Characterizing Electromagnetic Fields of Common Electronic Article Surveillance Systems

During the past decade, the U.S. Food and Drug Administration (FDA) has received more than 28 medical device reporting incidents of adverse interactions between medical devices and electronic article surveillance (EAS) systems, metal detectors, and security systems.¹ Several case reports and four peer-reviewed studies document adverse interactions between EAS systems and implanted pacemakers, implanted automatic cardiac defibrillators, implanted neurostimulators, and other ambulatory medical devices.^2,3,4–7 Anecdotal reports and many newspaper articles suggest that many more device interactions have occurred and gone unreported.

Each year millions of people enter establishments protected by EAS systems. Because more people are using electronic implants and ambulatory medical devices, adverse interactions with EAS systems are of increasing concern. FDA conducted a study to provide data to characterize electromagnetic fields generated by EAS systems. The data presented in the study are being used for susceptibility testing of various implanted cardiac devices and other ambulatory medical devices to magnetic fields emitted from EAS systems.

FDA's Center for Devices and Radiological Health (CDRH) identified EAS systems in common use and collected samples of the most popular technologies used in the United States. These included extremely low frequency and voice-frequency (both continuous-wave magnetic), low-frequency pulsed magnetic, and medium-frequency and high-frequency swept radio-frequency systems. An EAS system installer (Sentec EAS Corp.; Deerfield Beach, FL) and an EAS system manufacturer (Checkpoint Systems Inc.; Thorofare, NJ) loaned sample systems to CDRH for the study. The seven sample systems, including one duplicate, were from three different manufacturers: Sensormatic Electronics Corp. (Boca Raton, FL); Knogo North America (Hauppauge, NY); and Checkpoint Systems Inc.

Each EAS system was mounted on a simple wooden platform. The purpose of the platform was to fix the separation distances between the transmitter and receiver pylons to separation distances found in the typical installation for each type of system. A magnetic loop antenna connected to a Hewlett-Packard (Palo Alto, CA) Model 8560E spectrum analyzer and a Tektronix Inc. (Portland, OR) Model TDS 380 oscilloscope measured the frequency of operation, modulation type, and duty cycle of each EAS system.

The spatial magnetic flux density distributions were mapped using a scanning system capable of positioning a probe anywhere in a volume measuring 2 m wide x 2 m deep x 1 m high. The software to control the scanning system was developed by Sonix Inc. (Springfield, VA). Data were recorded via a Datel Inc. (Mansfield, MA) PC414A high-speed analog input board with 12/14-bit analog-to-digital resolution. The three-axis scanning system was bolted to the ceiling of a shielded room. The z-axis structure (representing up and down scanning) was constructed of nonconducting materials that are minimally perturbing to electromagnetic fields.

Different electromagnetic field measurement systems were used for different frequency ranges (see Table I). All except one were three-axis (isotropic) probes that measured the total magnetic field at a given point. A single-axis Deno electric field measurements (EFM) (West Stockbridge, MA) Model 116-3-60-0367 magnetic field probe was used for the extremely low frequency field mapping (219–535 Hz). The extremely low frequency magnetic fields were mapped in three separate scans for each plane. In each scan, the Deno magnetic field probe was oriented along a different orthogonal axis. At each point, the data were then combined using the square root of the sum of the squares in each of the three orthogonal field components. Extremely low frequency and very low frequency magnetic fields were also measured using a Holaday Industuries Inc. (Eden Prairie, MN) Model HI-3627 extremely low frequency magnetic field meter (5 Hz–2 kHz).

The voice-frequency magnetic fields were mapped using a Wandell and Goltermann (Research Triangle Park, NC) Model EFA-2 (5 Hz–30 kHz) field analyzer that can make isotropic measurements. Low-frequency pulsed magnetic fields were mapped using a Holaday Industries Inc. Model HI-3637 very-low frequency (2–400 kHz) isotropic magnetic field meter. The medium- and high-frequency swept radio-frequency magnetic fields were mapped using an HI-4433-LFH broadband (0.3–10 MHz) magnetic field probe. The electric fields were mapped using an HI-4433-HSE (0.5 MHz–1.5 GHz) isotropic electric field probe. Prior to making these magnetic field measurements, the probes were calibrated using an EMCO (Austin, TX) Helmholtz coil, driven by a Hewlett-Packard Model 33120A function/arbitrary waveform generator connected to an ADCOM (East Brunswick, NJ) GFA-555II high-current power amplifier. For the low-frequency range (30–300 kHz) and higher-frequency ranges, the arbitrary waveform generator was connected directly to the Helmholtz coil. Current to the Helmholtz coil was monitored across a precision 1% 20-W 0.866- resistor with a Fluke (Everett, WA) Model 87 True rms multimeter, a Keithley Model 197 autoranging microvolt digital multimeter, or a Tektronix Model TDS 380 digital real-time oscilloscope, depending upon the frequencies being calibrated.

The low-frequency to high-frequency magnetic field probe accuracy was verified by using spot measurements with a single-loop antenna and the HP-8560E spectrum analyzer. A single turn loop of area 21.5 cm² was made from a section of 0.325-in. (8.26 mm) diam 50- semirigid cable measuring 86 cm from the center of the loop to the end of the connector. The outer copper jacket of the loop is cut circumferentially around the portion of the cable jacket centered on the section of the loop farthest from the main shaft leading to the N-type connector to cancel electric field–induced currents. This loop was placed in the magnetic field at the same location as the low-frequency and high-frequency magnetic field probes; it was rotated to achieve the maximum reading on the spectrum analyzer. The calculated field strength from the single turn loop was then compared with the measured field strength from the HI-4433-LFH broadband isotropic magnetic field probe. When the measurements from the two instruments were within ±1.5 dB, they were considered to agree.

Figure 1. The horizontal plane, normal to the face of the EAS system transmitter pylon. H = the height from the base of the EAS system pylon, 130 cm and x cm, where x = the height of the maximum field strength. S is 6 cm from the transmitter pylon face.

Figure 2. The magnetic EAS system #1 with both system pylons transmitting magnetic fields. The distances noted are with reference to the inside face of one pylon (left side) as indicated by the distance from pylon axis.

Figure 3. Magnetic field strengths for all EAS systems in the horizontal plane 130 cm from the floor.

A scanning protocol specified that electromagnetic measurements would be made in two horizontal planes and three vertical planes around the transmitter pylon resulting in five data sets. Measurements were made at two heights in the horizontal plane (see Figure 1) for each EAS system. An electromagnetic field map was made at a height of 130 cm and at the height of the maximum flux density for each EAS system (as determined by a scan of the vertical normal plane). Vertical measurements were made in three different planes. Two vertical-plane measures were performed, each parallel with the face of the EAS system transmitter pylon. The planes were located 6 and 36 cm from the transmitter pylon face. The vertical planes were 100 cm high starting 55 cm from the base of the transmitter pylon. The third vertical plane was normal to and centered on the face of the transmitter pylon (see Figure 3). The nearest edge of this plane was 6 cm away from the pylon of the EAS system.

Figure 2 shows an example plot for the horizontal plane 130 cm from the floor. Table II provides a summary for the eight systems of the maximum magnetic flux density measured at a single point 36 cm from the EAS system transmitter pylon face at a height of 130 cm from the ground, along the centerline of the pylon. The distance from the floor roughly approximates that of an implanted pacemaker in a standing adult. The horizontal distance from the transmitter pylon was chosen to minimize mutual inductance between the magnetic field probe and the transmitter pylon. Also, at distances greater than three probe diameters, the error between the idealized point measurement reported and the isotropic volume measurement of an ideal magnetic field probe is less than 0.04 dB (1%). This determination was made by a simple spreadsheet model of a field gradient comparing point measurements to an idealized measurement probe surface.

Analysis and Discussion

Data are presented as peak magnetic flux density. This format was chosen so that the magnitude of the magnetic flux density emitted from these EAS systems could be compared. Time-varying magnetic fields (dB/dt) induce voltages in tissues and medical device leads that can be readily calculated from peak magnetic flux density, frequency, and modulation. Using these calculations, medical device designers can determine if interactions are likely when the device is exposed to given magnetic flux densities and waveforms.

Converting continuous-wave modulation measurements from instruments reporting true root-mean-squared (rms) values is straightforward. Data for the pulsed EAS systems required special analysis because of the nature of the instruments used. The magnetic flux densities displayed by the Holaday Industries isotropic field instruments for extremely low frequency, voice-frequency, very-low frequency, and low-frequency magnetic fields are generated by true rms converter-integrated circuits. The true rms values from the circuits are combined from each of the three orthogonal magnetic field sensors through vector addition to produce the resultant magnetic field magnitude displayed by the instrument. In our study, we converted the rms values (B^_rms) for magnetic flux densities to peak measurements for the gated low-frequency sinusoid as shown in the following list using the definition of rms as given in equation (1).

Substituting the equation of a sinusoid for f(t) yields equation (2), where = 2f, f = the frequency of the sinusoid, T = the period of the gated sinusoid, and B₀ = the peak magnetic flux density.

The integral is evaluated from 0 t T/10, where T/10 is the duty cycle of the gated sinusoid from the EAS units we calculated. Since the amplitude of the gated sinusoid is zero from T/10 t T, the second term in equation (2) is equal to zero.
The simplified solution for the integral is shown below in
equation (3).

For >>1/T, equation (3) can be approximated as

Peak Magnetic Flux Density. Figure 3 is a summary of the peak magnetic flux densities measured in the horizontal plane 130 cm from the floor (as shown in Figure 1). The closest measurements shown in Figure 3 are 15 cm from the transmitter face. Data measured closer than this distance are likely to contain large errors due to mutual inductance between the coils of the EAS system and the magnetic field instrument. The EAS system technologies measured indicate that the magnetic flux densities are highest for the extremely low frequency, voice-frequency, and low-frequency systems. The radio-frequency systems (medium frequency and high frequency) had weaker magnetic flux densities. The extremely low frequency system that we measured had two transmitting pylons. In this system, the exposure to the magnetic fields remains relatively higher anywhere between the pylons than was measured between the single-transmitter pylon systems. Table II shows a numerical summary of the flux densities measured 36 cm from each of the transmitter pylons at a single point on the centerline of the EAS systems.

System Installation Specifications. Each EAS system manufacturer specifies the nominal environmental conditions in which their units should be installed. Because most of the systems studied were installed by a third party and were acquired as used systems, some of the units may not have flux densities representative of a system installed by the manufacturer's representatives. However, since these EAS systems are available on the market, the magnetic flux densities measured would represent exposures that may be encountered by the public. Subsequent transmitter coil current measurements indicated that two of the low-frequency systems (PM #1 and PM #2) produced magnetic fields more than 10% stronger than the manufacturer reported as nominal at the specified supply voltage. These units also have unregulated power supplies. The building power supplied to these units was not regulated during testing. The typical supply voltage to the systems during our testing was 126 V ac, which was 16 V ac (15%) more than the manufacturer's designed nominal ac voltage of 110 V ac. These two factors combine to produce reported magnetic flux densities that may be stronger than nominal for many units installed by the manufacturer and operated at the nominal power line voltages (110 V ac).

FDA's Winchester Engineering and Analytical Center (WEAC) measured a number of EAS systems in use in retail stores and libraries in and around the greater Boston area.⁸ Results of this study were compared with measurements reported in this paper for similar systems based on the EAS system emission frequency (Table III). WEAC measured two extremely low frequency systems in music stores. The separation distances between EAS system pylons for these units were greater than the pylon separation distances that were recommended for the extremely low frequency EAS system installed in the CDRH laboratory. The flux densities measured were generally higher for the field systems than for the unit measured in the laboratory by 3 and 1.6 dB respectively at a horizontal distance of 40 cm from the transmitter pylon and 130 cm above the floor. When the maximum values of the measurements made in the field and in the laboratory are normalized to one and compared, the data show differences of –1.2 and –0.4 dB respectively at 40 cm from the transmitter pylon and 130 cm from the floor.

Comparison of flux densities for voice-frequency EAS systems at 40 cm from the transmitter pylon and 130 cm from the floor showed measurement differences of 0.2 dB. Flux densities for low-frequency EAS systems at 40 cm from the transmitter pylon and 130 cm from the floor showed measurement differences of 3 dB; normalized data show a difference of 2.1 dB. The radio-frequency EAS systems have relatively weak magnetic fields, and, therefore, the instruments used by both groups were not able to measure the magnetic fields for the laboratory EAS systems at distances of 40 cm or greater. The radio-frequency systems measured by WEAC were library systems.

Differences between electromagnetic flux densities measured in the laboratory by CDRH and field measurements by WEAC would be expected. The make and model numbers of some of the systems measured by WEAC were not the same as the models measured in the laboratory. Differences in commercial ac power levels supplied to the systems and the presence of conductive and magnetic materials in proximity to the systems or the fields being measured could change the electromagnetic field levels and patterns generated by the systems (Table III). Measurements made where the flux densities are weak and where flux densities are close to the minimum sensitivity of the measurement instrument are subject to greater measurement error. Under these conditions, small perturbations will indicate larger differences in electromagnetic field measurements. Probe positional differences can also significantly affect the repeatability of measurements, particularly for those made close to the source. Here the flux densities are changing rapidly with distance. While efforts can be made to control these variables in the laboratory, they are very difficult to control during on-site measurements.

FDA studied eight EAS systems representing seven different models from three manufacturers. The operating frequencies of these units varied from about 200 Hz to about 10 MHz. Spatial maps of electromagnetic fields indicate that lower frequency systems generally have stronger magnetic fields than the high frequency systems. Magnetic fields also fall off rapidly as distance increases from the transmitting coil. Measurements indicate that systems installed in retail stores and libraries may have flux densities that vary within ±3 dB of the laboratory measurements. Magnetic fields emitted from individual EAS units of the same model may vary.

There have been a number of reports of implanted and other ambulatory electronic medical devices interacting with the electromagnetic fields emitted from EAS systems. The information provided in this paper may be useful in studying such EMI and in helping engineers employ design techniques to minimize the vulnerability of implanted equipment.

The author would like to thank Leah M. Shrupp, a senior in biomechanical engineering minoring in biology and chemistry, and Michael Cobb, a senior in biomedical/electrical engineering, both at Marquette University (Milwaukee, WI), for contributing to this project. They participated via an internship arranged through a cooperative program between Marquette University, the Les Aspen Center (Washington, DC), and FDA.

Jon P. Casamento is working at FDA as a senior regulatory research officer and electronics engineer with the Center for Devices and Radiological Health. He received his BS in electrical engineering from the University of Maryland, College Park, in 1979. He can reached by phone at 301/827-4959 or by e-mail at JPC@cdrh.fda.gov.