If you need to measure a tiny change in resistance—like a strain gauge bending or a thermistor tracking ambient temperature—a simple voltage divider will leave you fighting noise and lead-wire errors. The solution is the Wheatstone bridge. By applying fundamental circuit formulas physics principles (specifically Kirchhoff’s Voltage Law and Ohm’s Law), we can translate micro-ohm or milli-ohm resistance shifts into clean, differential voltage signals that an ADC or instrumentation amplifier can actually read.
This guide walks through the physical topology, selects real-world components for a temperature-sensing bridge, maps out exactly what happens when components fail, and provides a step-by-step breadboard verification process.
The Topology: Nodes, Legs, and the Core Physics Formula
The Wheatstone bridge consists of two parallel voltage dividers sharing a common excitation voltage. Let’s label the nodes to keep the physics grounded:
- Node A (Top): Excitation voltage ($V_{ex}$), typically 3.3V or 5.0V DC.
- Node C (Bottom): Ground reference (0V).
- Node B (Right Mid): Output positive ($V_{out+}$), between fixed resistors $R_3$ and $R_4$.
- Node D (Left Mid): Output negative ($V_{out-}$), between $R_1$ and the sensor $R_2$.
The governing circuit formulas physics equation for the differential output voltage ($V_{out} = V_B - V_D$) is derived by calculating the voltage at each mid-node using the voltage divider rule, then subtracting them:
$V_{out} = V_{ex} \left( \frac{R_4}{R_3 + R_4} - \frac{R_2}{R_1 + R_2} \right)$
When the bridge is perfectly balanced ($\frac{R_1}{R_2} = \frac{R_3}{R_4}$), $V_{out}$ is exactly 0V. Any physical change in the sensor ($R_2$) unbalances the bridge, creating a differential voltage proportional to the resistance delta. According to HyperPhysics at Georgia State University, this null-measurement principle is what makes the bridge immune to fluctuations in the excitation voltage, provided the ADC reads the ratio rather than absolute voltage.
Design Walkthrough: Selecting Real Component Values
Let’s design a bridge to measure room temperature (15°C to 35°C) using a 10kΩ NTC thermistor. We need to pick components that minimize self-heating while providing a measurable voltage swing.
| Designator | Role | Real Part Number / Spec | Value | Tolerance / Tempco |
|---|---|---|---|---|
| $R_1, R_3, R_4$ | Fixed Reference Legs | Vishay Dale CMF5510K000FHEB | 10.0 kΩ | 1%, ±50 ppm/°C |
| $R_2$ | Sensor (Active Element) | Vishay NTCLE100E3103 | 10.0 kΩ @ 25°C | Beta = 3977K, ±2% |
| $V_{ex}$ | Excitation Source | LM7805 or Bench Supply | 5.00 V DC | Line reg: 5mV |
At 25°C, $R_2$ is exactly 10kΩ. The bridge is balanced, and $V_{out}$ is 0mV. If the temperature drops to 15°C, the NTC resistance rises to ~15.7kΩ. Plugging this into our physics formula yields a $V_{out}$ of roughly -330mV. If the temperature rises to 35°C, $R_2$ drops to ~6.5kΩ, yielding a $V_{out}$ of +380mV. This gives us a highly linear-ish swing of ~710mV across a 20°C range, easily readable by a standard 12-bit ADC.
Behavior and Failure Modes: What Happens When Things Break
Understanding how a circuit behaves at its extremes is where textbook theory meets jobsite reality. If a solder joint cracks (open) or a component fails short, the bridge output will slam to a rail. Here is the failure-mode contrast you need to program your microcontroller's fault-detection logic.
| Element Changed | Direction of Change | Effect on $V_{out}$ ($V_B - V_D$) | Physical Meaning / Fault Condition |
|---|---|---|---|
| $R_2$ (Sensor) | Resistance Increases | Swings Negative | Temperature dropping (normal operation). |
| $R_2$ (Sensor) | Resistance Decreases | Swings Positive | Temperature rising (normal operation). |
| $R_1$ (Top Left) | Shorted (0Ω) | Slams to Maximum Positive ($+V_{ex}/2$) | Solder bridge across $R_1$. Node D pulled to $V_{ex}$. |
| $R_1$ (Top Left) | Open (∞Ω) | Slams to Maximum Negative ($-V_{ex}/2$) | Broken trace. Node D pulled to GND via $R_2$. |
| $R_4$ (Bottom Right) | Shorted (0Ω) | Slams to Maximum Negative ($-V_{ex}/2$) | Solder bridge across $R_4$. Node B pulled to GND. |
| $R_3$ (Top Right) | Open (∞Ω) | Slams to Maximum Negative ($-V_{ex}/2$) | Broken trace. Node B pulled to GND via $R_4$. |
Notice that an open $R_1$ and a shorted $R_4$ produce the exact same output voltage. This is a classic troubleshooting trap. If your ADC reads -2.5V (assuming a 5V excitation), you cannot tell if the top-left resistor broke or the bottom-right resistor shorted without taking a multimeter to the physical nodes. Always design your firmware to flag any $V_{out}$ beyond the expected physical range (e.g., > ±450mV for our thermistor) as a "Sensor Fault" rather than an extreme temperature reading.
Breadboard Testing: Step-by-Step Verification
Before committing this topology to a PCB, validate the physics on a solderless breadboard. Parasitic contact resistance in breadboards (often 0.1Ω to 0.5Ω per contact) won't ruin a 10kΩ bridge, but it will destroy a 120Ω strain gauge bridge. For our 10kΩ NTC design, a standard 830-point breadboard is perfectly adequate.
- Seat the Fixed Resistors: Insert the three Vishay CMF55 10kΩ resistors ($R_1, R_3, R_4$). Place $R_1$ and $R_2$ (the thermistor) in the left column, sharing Node D. Place $R_3$ and $R_4$ in the right column, sharing Node B.
- Route Excitation and Ground: Connect the top rails to your 5.0V bench supply. Connect Node A (top of $R_1$ and $R_3$) to the positive rail. Connect Node C (bottom of $R_2$ and $R_4$) to the ground rail.
- Verify the Null Point: Set your multimeter (e.g., Fluke 87V) to DC millivolts. Probe Node B with the red lead and Node D with the black lead. At exactly 25°C, you should read between -15mV and +15mV. (The offset is due to the 1% tolerance stack-up of the fixed resistors; a true 0.00mV requires 0.1% precision resistors or a trimming potentiometer).
- Apply Thermal Stimulus: Pinch the thermistor bead between your fingers. The body heat (approx 32°C) will drop the NTC resistance. Watch the multimeter; $V_{out}$ should smoothly swing positive, crossing +200mV within a few seconds.
- Simulate a Fault: Pull $R_1$ out of the breadboard to simulate an open circuit. The meter should immediately snap to roughly -2.5V. Re-insert $R_1$ and verify the voltage returns to the ambient baseline.
Why a Bridge Over a Simple Voltage Divider?
If you only have one active sensor, why not just put it in a single voltage divider ($V_{out} = V_{ex} \times \frac{R_{sensor}}{R_{fixed} + R_{sensor}}$) and feed it straight to an Arduino analog pin?
The bridge wins on three critical fronts, heavily detailed in Texas Instruments' application notes on bridge sensors:
- Common-Mode Rejection: If your 5V supply sags to 4.8V due to a motor turning on elsewhere in the system, a single divider's output sags proportionally, causing a false temperature reading. In a balanced bridge, both legs sag equally. The differential voltage ($V_B - V_D$) remains virtually unchanged, rejecting the power supply noise.
- Zero-Point Calibration: A single divider outputs ~2.5V at room temperature. You are trying to measure a 10mV change riding on top of a 2.5V common-mode voltage, which wastes the dynamic range of your ADC. The bridge outputs 0V at room temperature, allowing you to amplify the ±300mV delta by 10x using an instrumentation amplifier (like the INA128) to perfectly span a 0-5V ADC range.
- Lead Wire Cancellation: In remote sensing (e.g., the thermistor is 3 meters away), the copper wire adds resistance. In a 3-wire or 4-wire bridge configuration, the lead resistances are distributed symmetrically into adjacent bridge arms, mathematically canceling their effect out of the final physics equation.
When you are designing sensor front-ends, relying on basic circuit formulas physics isn't just an academic exercise. It is the exact mechanism that separates a noisy, drifting prototype from a robust, calibration-stable instrument. Pick your resistors carefully, map your failure modes in firmware, and always verify the null point on the bench before writing your final ADC conversion code.
