Shunt Resistor Packages: Power and Thermal Limits

A shunt resistor is the component that turns current into a voltage a circuit can measure, and it is chosen on the assumption that it will stay at the value printed on the reel. That assumption holds only if the part stays cool enough, and staying cool is a joint property of the resistor, the copper it is soldered to and the air around it. Package selection is therefore a thermal decision as much as an electrical one.

What a Shunt Resistor Has to Do

The device converts current into a small voltage through a known resistance, and everything downstream depends on that resistance being stable. The value moves with temperature, with the mechanical stress that the board applies to the part, and with the soldering process. A shunt resistor that measures correctly on the bench but drifts in service is usually a thermally inadequate choice rather than a defective part.

The second requirement is that the voltage developed across the part is measured at exactly the points where the resistance is defined. Any copper shared between the current path and the sense path adds an unknown term, and that term is proportional to current. The four-terminal, or kelvin, connection exists to remove it.

Package Size and Power Dissipation

Package size determines the surface available to dissipate heat, and the relationship is not proportional to the footprint. A larger part has more area in contact with the board and more area exposed to air, but it also has a longer thermal path through its own body, and the metal alloy used has a lower thermal conductivity than copper. The datasheet rating is therefore a starting point defined under a specific mounting condition.

The practical consequence is that a part rated for one watt in a laboratory test may be limited to a fraction of that on a real board. The limiting factor is usually the temperature of the solder joint, because the joint is what couples the part to the copper, and the joint is also what fails when the temperature cycles.

Thermal Resistance and Derating

Power dissipation in a shunt resistor produces heat, and heat leaves through two paths: into the board through the terminations, and into the air from the body. Each path has a thermal resistance, and the parallel combination determines the temperature rise for a given power. Because the board path usually dominates, the copper geometry around the part is more important than the airflow over it.

Derating curves are published against a mounting standard, and it is worth checking which one. A curve measured on a high thermal conductivity test board with generous copper will be optimistic for a design that mounts the part on a narrow trace with a plane underneath. Our notes on PCB thermal design and cooling describe how to estimate the real thermal path rather than relying on the headline number.

Kelvin Connection and Measurement Accuracy

Accuracy depends on where the sense voltage is taken. If the sense traces leave from the same pads that carry the current, they pick up the voltage drop in the solder joint and in the pad copper, and that drop changes with temperature and with solder volume. Moving the sense points to the inner edge of the resistor terminations removes most of it.

The layout rule that follows is simple to state and easy to break. The current path should enter and leave the pads, the sense traces should run from inside the pad area with no current flowing in them, and the two sense traces should be routed as a pair so that they see the same environment. Any asymmetry between them becomes a measurement error that no calibration can separate from the signal.

Copper Area as a Heat Sink

Because most of the heat leaves through the terminations, the copper attached to them is the real heat sink. A generous pad, a short wide connection to a plane, and thermal vias beneath the body all reduce the temperature rise. Where the current is large, it is common to see the sense resistor placed over a small copper island connected to inner planes by a via array.

There is a trade-off with soldering. A very large copper area draws heat away during reflow and can leave a cold joint, so the layout often uses thermal reliefs on the plane connection or a deliberate split to control the heating. Our notes on trace width current calculation cover the related question of sizing the conductors that carry the current into and out of the part.

Parasitic Inductance at High Frequency

A shunt resistor is not a pure resistance. Its geometry and the loop formed with the sense traces introduce inductance, and the inductance produces a voltage proportional to the rate of change of current rather than to the current itself. In a switching circuit with fast edges, that term can dominate the measurement during the transition.

Keeping the sense loop small is the principal remedy. A short, tight pair of sense traces, routed together and referenced to a plane, reduces the loop area and the inductance. Where the measurement must be accurate during a switching event, a coaxial or twisted sense connection may be justified even at the cost of routing complexity.

>

Choosing Between Packages

The choice is usually made between a small chip resistor, a larger chip, and a metal element part with a defined four-terminal construction. Small chips are cheap and easy to place but have limited power capability and a temperature coefficient that is adequate for monitoring rather than for precision measurement. Metal element parts offer a lower temperature coefficient, better long-term stability and a defined sense point.

Where the current is high and the measurement matters, the metal element part usually wins despite its cost, because the alternative is a calibration procedure that has to be repeated as the part ages. Where the current is modest and the requirement is protection rather than measurement, a standard chip is entirely adequate.

Manufacturing and Assembly Notes

Assembly affects the result as much as layout. Solder volume at the terminations sets the thermal path and also applies mechanical stress to the part as it cools, and an asymmetric solder fillet can tilt a small resistor enough to shift its effective value. Our notes on PCB pad design standards cover the pad geometry that produces a symmetric joint.

Where the part is used for measurement rather than protection, it is worth measuring a sample after assembly and comparing the result with the value measured on the unpopulated board. The difference includes the solder joint, the placement and any stress the board applies, and it is the number that matters for the accuracy of the finished product.

FAQ

Can a shunt resistor be run at its rated power? Rarely on a real board. The rating assumes a specific mounting and ambient condition, and a conservative design usually runs the part at a fraction of it to keep the sense voltage stable over temperature.

Is a four-terminal part always necessary? No. It is necessary when the accuracy target is tighter than the voltage drop in the terminations allows. For protection and coarse monitoring, a two-terminal part is often sufficient.

How much copper should be attached to the pads? As much as the layout allows while keeping the joint manufacturable. The copper is the heat sink, so reducing it raises the part temperature directly.