Chapter 4 - Electromagnetic Compatibility And Medical Devices: Conducted Emissions

CONDUCTED EMISSIONS

Conducted emissions measurements are made to determine the line-to-ground radio noise
from each power-input terminal of a line-powered medical device. Measurements are
taken using a line impedance stabilization network (LISN). A spectrum analyzer and a
quasi-peak adapter with a measurement bandwidth of 9 kHz are typically used to record
the conducted emissions. As shown in Figure 4.15, tests are performed in a shielded
room.

A LISN is a passive RCL network that connects between the ac power line and the
device under test. The purpose of the LISN is to present a standard line impedance to
the device under test regardless of local power line impedance conditions. The LISN also

isolates the device under test from unwanted interference signals on the power line and
provides a test point to probe emissions conducted from the device under test toward the
power line. Figure 4.16 presents the circuit for a 50Ω/50 μH LISN following the definition
of standard CISPR-16-1. This circuit provides a 50-Ω output impedance for measurement
of RF emissions produced by the device under test. This impedance was selected because
theoretical and empirical data have shown that the power circuitry statistically looks like a
50-Ω impedance to standard electronic equipment, and RF test equipment is typically
designed for 50-Ω input. The bandwidth is typically determined by the operating frequency
of the potential victims of the device under test’s conducted emissions. For the
majority of medical devices, emission measurements are carried out from 150 kHz to
30 MHz. This ensures that devices do not interfere with VLF or HF radio communication
systems and other electronic devices operating at these frequencies.

Each of the 50-μH inductors and 1-μF capacitors form an unbalanced filter. The inductors
must be of sufficiently large wire gauge to carry the full ac current demanded by the
device under test with less than a 2-V drop. Conducted emissions are then measured using
a spectrum analyzer with quasi-peak detection. Measurements are taken between hot to
ground and then between neutral to ground. A 50-Ω resistor needs to be connected across
the 1-kΩ resistor, which is not connected to the spectrum analyzer’s 50-Ω input. Switch
SW1 accomplishes phase selection and automatic shunting of the LISN leg not being
observed.

Note that the LISN established by the standards presents a 1-μF capacitance between
the hot line and the LISN and device under the test’s safety ground. A ground fault could
lead to potentially lethal currents to operators in contact with the LISN or the device under
test. For this reason, it is advisable to wire the LISN’s ground terminal permanently to
ground.

A safer way of running design-time tests is to use a LISN made from a modified power
protector designed to filter power line glitches prior to supplying power to computers and
other electronic equipment. The circuit for this LISN is shown in Figure 4.17. The input
connector and power cord, circuit breaker F1, and neon light are found almost universally
in power protector strips. You may also leave any MOVs that you find in the power strip.

Modify the filter circuitry and configure it as shown in the schematic diagram. You may
remove all except one of the power outlets to accommodate the components. The power
outlet used to connect to the device under test (J2) would be one of the power strip’s original
power outlets.

Emissions radiated from this LISN are coupled inductively to a spectrum analyzer. L3
is a single loop of No. 14 stranded insulated wire that exits and reenters the power strip’s
enclosure. The H-field probe made of a VCR head, described above, would then be used
to pick up conducted emissions. Although this LISN does not yield results identical to
those of the standards, it makes it easy to detect emissions conducted by the device under
test into the power line. In addition, although conducted emissions tests should be performed
on both phases (hot and neutral) of the power line, most offensive units reveal
themselves with just the hot-to-ground measurement provided by this LISN.

When the device is prepared for testing, the power cord in excess of the distance is
folded back and forth, forming a bundle 30 to 40 cm long in the approximate center of the
cable. Power supply cords for any peripheral equipment should be powered from an auxiliary
LISN. Excess interface cable lengths should be bundled separately in a noninductive
arrangement at the approximate center of the cable with the bundle 30 to 40 cm in
length. The emissions conducted are maximized by varying the operating states and
configuration of the device under test. The limits for conducted emissions per EN-55011
for group 1 devices are shown in Table 4.4. As an example, Table 4.5 shows the results
we obtained recently when testing an implantable-device programmer for conducted
emissions.

CONDUCTED EMISSIONS

Conducted emissions measurements are made to determine the line-to-ground radio noise
from each power-input terminal of a line-powered medical device. Measurements are
taken using a line impedance stabilization network (LISN). A spectrum analyzer and a
quasi-peak adapter with a measurement bandwidth of 9 kHz are typically used to record
the conducted emissions. As shown in Figure 4.15, tests are performed in a shielded
room.

A LISN is a passive RCL network that connects between the ac power line and the
device under test. The purpose of the LISN is to present a standard line impedance to
the device under test regardless of local power line impedance conditions. The LISN also

isolates the device under test from unwanted interference signals on the power line and
provides a test point to probe emissions conducted from the device under test toward the
power line. Figure 4.16 presents the circuit for a 50Ω/50 μH LISN following the definition
of standard CISPR-16-1. This circuit provides a 50-Ω output impedance for measurement
of RF emissions produced by the device under test. This impedance was selected because
theoretical and empirical data have shown...