Discussion: Figure 1
shows a
shielded magnetic loop probe
constructed using semi-rigid coax cable (note the small gap in the
shield at the midpoint of the right side of the loop). Figure 2 shows
an example of an unshielded magnetic loop probe constructed from stiff
wire and heat shrink tubing on the end of a BNC barrel adapter. Either
of these loops can be used to measure possible effects of a local
magnetic field that can couple onto cables and thus causes far field
emissions from the cable. I prefer square loops to round ones as most
cables are relatively straight, not curved in an arc that matches the
curvature of a specific round loop probe.
Figure 2. A Simple Wire Loop
The situation of
concers is illustrated in Figure 3. In this case, a physically small
magnetic field couples from an IC to a system cable. The magnetic field
emitted from the IC induces a series voltage in the cable which then
causes a current to flow resulting in emissions. If we assume the
driving point impedance of the cable at the IC is 100 Ohms (the driving
point impedance of a dipole at resonance is about 70 Ohms, but I will
use 100 Ohms as a round number) then only about 1 millivolt of induced
signal will generate current on the cable of 10 microamps, enough
to be a potential class B (consumer equipment) emissions problem. See my
March 2006 Technical Tidbit, Predicting Cable Emissions from Common Mode Current for more information on common mode currents and emissions.
Figure 3. System Cable Near an IC
A situation like this can be predicted by using a magnetic loop probe.
First construct a square probe with the length of a side approximately
equal to the distance across the IC. A plain wire loop like that shown
in Figure 2 will do. If the probe is
positioned over the IC, perpendicular to the plane of the IC package
and with the side of the loop opposite its cable against the IC
package, the resulting loop output
signal can be used to predict emissions if an appropriate length cable
scan
intercept the field of the IC. The loop should be rotated and moved
over the package to pick up the maximum signal possible at each
frequency of interest. If a magnetic field loop probe is
positioned over the loop and picks up two millivolts into the 50 Ohm
input of a spectrum analyzer, then a cable might pick up a similar
voltage if positioned over the IC. Two millivolts driving 100 Ohms will
result in 20 micropamps of current, enough to be even a Class A
(industrial equipment) emission problem.
Similarly, a wire loop can measure the possible induction into cables
from other sources, such as a switching power supply. I generally find
a one inch square loop to be a good size. The side of the loop opposite
its feed cable is placed where a system cable could be located to
estimate the signal induced into such a cable. Larger loops could be
used, but when the circumference of the loop is 1/2 wavelength at the
frequency of interest, the loop goes into resonance itself and becomes
less useful.
The surface of a circuit board can be scanned using a square magnetic loop
probe and areas of the board than
can induce a millivolt at any frequency into the loop can be dangerous
if a cable can drop onto the board at that point and then exit the
enclosure.
Cable radiation due to interception of local fields is easy to
demonstrate. Just put a current probe on a cable and place the cable
near the current probe on top of an active IC of good size and watch
the common mode current on the cable increase as the cable is brought
near the IC.