the path the current takes,
and this sets up a slight charge
imbalance across the width
of the conductor. The Hall ef-
fect occurs in all conductors,
but the voltage produced is
downright minuscule, even in
special semiconductors made
for the job: we’re talking a few
microvolts to millivolts of volt-
age difference per milliTesla
of magnetic field intensity, so
follow-on amplification is de
rigueur.
The output of the Hall sen-
sor can be amplified directly
for use by external circuits
- which is called open-loop
operation - and while the ac-
curacy and linearity of this ap-
proach is poor, it is capable of
a higher bandwidth for a given
peak current. Alternatively,
the Hall sensor can be used to
drive a counteracting current
through a coil wound on the
same core to cancel out the
field from the monitored con-
ductor, and the output made
proportional to the control
current, which is called closed-
loop operation. The advantage
of the latter approach is that
any non-linearity caused by
the core material is nulled
out. Also, because the net field
in the core is close to zero, a
closed-loop sensor can gener-
ally handle a higher current
for a given core area (at least
until the nulling circuit runs
out of headroom). Still, ac-
curacy and linearity aren’t on
par with a good shunt-based
design: one popular manufac-
turer quotes an accuracy of
1%, a temperature coefficient
of +/- 0.1% / K, and initial
and hysteresis offsets (which