Benchtop EMC Testing Techniques for Medical Equipment

Medical Device & Diagnostic Industry Magazine
MDDI Article Index

An MD&DI  January 1998 Column

EMC TESTING

Using loop probes to help test devices for electromagnetic compatibility saves costly redesign and complements open area and chamber tests.

A design may seem perfect; however, when electromagnetic compatibility (EMC) performance hasn't been considered, too often the product will fail at the point in the development process, the final testing phase, where redesign is most expensive and difficult. Designing medical electronic products to be electromagnetically compatible means establishing confidence that the design will comply with regulatory agency requirements and be compatible with associated equipment.

Loop probes are easily constructed and facilitate benchtop testing for electromagnetic compatibility during product design, testing, and production phases.

A cost-effective approach to product development is to gradually increase confidence in a design's EMC performance during the design, testing, and production phases, rather than to defer testing to the very end. Potential problems and contingency solutions should be identified as early as possible so that problems encountered later in development can be handled with relative ease. While some product designs might seem straightforward—apparently negating the need for thorough process testing—the cost and effort required to redesign a product late in development and the potential loss of revenue caused by delayed market entry, are greater than the costs of early EMC testing.

For example, to minimize cost a manufacturer might design a printed circuit board (PCB) with only two layers. Attaining good grounding on a two-layer board is far more difficult than doing so on a multilayer board with a dedicated ground plane. If the two-layer board passes the final EMC test, everything is fine. However, if it fails, redesign may delay production and shipments. A more prudent approach would be to design a two-layer board with the best possible ground system while simultaneously pursuing a multilayer design. Because it has a better ground system, the apparently more expensive multilayer approach may work with fewer bypass capacitors or require less-stringent shielding. When alternatives are developed in parallel, the choice of which design to use can be made after final EMC testing with little or no effect on production scheduling.

Final EMC testing of products with embedded microprocessors is usually done at an open area test site (OATS) or in a specially designed absorber-lined chamber (ALC). These tests are the most accurate available but also the most time-consuming. In comparison, benchtop EMC measurements are faster but focus on individual sources of interference instead of system emissions. Benchtop tests can be powerful complements to OATS or ALC measurements. For example, a preliminary OATS test can be carried out on a prototype, and then benchtop tests can be used to identify the specific source of the emissions. Once the emissions source is identified, specific design changes can be evaluated on the design bench and clarified with further OATS or ALC measurements.

Identifying radiated emissions sources close to their origins is usually done with a sensor or transducer that converts fields into voltages for measurement by a receiver or spectrum analyzer. This sort of transducer, or antenna, comes in two basic types: dipolelike structures that sense electric fields and looplike structures that sense magnetic fields. Derivations of the classic loop antenna are often used for benchtop EMC work because it is usually easier to identify and characterize radiated emissions sources by examining magnetic fields produced close to their origin than it is by studying electric fields.

MAGNETIC FIELD LOOP PROBES

Any electrical circuit can produce magnetic fields and radiate radio-frequency (RF) energy.In digital circuits, RF currents usually come from the high-order harmonics of the digital signals. Magnetic fields will be strongest wherever RF currents are forced to flow in other than straight lines. For example, a cutout or opening in an otherwise continuous ground-plane layer of a PCB will cause RF currents in that layer to flow in curved paths around the cutout, just as water flows around a rock in a stream. These currents produce magnetic fields in the cutout.¹ A similar situation can occur near apertures or seams in otherwise continuous shields.²

Essentially identical to the loops used as radio direction-finding antennas, but much smaller, electrically small shielded loop antennas are effective near-field probes for characterizing magnetic field sources on PCBs or other electronic structures.³

Theoretical Basis. An electrically small loop antenna will produce an output voltage proportional to a perpendicular incident magnetic field, as shown in Figure 1. This voltage can be calculated from Faraday's law as

where n = number of turns in loop (typically one for small probes), = angular frequency = 2 ¼ x frequency, B = incident magnetic field, A = area enclosed by loop, and = angle between perpendicular to loop plane and the B-field vector.

Figure 1. A loop probe's output voltage is proportional to the perpendicular component of the impinging RF magnetic field. Equivalent circuit impedance is very small.

The "electrically small" assumption means the loop probe is small compared to the wavelength at the frequency of interest, so the phase shift of the current flowing around the loop is negligible. For benchtop EMC testing, there is no need to match the probe's impedance to the receiver or spectrum analyzer to which it's connected.

The probe's equivalent circuit is composed of the radiation resistance (R_r), the simple bulk resistance of the loop (R_L), and the inductance of the loop (L). All three of these quantities are usually very small, and the sum of their impedances is usually much smaller than the typical 50-‡ input impedance of most receivers or spectrum analyzers. Consequently, in most cases the impedance mismatch can be ignored, and the probe acts as an ideal voltage source.

Construction Features. Magnetic field loop probes can be constructed or purchased. To be effective, the loop probe's response to incident electric fields must be minimized. Otherwise, the probe's response to magnetic fields will be difficult to differentiate from its response to electric fields. The classic way to reduce a loop antenna's electric field response is to add an electric field shield with a small break at one point to keep the circulating currents from shielding the loop from magnetic fields, too.

Another way to minimize electric field responses in loop antennas is to use their inherent balance. Electric fields produce a common mode current at the probe's output terminals, whereas magnetic fields produce a differential response. A balanced-to-unbalanced transformer, or balun, attenuates the electric field response but has less effect on the differential magnetic field response. Figure 2 shows how a common mode transformer or choke can be used as a simple balun.

Figure 2. Small magnetic field probes should have E-field shields and a balun or other means to inhibit common-mode E-field responses.A loop probe tests for magnetic fields, which are most likely to be found near a product's seam fasteners or corners.

Finally, antennas must be insulated when used as field probes. It's easy to inadvertently touch the probe to electronic components and live circuitry, creating a short to ground if the probe isn't insulated. If high voltages are exposed or high currents available, such as in a power supply, this can be very hazardous to the circuitry and the user. Loop probes should be covered with a durable insulating layer of plastic or rubber.

PROBING FOR MAGNETIC FIELD SOURCES

Loop probes help designers visualize the structure of the magnetic fields produced by electronic circuitry. The probe responds to the portion of the field that is perpendicular to the plane of the loop. For example, a circuit loop on a PCB acts as a small electromagnet, producing a field that is vertical directly over the board but then curves over to return to the other side of the board. Consequently, when probing PCBs, the loop should first be held with its plane parallel to the board (perpendicular to vertically oriented fields). In the vicinity of a hot spot, the probe should be turned 90° to see how the field behaves.

Loop probe tests for magnetic fields, which are most likely to be found near a product's seam fasteners or corners.

When probing shields for leakage, the strongest magnetic fields will be found where shield currents are constricted and forced to flow in a curved path—most likely near seam fasteners or corners, especially at seam gaps. Again, the loop probe should be held parallel to the shield's surface to search for hot spots.

INDUCING MAGNETIC FIELD SOURCES

Loop probes can also use a signal generator's output to produce magnetic fields for EMC troubleshooting. However, loop probes are almost the electrical equivalent of a dead short, so the absolute minimum signal power needed for the job should be used. To protect the signal generator, it's a good idea to add a 50-(omega) resistor in series with the probe.

When using the loop probe as a field source, be aware of the structure of the magnetic field around the probe. The probe should be rotated to control the field orientation and induced currents when it is near sensitive circuitry.

STRUCTURAL RESONANCES

In some cases, passive structures within electronic equipment have naturally resonant frequencies that coincide with internal RF sources, such as clock harmonics. When excited, resonant structures can act as antennas, exacerbating radiated emissions. A spectrum analyzer, tracking generator, directional coupler, and simple loop probe can be configured as an absorption wave meter to locate such resonant structures.⁴ Once they are located, the product's design can be altered to dampen the resonance (e.g., by adding a ferrite to a cable) or to move it to a frequency where an internal excitation source doesn't exist.

Structural resonances can be found by using a spectrum analyzer with an internal tracking generator (Figure 3). The tracking generator's output frequency follows the spectrum analyzer's swept input frequency. The generator's output is connected to the loop probe through a directional coupler. The loop probe reflects most of the incident energy it receives from the tracking generator through the directional coupler, which then routes this reflection to the spectrum analyzer.

Figure 3. A system for finding structural resonances can be assembled from a spectrum analyzer with a tracking generator, directional coupler, and loop probe.

In operation, the spectrum analyzer sweeps some band of interest, such as a portion of the 30—1000-MHz range used for commercial EMC testing. The reference level is adjusted so the loop reflection appears near the top of the screen. When the loop is brought near a structure with a natural resonant frequency within the swept-frequency range, some of the incident RF is absorbed and lost. This absorption appears as a dip in the spectrum analyzer's trace, and the structure's resonant frequency can be read from the screen.

This effect is subtle and can easily be missed. The spectrum analyzer's vertical sensitivity should be set to 1 dB per division and the probe moved very slowly. Frequency spans should be no larger than necessary, e.g., 30—130 MHz.

Many structures with measurable electromagnetic resonances may be discovered this way. The probability of finding such resonant structures is likely to increase as clock speeds increase, because progressively higher RF frequencies will more often excite smaller structures.

CONCLUSION

Anticipating problems and preparing contingency solutions helps keep new product development on schedule. Early EMC information is useful even when somewhat qualitative. Benchtop EMC techniques using small magnetic field loop probes sacrifice some accuracy for speed but help identify potential EMC problems and cost-effective solutions.

REFERENCES

1. Kimmel WD, and Gerke DD, Electromagnetic Compatibility in Medical Equipment, IEEE and Interpharm Press, Buffalo Grove, IL, 1995.

2. Ott HW, Noise Reduction Techniques in Electronic Systems, 2nd ed, New York, John Wiley, 1988.

3. Roleson S, "Evaluate EMI Reduction Schemes with Shielded-Loop Antennas," EDN, 29(10):203—207, 1984.

4. Roleson S, "Finding EMI Resonances in Structures," EMC Test Design, 3(1):25—28, 1992.

Scott Roleson is the lead engineer for Hewlett-Packard's Telecom Test Center (San Diego). Photos courtesy of Scott Roleson

Copyright ©1998 Medical Device & Diagnostic Industry