A piezoresistive pressure sensor converts pressure into a small electrical signal in four steps: pressure deflects a diaphragm; the resulting strain changes the resistance of sensing elements; a Wheatstone bridge turns those changes into a differential voltage; and electronics amplify, digitize, or transmit that voltage. The raw bridge output is commonly measured in millivolts, not directly in bar or psi. It needs an excitation source and a suitable readout circuit before an OEM controller can use it reliably.
This guide follows the electrical path rather than repeating a broad comparison of sensing technologies. It defines the four bridge nodes, works through an explicitly hypothetical resistance example, explains mV/V sensitivity, and identifies the questions to ask when integrating a real sensor core. For technology trade-offs, see our capacitive versus piezoresistive pressure sensor guide.
From pressure to a resistance change
The process pressure acts on a thin diaphragm. As it bends, regions near its edges experience different mechanical stresses. Piezoresistive elements placed in appropriate regions respond with changes in electrical resistance. In a silicon sensor, the resistors may be formed on the sensing die; ceramic thick-film designs use a different material and construction but can also employ a resistive bridge. The details of the stress field, resistor placement and package belong to the specific sensor design.
For a common four-active-element arrangement, two elements increase in resistance under the selected pressure direction and two decrease. Correctly placing those elements in the bridge makes their effects add in the differential output. A Kistler explanation of the piezoresistive effect describes the four-resistor silicon diaphragm; a Texas Instruments pressure-bridge design note follows the resistance change through the signal chain. The arrangement described here is representative, not a statement that every pressure sensor has the same die layout.
The pressure medium does not necessarily touch the silicon die. An isolation diaphragm and fill fluid may transfer pressure to it. That separation affects media compatibility and the package construction, but the electrical explanation begins at the strain-sensitive resistors. A raw core, a conditioned module and a field transmitter expose different signals; our sensor, transducer and transmitter guide explains those integration levels.
How are the four resistors connected?
Think of the Wheatstone bridge as two voltage dividers supplied by the same excitation. To make the equation unambiguous, use this specific node convention:
| Bridge position | Resistor | Terminal relationship |
|---|---|---|
| Top left | R1 | Between EXC+ and left midpoint |
| Bottom left | R2 | Between left midpoint and EXC− |
| Top right | R3 | Between EXC+ and right midpoint |
| Bottom right | R4 | Between right midpoint and EXC− |
Apply an excitation voltage VEX = V(EXC+) − V(EXC−) across the top and bottom terminals. Define V+ as the left midpoint and V− as the right midpoint. With negligible load on those output nodes, each side is an ordinary voltage divider:
V+ = VEX × R2 / (R1 + R2)
V− = VEX × R4 / (R3 + R4)
Therefore, the differential bridge output under this chosen polarity is:
VOUT = V+ − V− = VEX × [R2/(R1 + R2) − R4/(R3 + R4)].
If all four resistors are exactly equal in an ideal model, both midpoints sit at half the excitation voltage and VOUT = 0. Real devices can have nonzero zero-pressure output because matching is imperfect and mechanical, thermal or electrical offsets remain. Swapping OUT+ and OUT− reverses the sign of the measured output; a sign reversal alone does not mean the sensor is damaged. Use the ordered model’s pinout and pressure-reference definition, not this illustrative node convention, when wiring hardware.
A fully stated resistor example
Consider a teaching circuit with four nominal 10,000 Ω resistors and 5 V voltage excitation. Assume pressure changes each active resistance by 10 Ω in a symmetric pattern: R1 = R4 = 9,990 Ω og R2 = R3 = 10,010 Ω. This is a mathematical illustration, ikke a measurement or specification of a Winsen product.
The left midpoint is 5 × 10,010 / 20,000 = 2.5025 V; the right midpoint is 5 × 9,990 / 20,000 = 2.4975 V. Their difference is +0.005 V, or +5 mV. Equivalently, for this symmetric arrangement, VOUT = VEX × ΔR/R = 5 V × 10/10,000 = 5 mV.
This arithmetic shows why a small resistance change can be sensed without measuring each resistor separately. It also shows why the output is a difference between two nodes, even though both nodes are near 2.5 V. At another pressure, resistor changes and output need not stay perfectly linear; calibration determines the useful pressure-to-output mapping.
Excitation and mV/V sensitivity
A passive bridge needs an electrical drive. Many raw cores are characterized using a specified constant voltage eller specified constant current. Follow the exact model data sheet: a 10 V test condition does not authorize driving an unrelated sensor at 10 V, and a 1.5 mA current condition is not equivalent to 10 V. Input resistance, power dissipation, self-heating, supply limits and cable voltage drop all matter.
Under constant-voltage excitation, bridge output is proportional to excitation in the ideal model. That is why a product may state sensitivity in millivolts per volt of excitation (mV/V). For example, the official Winsen WPAH01 ceramic pressure sensor page lists a 2–20 V working-voltage interval and a sensitivity range of 1.5–4 mV/V, with a stated typical value of 2.5±1.0 mV/V. Using 2.5 mV/V as an arithmetic example and 5 V excitation gives 2.5 mV/V × 5 V = 12.5 mV nominal full-scale output. The unit “per volt” cancels the excitation volt. This calculation is not a guaranteed output for a particular ordered WPAH01 or an instruction to use 5 V without checking the current data sheet and selected range.
Winsen’s WPAK63 isolation-film pressure sensor page lists 1.5 mA or 10 V excitation options and full-range output values that vary with range and drive condition. Its official manual identifies excitation and differential output terminals. For instance, the manual’s 10 V column and the 1.5 mA column report different millivolt quantities. Choose the ordered variant and its matching electrical table before sizing an amplifier. Do not substitute a generic mV/V number for a constant-current output specification.
From raw millivolts to a controller value
A raw bridge signal is usually too small and too close to a substantial common-mode voltage for a typical single-ended controller input. In the ideal 5 V example above, each sense terminal is around 2.5 V while the useful difference is just 5 mV. The front end must accept the midpoints’ common-mode range and resolve their difference without excessive offset or noise.
A practical signal path typically has these stages:
- Excitation: a supply or current source within the sensor’s specified limits, measured at the bridge if lead resistance is significant.
- Differential input: a high-input-impedance instrumentation amplifier, programmable-gain amplifier or suitable differential ADC input that does not load the bridge appreciably.
- Filtering and conversion: bandwidth chosen for the actual pressure dynamics, followed by an ADC with enough noise-free resolution for the permitted pressure error.
- Calibration and scaling: zero/span corrections and temperature compensation as required by the core and system specification.
- System output: pressure units, a conditioned voltage, digital data, or an industrial current interface implemented by the module or transmitter electronics.
An Analog Devices application note on bridge measurements explains a ratiometric ADC arrangement: using a reference derived from the same voltage that excites the bridge can reduce first-order sensitivity to excitation-voltage changes. The example in that note is a load cell, but the voltage-bridge principle is applicable here. Ratiometric conversion does not cure bridge offset, nonlinearity, temperature drift, wire resistance, poor input common-mode range or inadequate filtering. Constant-current excitation requires its own system analysis.
Do not infer that a raw mV bridge can be connected directly to a 4–20 mA PLC input. The loop interface needs conditioning and transmitter electronics. Conversely, a transmitter that already provides a conditioned signal should not be treated as though its external terminals expose the raw bridge.
Why does a bridge show output at zero pressure?
The balanced equations describe an ideal circuit. In practice, resistor mismatch, diaphragm residual stress, package assembly, amplifier input offset and changing temperature can create a zero offset. The output may also change with temperature because bridge resistance and sensitivity are temperature dependent. Lead and contact resistance can change the effective excitation at the bridge; power dissipation can warm a small sensing element.
The remedy is not to assume every offset is “noise” and filter it away. Determine whether the zero value is within the data sheet, whether the supply and input configuration match the specified test condition, and whether the system requires zero adjustment or multi-point calibration. Separate the raw core’s specification from the finished assembly’s full error budget. A compensated transmitter may include adjustments that a bare core leaves to the OEM.
When evaluating samples, record raw OUT+ − OUT− at zero, a middle pressure and full scale, then repeat as pressure decreases. Repeat at relevant temperatures if the product must work outside a laboratory reference condition. Record the actual voltage across EXC+ and EXC− at the sensor, not only the power supply setting. Use a pressure reference appropriate to the desired uncertainty and reference type. Our Day02 accuracy guide explains why %FS, %reading and temperature terms should not be silently interchanged.
How do WPAK63 and WPAH01 fit an OEM design?
Both model pages are useful starting points, but they represent different core constructions and electrical specification styles. The bridge discussion above explains what to ask of each; it does not transfer dimensions, pinouts or performance between them.
| Model | Verified manufacturer information | Integration decision |
|---|---|---|
| WPAK63 | Isolation-film pressure sensor core with 316L diaphragm and housing, silicone oil fill, differential millivolt output and 1.5 mA or 10 V excitation choices; the manual provides pin/lead definitions and range-dependent output data. | Useful where an OEM wants a raw pressure core for its own amplifier and calibration chain. Confirm ordered range, reference, excitation option, pinout, media compatibility and exact full-scale output. |
| WPAH01 | Ceramic piezoresistive pressure sensor; official page gives 2–20 V working voltage and 1.5–4 mV/V sensitivity with a typical 2.5±1.0 mV/V figure. | Useful when a ceramic core and mV/V-style voltage-excited interface fit the assembly. Confirm ordered range, mechanical sealing, exact sensitivity and compensation/calibration requirements. |
WPAK63 is ikke a drop-in equivalent of WPAH01. The manufacturer page and an older WPAH01 manual differ in the breadth of pressure ranges listed, so request current variant-specific documentation before writing a pressure-range claim into a purchase specification. Neither AI image in this article is an official photo, drawing or pinout of either model.
An OEM bridge-interface handoff checklist
Before releasing a circuit or an RFQ, ask the supplier for the following information for the exact ordered core:
- Pressure reference and calibrated range, plus proof and burst limits.
- Excitation mode and limits, bridge input resistance, supply tolerance, and maximum permissible power.
- De actual EXC+/EXC− and OUT+/OUT− pin or wire assignment, polarity, and cable length.
- Zero output, full-scale output or mV/V sensitivity, with minimum, typical and maximum values clearly separated.
- Temperature compensation interval, zero and span drift, hysteresis, repeatability and long-term stability definitions.
- Media-wetted materials, seal choice, mounting and any isolation fluid.
- Sample test method: pressure points, temperature points, source impedance, amplifier/ADC configuration and acceptance limits.
The most useful first calculation is often simple: convert the model’s output rating to actual millivolts at the chosen excitation, then check the amplifier input range and the ADC’s usable resolution. The most useful first experiment is equally simple: measure both midpoint voltages, their difference, and the bridge excitation at a known pressure. Those four quantities reveal more than a single controller reading when an integration is not behaving as expected.
Ofte stillede spørgsmål
Does a Wheatstone bridge produce pressure directly?
No. It produces an electrical differential voltage related to pressure-induced resistance changes. Calibration and signal processing map that voltage into pressure units under stated conditions.
Why are four resistors often used?
Four active elements can place opposite strain responses in bridge arms so their differential effects add. The exact physical arrangement depends on the sensing die or ceramic element; a pressure sensor can have a different topology or extra compensation components.
What does mV/V mean on a pressure-sensor data sheet?
It is millivolts of output per volt of specified excitation, generally at a defined pressure span and condition. Multiply the stated mV/V figure by the actual allowed excitation voltage to obtain a nominal millivolt span, then check tolerance, offset and temperature terms.
Can I read only OUT+ with a single-ended ADC?
Not reliably without an appropriate reference and input design. The useful signal is OUT+ − OUT−, while each node may sit near half the bridge excitation. Use a differential front end or a manufacturer-approved conditioned output.
Does ratiometric conversion remove all supply error?
No. Sharing excitation and ADC reference can cancel a first-order change in excitation for a voltage-driven bridge. It does not eliminate offset, temperature effects, lead drops, amplifier limitations or other sensor errors.
Is a zero-pressure reading of a few millivolts automatically a fault?
No. Real bridges may have a specified zero-output offset. Compare it with the exact model’s limits and test condition, then calibrate or troubleshoot if it exceeds them.
Request the right bridge data for your design
If your equipment will use a raw pressure core, send Winsen the pressure range and reference, process medium, mechanical space, target excitation, available ADC or amplifier, temperature interval, allowable pressure error and expected quantity. Ask for the current WPAK63 or WPAH01 variant data sheet, pinout, zero/span distributions and sample test conditions through the Winsen contact page. That makes the electrical interface review concrete before a PCB or housing is frozen.
