Choosing Kelvin Pickup Points on a Current-Shunt Footprint

At low shunt resistance, a small amount of unintended copper resistance can become a significant measurement error. A pair of thin sense traces is not automatically a Kelvin connection: their origin on the footprint determines which voltage drop the amplifier actually measures. Review those pickup points before judging the amplifier’s accuracy.

Define the measured resistance

Analog Devices shows how pad layout affects high-current measurement with low-value shunts and discusses four-terminal sensing arrangements. Its examples demonstrate why the resistor and footprint must be considered together. Use the resistor manufacturer’s recommended connection geometry rather than assuming every two-pad device can share the same split-pad pattern. Analog Devices: shunt pad layout and accuracy.

Draw the load-current path through the resistor and mark where the voltage-sense traces branch from it. Identify any copper segment that carries load current before the sense pickup. Its voltage drop may become part of the measured signal. For a four-terminal resistor, verify that the schematic and footprint distinguish power and sense terminals correctly.

Estimate why the geometry matters

As an illustrative calculation, a 1 milliohm shunt carrying 10 A develops 10 mV. An unintended additional 0.1 milliohm included by the pickup contributes another 1 mV at that current. That is a 10 percent error relative to the ideal shunt signal, before amplifier errors are considered. These values are a calculation example, not a suggested operating point or component rating.

Audit the sense route

• Keep the pair associated with the intended shunt terminals through layer changes.
• Avoid sharing a sense pickup with an unrelated load-current connection.
• Review input filtering and routing with the amplifier’s application guidance.
• Check common-mode range and transient conditions separately from the copper layout.
• Consider temperature distribution around the shunt and its effect on the measurement.

A neat differential-looking pair routed from the outer power pours can still measure the wrong resistance. Conversely, the correct pickup points may require a carefully defined footprint even when the remaining route is short.

Verify the released implementation

Compare measured output with a suitable current reference at several operating points and temperatures relevant to the product. Record the shunt part number, footprint revision, copper construction, and sense connections. Calibration may compensate a repeatable gain error, but it does not justify an uncontrolled geometry or a temperature-dependent error that was never evaluated. Keep the accepted pickup arrangement explicit so a later library cleanup does not merge sense and power pads and silently change the measurement.