveal disturbances of 1 or 2 dB in the measurements made by one receiver as a result of a second, similarly polarised, physically separated antenna. Because radiated EMI measurements typically require uncertainties of less than 1 dB, this interference is clearly unacceptable.

The solution is to orient the antennas with different polarisations, one vertical and one horizontal. With this configuration, the antennas become mutually transparent even when both are in the same test chamber.

With this fundamental problem solved, there is now the real potential to drastically reduce the pre-scan test time. The antennas are located within the test chamber up to 10 metres from the product under test, and at an angle to each other as dictated by the test procedure for the product. The standards dictate that you make the tests at different rotations of the turntable. By positioning the antennas at an appropriate angle to each other, typically 368, you can make two sets of measurements at each point of rotation. Because the mechanical movement of the product dominates the pre-scan test time, this simple approach means that you can halve the pre-scan test time.

To implement the test requires each antenna to have a separate test receiver (see Figure 1). The receivers can either sweep the same or different frequency ranges to collect the measurement data. This arrangement produces two separate sets of tabulated results. The challenge now is to reconcile these two sets of data into one table of results so that you can make decisions on which frequencies require further investigation.

Figure 1.  You can conduct EMC pre-scan tests in parallel because horizontal and vertical plane antennas within a ten-metre test chamber do not interact.

Software Manages Dual Channels
Managing this data interpretation is the function of software developed at Hewlett-Packard’s Roseville, Vancouver, and Fort Collins, USA, divisions. The software fulfils a number of roles including testing, scheduling, reporting, and billing.

The software consists of two main modules: a management Test Centre Software Environment (TCSE) module and a Test Executive (TE) module. The TCSE provides a database of resources that you can use for scheduling and reporting statistics. The TE manages the users, test equipment, configuration, and the selection of the specific test procedure. The TE also includes a Radiated Emissions Module (REM), which is important for product qualification testing. Other modules exist for conducted emissions and radiated immunity testing.

The REM performs a variety of radiated emissions measurements as a part of overall product compliance testing. The REM supports the most commonly used tests and conventional measurement techniques. It also specifically supports radiated emissions testing in 10-metre semi-anechoic chambers with the capability of supporting two EMI receivers (such as HP 8542E, HP8546A or HP8571/2A), or two spectrum analysers connected to two separate antennas that are offset by a minimum of 368.

The REM has five test modes — QuickScan, PreScan, Final, Elevation, and Polar — along with a real time display capability:

• QuickScan determines which bands of frequencies you should measure in a subsequent PreScan, or which emissions you should measure in Final mode.
• PreScan determines which frequencies to measure in Final mode.
• Final performs quasi-peak and average measurements of emissions at the frequencies you select during QuickScan or PreScan modes.
• Elevation plots data at a defined turntable angle as the antennas change height.
• Polar plots data at a defined antenna height as the turntable rotates firstly to 3608, and then after the antennas have changed polarity and as the turntable returns to 08.

During a test, each receiver and antenna set divides the frequency range of interest. The turntable rotates from 08 to 3608 with the antennas in one polarity and then from 3608 to 08 with the antennas in the opposite polarity. Each antenna/receiver measures the defined frequency range in a number of bands for a number of sweeps at specified antenna heights.

In Final mode, you can measure the emissions in either an “automatic” mode, an “automatic with pause” mode, or both. “Automatic with pause” allows you to maximise the emissions at a frequency before taking the final measurement. Final mode also sorts frequencies for optimal efficiency.

In a typical test procedure you may firstly use PreScan mode to select the frequencies of interest and follow by Final mode test using quasi-peak measurements with the “automatic with pause” option set. To test the ten highest emissions of a product-under-test from 30 MHz to 2000 MHz at eight turntable angles is possible in 90 minutes. Total test duration includes the time needed for the antennas to step from 1 metre to 4 metres in 200 mm steps in both polarities, as well as encompassing enough time to produce a draft test report. If the product-under-test description, model numbers, serial numbers, cable types, block diagram, and pictures are available, the system can produce a complete report almost immediately.

Looking Ahead
The regulatory environment and the demands of safety and security mean that comprehensive EMC testing is essential. The standards that exist are rigorous and wide-ranging. The challenge is to ensure that implementing comprehensive EMC testing is economically feasible for all manufacturers. One of the critical factors is how to reduce test times without compromising test quality. The parallel test approach is one way in which EMC testing becomes more affordable.

The concept of parallel testing using multiple antennas and receivers has been proven in practice both at Hewlett-Packard and at organisations such as LGAI. However, the concept does not have to stop at dual test sets. EMC engineers are undertaking trials to see if utilising more than two receivers in parallel can extend the concept and reduce test times even further.

Josep Grifol is an EMC consultant with Hewlett-Packard, Spain. He was project co-ordinator for the EMC lab at the Laborati General D’Assaig Investigacions (LGAI).