Look at the datasheet for a strain-gauge pressure transmitter or a load cell and you will find a line like full-bridge, 2 mV/V. That specification refers to a Wheatstone bridge, the four-resistor circuit that sits at the heart of nearly every strain-gauge sensor, RTD, and load cell. The bridge is what turns a tiny change in resistance into a measurable voltage, and it does so in a way that cancels much of the drift that would otherwise ruin the reading. This guide explains how a Wheatstone bridge works, what balanced and unbalanced mean, the difference between quarter, half, and full configurations, and why the millivolt-per-volt figure describes the output.
Wheatstone bridge in one line: A Wheatstone bridge is a circuit of four resistors arranged in two parallel voltage dividers powered by an excitation voltage, with the measured output taken as the voltage difference between the midpoints of the two dividers. When all four resistors are matched the bridge is balanced and the output is zero; when a sensing resistor changes, the bridge goes unbalanced and produces a small output voltage proportional to that change, expressed as millivolts of output per volt of excitation.
Picture four resistors wired into a diamond. An excitation voltage is applied across one diagonal of the diamond, and the output is measured across the other diagonal. Each side of the diamond is really a voltage divider - two resistors in series - so the excitation splits at the midpoint of each side according to the ratio of its two resistors. The bridge output is simply the difference between those two midpoint voltages. This arrangement is what makes the bridge so sensitive: it measures a difference between two dividers rather than an absolute voltage.
When the four resistors are matched so that both dividers split the excitation in the same ratio, the two midpoints sit at the same voltage and the output is zero. This is the balanced, or null, condition, and it is the bridge's great advantage. A well-designed sensor is set up so that at its resting state the bridge is balanced and reads zero, which means the measurement starts from a clean, stable reference rather than from some large offset voltage that would have to be subtracted.
Now change one of the resistors - which is exactly what a strain gauge does when it is stretched, an RTD does when it warms, or a load cell element does under load. That divider now splits the excitation differently, its midpoint voltage shifts, and the balance is broken. The bridge output moves away from zero by an amount proportional to how much the resistance changed. Because the starting point was a true null, that output is a clean signal representing the physical quantity being measured, sitting on top of nothing rather than buried in a large baseline.
The three common bridge configurations differ in how many of the four resistors are active sensing elements. In a quarter bridge, only one arm is a strain gauge and the other three are fixed reference resistors. It is the simplest and cheapest arrangement, but it is also the least sensitive and the most exposed to error, because only one arm responds to the measured quantity while temperature and lead effects act on it largely uncancelled. Quarter bridges are common in general strain measurement where simplicity matters more than the last bit of accuracy.
A half bridge makes two of the four arms active, and this is where the configuration starts to earn its keep. If the two active gauges are arranged so that the measured quantity pushes one up and the other down - for example one gauge in tension and one in compression on a bending element - their effects add in the output while common effects like temperature, which change both gauges the same way, tend to cancel. The half bridge roughly doubles the sensitivity of a quarter bridge and improves its rejection of unwanted influences.
A full bridge makes all four arms active, with pairs arranged to add their signal contributions while cancelling common-mode effects. This gives the highest sensitivity and the best cancellation of temperature and other shared errors, which is why precision load cells and quality pressure transducers use full bridges. The tradeoff is cost and complexity - four matched, properly placed gauges - but for custody-grade and high-accuracy measurement that cost is well spent, and it is why a serious transducer datasheet almost always specifies a full bridge.
The bridge output is not a fixed voltage; it scales with the excitation applied across it. Double the excitation and you double the output for the same physical input. That is why sensor makers rate the output as millivolts per volt of excitation - a figure like 2 mV/V means that at full-scale load or pressure, and with the bridge excited by a given voltage, the output is two millivolts for every volt of excitation. Excite a 2 mV/V full-bridge sensor at 10 volts and its full-scale output is 20 millivolts. Expressing output this way makes the specification independent of whatever excitation the instrument actually supplies.
This dependence on excitation is also why the excitation voltage has to be stable and why many systems use a technique to measure it - either regulating it tightly or sensing it and computing the ratio - so that a drift in excitation does not masquerade as a change in the measured quantity. It is also why the bridge cancels temperature drift so effectively: because the output is a ratio of the two dividers, effects that change all the arms together, such as ambient temperature acting on the excitation or on matched gauges, move both midpoints alike and largely cancel in the difference rather than showing up as error.
In a SCADA installation, the raw bridge output is a low-level differential signal of a few millivolts, so it is conditioned close to the sensor - amplified, sometimes digitized in a smart transmitter - and often converted to a standard 4-20 mA or digital value before it reaches the input module. A cloud SCADA platform such as Merobix receives that scaled signal and applies the sensor's calibration to present engineering units. Knowing that the number originated as a mV/V bridge output makes field problems legible: a reading that will not zero points to a bridge that is not balanced or a shifted reference, and a reading that scales wrong points to an excitation or calibration issue, both traceable to how the bridge produces its signal.
It is the bridge's full-scale sensitivity: at rated load or pressure, the sensor produces two millivolts of output for every volt of excitation applied to the bridge. Excite it at 10 volts and the full-scale output is 20 millivolts. Rating the output per volt of excitation makes the specification independent of the actual excitation voltage the instrument supplies, so the same sensor works across systems with different excitation.
The terms describe how many of the bridge's four arms are active sensing elements. A quarter bridge has one active gauge and three fixed resistors - simplest but least sensitive. A half bridge has two active arms arranged to add signal while cancelling common effects like temperature. A full bridge makes all four arms active for the highest sensitivity and best drift cancellation, which is why precision load cells and pressure transducers use it.
Because the output is the difference between two voltage dividers rather than an absolute voltage, any effect that changes all the arms in the same way moves both divider midpoints together and largely cancels in that difference. Temperature acting equally on matched gauges is the main example. Half and full bridges enhance this by arranging active gauges so the measured quantity adds while shared influences subtract out, leaving a cleaner signal.
Merobix reads your field devices into a cloud SCADA - the real thing behind these terms, live in days from any browser.