Figure 1 shows the
test setup with an off-the-shelf switching power supply. The work lamp contains a 12 Volt, 50 Watt lamp that acts as
a load on the power supply and the square wire loop is used to measure
the magnetic fields. The loop is the same one that appears in many of
my Technical Tidbits and is shown in Figure 2. It is made from AWG 16
gauge stiff brass wire that is first soldered to a BNC barrel adapter on one
end, then covered with heat shrink tubing, and finally the other end is pushed into the center pin receptacle of the
BNC connector. Sometimes I use this loop to inject
magnetic fields into equipment using high voltage pulses (electrical
fast transients, EFT) making the solder connection of the wire to the
BNC adapter critical for safety, but here the solder just makes the
connection of the wire to the outside of the BNC adapter more
Figure 2. Wire Loop Used for Magnetic Field Measurements
Many switching power supplies do not
have a specification on allowed magnetic field emissions and this can
cause a problem. Since the radiated magnetic fields are often due to
parasitics, they can increase significantly due to a seemingly
innocuous component or layout change. The magnetic field
specifications I have seen usually specify the magnetic field itself over a
specified bandwidth. Such a specification, although technically
correct, does not lead to a quick understanding of the implications of
the field. Ultimately, the magnetic field maybe converted to a voltage
in a circuit resulting in a problem. To calculate that voltage can be a
complicated, time consuming process.
An easier way to specify the magnetic field is to define the geometry
of a loop (such as the one in Figure 2), a termination for the loop (50
Ohms in this case), and the allowed induced voltage in the loop over a bandwidth.
This specification has immediate intuitive value. If a two cm loop
pickups up hundreds of millivolts, so will other conductors in the
area. To the extend other conductors define larger loops, maybe more
voltage would be induced into them. A loop output into 50 Ohms of
hundreds of millivolts peak or more indicates a problematic level of magnetic
emissions that can affect nearby circuits, especially circuits on two
layer PCBs which are common in consumer electronics.
Figure 3 shows the loop of Figure 2
positioned over the power transformer of the switching power supply in
Figure 1, a likely source of magnetic field emissions. Physical layout
of the PCB in a power supply and component connections can also generate magnetic field
emissions as well.
Figure 3. Magnetic Loop Positioned Over Power Transformer of Switching Power Supply
Figure 4 shows the resulting loop output into the 50 Ohm scope input
impedance (50 Ohm coax was used). The resulting peak of 400 mV
is large enough to affect nearby circuits and should be reduced by
either changing the design of the power transformer or using a metal
cover over the power supply to short out the magnetic fields emanating
from the supply. Any highly conductive metal that will allow eddy
currents to cancel the field from the power supply will do.
Although there may be some electric field pickup by the unshielded
loop, I estimate it to be around 50 mV or less, it is not enough to
change the conclusion that this power supply can be a problem if
circuits are mounted directly over the top of the supply or sensitive
analog circuits are within 10 cm or maybe even further away. If one were to
flip the loop over and compare waveforms, the difference in the two
plots would be mostly due to electric field coupling.
Scope Plot of Loop Output into 50 Ohms
(Vertical scale = 100 mV/div, Horizontal scale = 50 ns/div)
I usually like to ensure that the output from a 2.5 cm, one inch, loop
loaded by 50 Ohms
and passed over the surface of a switching power supply is less than
50 mV peak. At that level, problems should be rare. However, if you
sensitive circuits, even a field resulting in 50 mV in the loop can be
a problem. I have seen a small computer tape backup
drive that was affected by a switching supply module about 15 cm away.
The switching module induced 4 Volts peak into a 2 cm loop! The tape
head in such drives is a sensitive receiver for
fields originating in switching supplies.