12
Shunt Resistor
The current-sensing shunt resistor should have low re-
sistance (to minimize power dissipation), low inductance
(to minimize dI/dt induced voltage spikes, which could
adversely aect operation), and reasonable tolerance (to
maintain overall circuit accuracy). Choosing a particu-
lar value for the shunt is usually a compromise between
minimizing power dissipation and maximizing accuracy.
Smaller shunt resistances decrease power dissipa-
tion, while larger shunt resistances can improve circuit
accuracy by utilizing the full input range of the isolated
Sigma-Delta modulator.
The rst step in selecting a shunt is determining how
much current the shunt will be sensing. The RMS current
in each phase of a three-phase motor is a function of
average motor output power and motor drive supply
voltage. The maximum value of the shunt is determined
by the current being measured and the maximum rec-
ommended input voltage of the isolated modulator. The
maximum shunt resistance can be calculated by taking the
maximum recommended input voltage and dividing by
the peak current that the shunt should see during normal
operation. For example, if a sinusoids phase current motor
has a maximum RMS current of 10 A and can experience
up to 50% overloads during normal operation, then the
peak current is 21.1 A (= 10 × 1.414 × 1.5). Assuming a
maximum input voltage of 200 mV, the maximum value
of shunt resistance in this case would be about 10 mW(≈
200 mV/21.1 A).
The maximum average power dissipation in the shunt
can also be easily calculated by multiplying the shunt re-
sistance times the square of the maximum RMS current,
which is about 1 W in the previous example.
If the power dissipation in the shunt is too high, the resis-
tance of the shunt can be decreased below the maximum
value to decrease power dissipation. The minimum value
of the shunt is limited by precision and accuracy require-
ments of the design. As the shunt value is reduced, the
output voltage across the shunt is also reduced, which
means that the oset and noise, which are xed, become
a larger percentage of the signal amplitude. The selected
value of the shunt will fall somewhere between the
minimum and maximum values, depending on the par-
ticular requirements of a specic design.
When sensing currents are large enough to cause sig-
nicant heating of the shunt, the temperature coecient
(tempco) of the shunt can introduce nonlinearity due to
the signal-dependent temperature rise of the shunt. The
eect increases as the shunt-to-ambient thermal resis-
tance increases. This eect can be minimized either by
reducing the thermal resistance of the shunt, or by using
a shunt with a lower tempco. Lowering the thermal resis-
tance can be accomplished by repositioning the shunt
on the PC board, by using larger PC board traces to carry
away more heat, or by using a heat sink.
For a two-terminal shunt, as the value of shunt resistance
decreases, the resistance of the leads becomes a signi-
cant percentage of the total shunt resistance. This has
two primary eects on shunt accuracy. First, the eective
resistance of the shunt depends on factors such as how
long the leads are, how they are bent, how far they are
inserted into the board, and how far the solder wicks up
the lead during assembly (these issues will be discussed
in more detail shortly). Second, the leads are typically
made from a material such as copper, which has a much
higher tempco than the material from which the resistive
element itself is made, resulting in a higher tempco for the
shunt overall. Both of these eects are eliminated when a
four-terminal shunt is used. A four-terminal shunt has two
additional terminals that are Kelvin-connected directly
across the resistive element itself; these two terminals are
used to monitor the voltage across the resistive element
while the other two terminals are used to carry the load
current. Because of the Kelvin connection, any voltage
drops across the leads carrying the load current should
have no impact on the measured voltage.
When laying out a PC board for the shunts, keep these
points in mind. Make sure the Kelvin connections to
the shunt are brought together under the body of the
shunt and then run very close to each other to the input
pins 2 and 3 of the isolated Sigma-Delta modulator; this
minimizes the loop area of the connection and reduces
the possibility of stray magnetic elds from interfering
with the measured signal. If the shunt is not located on
the same PC board as the isolated Sigma-Delta modulator
circuit, then a tightly twisted pair of wires can accomplish
the same thing.
Also, multiple layers of the PC board can be used to
increase the current-carrying capacity. To help distribute
the current between the layers of the PC board, surround
each non-Kelvin terminal of the shunt with numerous
plated-through vias. Use 2 or 4 oz. per square feet of
copper for the layers of the PC board; this will result in a
current-carrying capacity in excess of 20 A. Making the
current-carrying traces on the PC board fairly large can
also improve the shunt’s power dissipation capability by
acting as a heat sink. Liberal use of vias where the load
current enters and exits the PC board is also recommend-
ed.