User manual
28
Subject to change without notice
Practical EMI problems Practical EMI problems
The majority of such probes suffer from a disa-
dvantage: their spatial resolution is very limited.
It is hence difculttolocatethesource of the
measured signal. Therefore, when shopping for a
probe, it is advisable to look especially for a probe
withhighresolution of the magnetic eld. This
becomes ever more important as the degree of
integration on EC boards increases so that loca-
lizing individual sources of interference requires
resolution down to millimeters.
Measurements on 4 layer EC boards
using HM554
The following describes how to extract interesting
details from the probe signals. Principally the
signals may be displayed in the time or frequency
domains. The displayvs. time may be more trans-
parent. The following measurements were taken
from a 4 layer EC board of „Europe” format 100
x 160 mm square. The power distribution on this
board is on individual layers. The distance bet-
ween the Vcc- and groundlayers is 100 µm. In the
middle of this board a set of capacitors is located
which connect both layers for AC.
Picture 1: Current signal in the Vcclayer and
close to the Vccpin of a 74AC163.
Picture 1 shows the current signal in the vicinity
of the Vccpin of a 74AC163. The signal amplitude is
proportional to the rate of change of the magnetic
eldandthusofthecurrentatthislocationofthe
layer. The rise and fall times are in the subnano-
second range.
The reason is that the high frequency currents will
owmostlyclosetotheV
cc
pin as they can only be
fed from the charge of the layers there. Such high
frequency components can not be fetched far as
the impedance would become too high. There is
no bypassing capacitor at the V
cc
pin as it would
not be able to deliver high frequency current. Of
course, the two layers V
cc
and ground have a set of
capacitors in the middle of the board. But this set
can only deliver the low frequency components.
Picture 2: Current changes in the power distribu-
tionlayers close to the set of capacitors.
Picture 2 shows the current changes in the vi-
cinity of this set of capacitors. It is obvious that
this signal is much slower than that of picture 1.
Here the rise and fall times are approx. 3ns. The
set of capacitors can only deliver current slowly
to the layers. Such details are only visible with
highresolutionprobessuchastheµHeldprobe.
The following example demonstrates the effect of
absorption measures. In picture 3 the signal was
taken directly at the V
cc
pin of a 74AC00 using the
µHeldprobe.ThisICispoweredbyaV
cc
-ground
system which is un damped. The changes of the
magneticeldarestrong.