Optimize High-Current (a) 6.

35mm

Sensing Accuracy by
Improving Pad Layout of 3.18mm

Low-Value Shunt Resistors

By Marcus O’Sullivan 2.68mm

Introduction
(b)
Current-sense resistors, which come in a variety of shapes and
8.35mm
sizes, are used to measure current in many automotive, power
control, and industrial systems. When using very low value
resistors (a few milliohms or less), the resistance of the solder
becomes a substantial portion of the sense element resistance
and adds significantly to the measurement error. High-accuracy 1.45mm

applications often use 4-terminal resistors and Kelvin sensing

to reduce this error, but these special-purpose resistors can be
1.8mm 1.0mm 3.9mm
expensive. In addition, the size and design of the resistor pads
play a crucial role in determining the sense accuracy when
measuring large currents. This article describes an alternative
approach that enables high-accuracy Kelvin sensing using a
standard, low-cost, 2-pad sense resistor with a 4-pad layout.
Figure 1 shows the test board used to characterize the errors 4.75mm
caused by five different layouts.
Figure 2. (a) ULRG3-2512-0M50-FLFSLT resistor
dimensions; (b) Standard 4-pad footprint.

Traditional Footprint
For Kelvin sensing, pads in the standard 2-wire footprint must be
divided to provide separate paths for the system currents and the
sense currents. Figure 3 shows an example of such a layout. The
system current takes the path shown by the red arrows. If a simple
2-pad layout was used, the total resistance would be:
RTOTAL = RSENSE + (2 × RSOLDER)
To avoid the additional resistance, voltage sensing traces need to
be routed right to the sense resistor pad. The system current will
still cause a significant voltage drop across the upper solder joints,
but the sense currents will cause a negligible voltage drop across
the lower solder joints. This split pad approach thus removes the
solder joint resistance from the measurement and improves the
total system accuracy.

PCB PAD SENSE RESISTOR PCB PAD

RSOLDER RSOLDER
SYSTEM CURRENT
Figure 1. Sense resistor layout test PCB.
RESISTOR PAD RESISTOR PAD
SENSE SENSE
Current-Sense Resistor CURRENT CURRENT
Available in resistance values as low as 0.5 mΩ, commonly used RSOLDER RSOLDER
current-sense resistors packaged in a 2512 case can dissipate up to PCB PAD PCB PAD
3 W. In order to highlight the worst-case errors, these experiments
employ a 0.5-mΩ, 3-W resistor with 1% tolerance (part number SENSE+ SENSE–
ULRG3-2512-0M50-FLFSLT from Welwyn/TTelectronics). Its
dimensions and the standard 4-wire footprint are shown in Figure 2. Figure 3. Kelvin sensing.

Analog Dialogue 46-06 Back Burner, June (2012) www.analog.com/analogdialogue 1

Optimizing the Kelvin Footprint at 25°C. The resulting differential voltage was measured less than
The layout shown in Figure 3 is a significant improvement on the 1 second after the load current was enabled to prevent the resistor
standard 2-pad approach, but with very low value resistors (0.5 mΩ temperature from rising by more than 1°C. The temperature
or less), the physical location of the sensing point on the pad and of each resistor was monitored to ensure the test results were
the symmetry of the current flow through the resistor become more recorded at 25°C. At 20 A, the ideal voltage drop across a 0.5-mΩ
significant. For example, the ULRG3-2512-0M50-FLFSLT is a resistor is 10 mV.
solid metal-alloy resistor, so every millimeter of the resistor along
the pad will influence the effective resistance. Using a calibrated T = 25°C
current, an optimum sensing layout was determined by comparing +
INPUT
A
OUTPUT

the voltage drops across five custom footprints. X Y Y X

Test PCB C
X Y Y X

Figure 4 shows five layout patterns, labeled A though E, created PSU X Y Y X

ELOAD
D
on a test PCB. Where possible, traces were routed to test points at V = 5V I = 20.00A
X X

different locations along the sense pads, as indicated by the color E

coded dots. The individual resistor footprints are: X Y

SENSE RESISTOR LAYOUT TEST

Y X

–
A. Standard 4-wire resistor based on 2512 recommended footprints
(see Figure 2(b)). Sense point pairs ( X and Y ) at the outer and
inner edges of the pads (x-axis). DMM

B. Similar to A but with pads elongated closer inwards to allow

better coverage of the pad area (see Figure 2(a)). Sense points
at the center and end of the pads. Figure 5. Test setup.
C. Provides more symmetrical system current flow by using both Test Results
sides of the pad. Also moves sense point to a more central Table 1 shows the measured data using the sense pad locations
location. Sense points are at the center and end of the pads. shown in Figure 4.
D. Similar to C but with system current pads joined at the innermost
point. Only the outer sense points are used. Table 1. Measured Voltages and Errors
E. Hybrid of A and B. The system current flows through the wider Measured
pads and the sense current flows through the smaller pads. Sense Footprint Sense Pad (mV) Error (%)
points are at the outer and inner edges of the pads. A Y 9.55 4.5

INPUT OUTPUT X 9.68 3.2

A B Y 9.50 5
X 9.55 4.5
X Y Y X
C Y 9.80 2
B X 9.90 1
D X 10.06 0.6
X Y Y X
E Y 9.59 4.1
C
X 9.60 4

X Y Y X Top pad* 12.28 22.8

*Without Kelvin sensing. Voltage was measured across the main high-current
D
pad to demonstrate the error associated with the solder resistance.

X X Observations
1. Footprints C and D exhibit the lowest errors, with comparable
E results and variation within the individual resistor tolerance.
Footprint C is preferred as it is less likely to cause problems
X Y Y X
related to component placement tolerances.
SENSE RESISTOR LAYOUT TEST 2. The sense points at the outer extremity of the resistors provide
the most accurate results in each case. This indicates the
Figure 4. Test PCB layout. resistors are sized by the manufacturer to the entire length.
3. Note the 22% error associated with the solder resistance without
Solder was applied using a stencil and reflowed in a reflow oven.
using Kelvin sensing. This is an equivalent solder resistance of
The ULRG3-2512-0M50-FLFSLT resistor was used.
about 0.144 mΩ.
Test Procedure 4. Footprint E demonstrates the effects of asymmetrical pad layout.
The test setup is shown in Figure 5. A calibrated current of 20 A During reflow, the component is pulled to the pad with the most
was passed through each resistor while the resistor was kept solder. This type of footprint should be avoided.

2 Analog Dialogue 46-06 Back Burner, June (2012)

Conclusion
Based on the results shown, the optimum footprint is C, with
an expected measurement error of less than 1%. Recommended
dimensions of this footprint are shown in Figure 6. RSENSE

8.40mm

1.60mm

1.20mm

0.60mm 0.50mm

4.00mm
Figure 7. Recommended PCB trace routing.
0.50mm
The data in this article may not be applicable to all resistors, and
1.20mm results may vary, depending on resistor composition and size.
Resistor manufacturers should be consulted. It is the responsibility
of the user to ensure that the layout dimensions and structure
of the footprint comply with individual SMT manufacturing
3.40mm 3.40mm
requirements. Analog Devices, Inc., does not accept responsibility
for any issues that may arise as a result of using this footprint.
Figure 6. Optimum footprint dimensions.
Author
The routing of the sense trace also had an impact on the measure- Marcus O’Sullivan [marcus.osullivan@analog.com]
ment accuracy. To achieve the highest accuracy, the sense voltage joined Analog Devices in 1999 and works as an
should be measured at the edge of the resistor. Figure 7 shows a applications engineer with the Power Management
recommended layout using vias to route the outer edge of the pad Group. He holds a BEng degree in electronic
to another layer, thus avoiding cutting the main power plane. engineering from the University of Limerick.

Analog Dialogue 46-06 Back Burner, June (2012) 3