A shunt turns current into a small voltage, and shunt resistor Kelvin sensing is what keeps that voltage honest. The four-terminal arrangement exists because the resistance of a bolted joint is comparable to the resistance being measured. Sense where the manufacturer marked the sense points, or your reading quietly includes hardware you never meant to measure.
Why a current shunt has four terminals instead of two
A shunt is a precision resistor, often only a few milliohms, placed in series with the load. Ohm’s law turns the current through it into a voltage an instrument can digitize. The difficulty is that a few milliohms is the same order of magnitude as the resistance of a bolted lug, a short length of busbar, or a washer that has begun to oxidize. If the voltmeter picks up its signal from the current-carrying bolts, it reads the shunt plus whatever metal sits between its two pickup points. That extra resistance is not stable either: it shifts when the bolt is retorqued, when the joint corrodes, and when the hardware warms. Four-terminal construction splits the job in two. The outer pair carries load current. The inner pair, the sense terminals, carries almost none, so the drop along them is negligible and the instrument sees the resistive element alone.
Where the sense connection changes the answer
Picture a shunt rated 50 A for 50 mV output, which is one milliohm. A joint contributing a hundred microhms of contact resistance is ten percent of that. Sense across the load bolts and every reading runs ten percent high, and because the error scales with current it looks exactly like a gain error in the instrument rather than a wiring mistake. Move the sense wires to the marked points and the same shunt reads correctly. This is why those points are physically identified, often as drilled holes or slots cut into the strap. Manufacturers place them where current distribution inside the element has settled into a uniform pattern, which is not always the geometric middle.
Temperature moves the number more than people expect
Pushing current through a resistor heats it, and heat changes resistance. Shunt alloys such as manganin are chosen because their temperature coefficient is small, not because it is zero. The practical consequence is that a shunt reads slightly differently one second after a load step than it does five minutes later, and the drift runs in the direction of its own self-heating. Manufacturers publish a temperature coefficient and a power derating curve for precisely this reason, and the derating is usually severe, because a shunt operated at full rating is a small space heater. Airflow, the mass of the busbar it bolts to, and the ambient temperature inside the enclosure all change where it settles. If you need accuracy at high current, size the shunt so it never approaches its maximum dissipation, and let the reading stabilize before you trust it.
Wiring mistakes that show up as steady gain error
Sense leads carry almost no current, but they are not immune to trouble. Twist the pair so it does not enclose a loop that magnetic fields from the nearby load conductor can drive a voltage into. Route it away from power cabling rather than bundling everything into one loom. A junction of dissimilar metals in the sense path behaves as a thermocouple, and a temperature gradient across an unbalanced pair produces a wandering microvolt offset that reads as drift. Keep both sense terminations physically close and at similar temperature so those thermal voltages tend to cancel. On the instrument side, input impedance is finite; a high sense-lead resistance working into a modest input impedance forms a divider that quietly attenuates the signal. Check the specified input impedance rather than assuming it is effectively infinite.
How to read a shunt’s datasheet without guessing
Every genuine specification a shunt has lives in its datasheet, and none of it can be inferred from how the part looks. The figures that matter are nominal resistance and its tolerance, full-scale output voltage, accuracy class, temperature coefficient, continuous current rating, and the derating curve along with whatever busbar and airflow that curve assumes. A shunt stamped 100 A and 100 mV tells you the ratio, nothing more. Two shunts with identical markings can belong to different accuracy classes and behave very differently across a temperature range. If the shunt feeds a battery monitor or an energy meter, that instrument also has to be configured for the shunt in front of it; a mismatched setting produces numbers that are perfectly repeatable and perfectly wrong. When a label is missing or unreadable, treat the part as unknown rather than assuming a common value.
What to check before you act
- Land the sense leads on the shunt’s marked sense points, never on the load bolts.
- Twist the sense pair and route it clear of power conductors.
- Confirm resistance, rating and accuracy class from the datasheet, not the markings.
- Let the reading settle so self-heating stops moving the number.
Common questions
Can I use an ordinary power resistor as a current shunt?
You can sense current across one, but its tolerance and temperature coefficient are usually far worse than a purpose-built shunt, so the reading drifts with load. Treat it as a rough indication, not as metering.
Why does my battery monitor read high with a brand-new shunt?
The two usual causes are sense leads landed outside the marked sense points and a shunt value in the monitor’s configuration that does not match the part installed. Check both before suspecting the resistor.
The practical takeaway
A shunt is only as accurate as the two points its sense leads touch. Land them on the marked sense terminals, keep the pair twisted and away from power conductors, and configure the monitor for the shunt that is actually fitted. If a reading is still off by a consistent percentage, suspect the sense connection long before you suspect the resistor.
References and further reading
- CDC/NIOSH: overview of electrical hazards and safe work practices
- NFPA: electrical safety resources and code background
- Department of Energy: estimating appliance and electronic energy use
Numerical scenarios are illustrative unless identified otherwise. Follow the exact product instructions; component ratings and local installation requirements can differ.
