The Wheatstone bridge is the gold standard direct current circuit topology for precision resistance measurement and sensor interfacing. While a simple series voltage divider can technically convert a resistance change into a voltage, it suffers from severe common-mode drift and poor resolution at the extremes. A balanced bridge outputs exactly 0V at its null point, and an unbalanced bridge outputs a proportional millivolt differential signal that rejects power supply noise. This guide walks through the exact node architecture, component selection, failure modes, and breadboard verification for a thermistor-based bridge.
Topology Architecture and Node Labels
A standard Wheatstone bridge direct current circuit consists of two parallel voltage divider legs driven by a common excitation voltage. To analyze it on the bench, we assign four distinct nodes:
- Node A (Excitation+): The top rail where the positive DC reference voltage is applied.
- Node B (Sense+): The midpoint of the left leg (between R1 and R2). This is the positive differential output.
- Node C (Excitation- / GND): The bottom rail, tied to circuit ground.
- Node D (Sense-): The midpoint of the right leg (between R3 and R4). This is the negative differential output.
The left leg is typically the 'reference' leg, built with fixed, high-precision resistors. The right leg is the 'sense' leg, where R4 is replaced by a variable resistive sensor (like a strain gauge, RTD, or thermistor). The measurable output is the differential voltage: V_out = V_B - V_D.
Component Selection and Design Walkthrough
Let us design a direct current circuit to measure ambient temperature using a 10k NTC thermistor. We will balance the bridge at exactly 25°C. At 25°C, the thermistor reads 10,000Ω. To keep self-heating errors below 0.1°C, we must limit the current through the thermistor to less than 100µA. Using a 5.00V excitation source, the total leg resistance must be at least 50kΩ. However, standard 10k thermistors dictate our baseline. We will accept a slight self-heating trade-off for better signal-to-noise ratio, or we can use a lower excitation voltage. For this bench design, we will use a precision 2.500V shunt reference to keep power dissipation in the 10k NTC below 0.6mW.
| Designator | Component / Part Number | Value / Spec | Tolerance | Power Rating |
|---|---|---|---|---|
| V_ex | TI LM4040AIM5-2.5 | 2.500V Shunt Reference | 0.1% Initial | N/A |
| R1 | Vishay MRS25000C1002 | 10.0kΩ Metal Film | 1% | 0.6W |
| R2 | Vishay MRS25000C1002 | 10.0kΩ Metal Film | 1% | 0.6W |
| R3 | Vishay MRS25000C1002 | 10.0kΩ Metal Film | 1% | 0.6W |
| R4 | Murata NCP18XH103F03RB | 10kΩ NTC (Beta 3950) | 1% @ 25°C | 100mW Max |
With these values, at exactly 25°C, R1=R2=R3=R4=10kΩ. Node B sits at exactly 1.250V. Node D sits at exactly 1.250V. The differential output V_B - V_D is 0.000mV. As the temperature rises, the NTC resistance drops, pulling Node D lower and creating a positive differential voltage.
Behavioral Analysis: What Changes When Elements Shift
Understanding how the bridge reacts to component drift is critical for selecting an instrumentation amplifier downstream. The table below maps the exact node behavior when the thermistor (R4) shifts due to temperature changes, assuming a perfect 2.500V excitation.
| Condition | R4 Value | Node B (Sense+) | Node D (Sense-) | V_out (V_B - V_D) |
|---|---|---|---|---|
| Baseline (25°C) | 10.00 kΩ | 1.2500 V | 1.2500 V | 0.000 mV |
| Temp Rises (35°C) | 6.53 kΩ | 1.2500 V | 0.9874 V | +262.6 mV |
| Temp Drops (15°C) | 15.76 kΩ | 1.2500 V | 1.5241 V | -274.1 mV |
| Extreme Heat (85°C) | 1.15 kΩ | 1.2500 V | 0.2575 V | +992.5 mV |
Notice that Node B remains completely static. Because the left leg is isolated from the right leg, changes in R4 do not load down Node B. This is a massive advantage over single-ended topologies where the sensor itself forms the bottom half of a single divider, causing the source impedance to vary wildly and destabilizing the ADC sampling capacitor.
Failure Modes and Topology Justification
Why choose a four-resistor bridge direct current circuit over a simple two-resistor series voltage divider? The answer is Common-Mode Rejection Ratio (CMRR). If you use a simple divider (10k fixed, 10k NTC) powered directly from a 5V USB rail, and that USB rail sags to 4.8V under load, your output voltage shifts dramatically, faking a temperature reading. In a bridge, if V_ex sags from 2.500V to 2.400V, both Node B and Node D drop by the exact same percentage. The differential voltage (V_B - V_D) remains virtually unchanged. The bridge rejects power supply noise inherently.
When building this direct current circuit, you must anticipate hard failures. Here is what breaks at the extremes:
| Failure Event | Node B Impact | Node D Impact | System Result |
|---|---|---|---|
| R1 Opens | Drops to 0V (pulled by R2) | Unchanged | Massive negative V_out; reads as extreme cold. |
| R2 Shorts | Clamps to GND (0V) | Unchanged | Massive negative V_out; reads as extreme cold. |
| R3 Opens | Unchanged | Drops to 0V (pulled by R4) | Massive positive V_out; reads as extreme heat. |
| R4 Shorts | Unchanged | Clamps to GND (0V) | V_out equals V_B (1.25V); ADC saturates high. |
According to All About Circuits, the bridge's sensitivity is highest when all four arms are roughly equal in resistance. If you must use a 100Ω PT100 RTD instead of a 10k NTC, you must scale R1, R2, and R3 down to 100Ω as well, which drastically increases current draw and requires careful thermal management to prevent the fixed resistors from heating the board and skewing the RTD.
Step-by-Step Breadboard Verification
Before committing this direct current circuit to a PCB, verify the null point and differential swing on a solderless breadboard. Parasitic contact resistance in breadboards (often 0.1Ω to 0.5Ω per spring clip) can introduce microvolt errors, but for a millivolt-scale thermistor bridge, it is acceptable for initial validation.
- Establish the Excitation Rail: Insert the TI LM4040AIM5-2.5 shunt regulator. Connect its cathode to your bench supply (set to 5.0V) via a 1kΩ current-limiting resistor. Connect its anode to the breadboard ground rail. Verify the cathode reads exactly 2.500V ±2.5mV with a 4.5-digit multimeter.
- Build the Reference Leg (Left): Insert R1 (10kΩ) from the 2.5V rail to Node B. Insert R2 (10kΩ) from Node B to the ground rail. Measure Node B. It must read 1.250V. If it reads 1.256V, your resistors are mismatched by 1%; swap R2 until you are within 2mV of 1.250V.
- Build the Sense Leg (Right): Insert R3 (10kΩ) from the 2.5V rail to Node D. Insert the Murata 10k NTC thermistor from Node D to ground. Ensure the thermistor leads are fully seated to avoid intermittent contact resistance.
- Measure the Null: Set your multimeter to the millivolt DC range. Place the red probe on Node B and the black probe on Node D. At a standard 22°C room temperature, the NTC will read roughly 11.2kΩ, yielding a slight negative differential (approx -35mV). Touch the thermistor bead with your finger; the voltage should immediately swing positive as your body heat drops the NTC resistance.
- Inject a Precision Shunt (Optional): To simulate a specific temperature without a thermal chamber, remove the NTC and insert a precision 10.00kΩ decade box or a matched 1% metal film resistor. The differential voltage should collapse to within ±5mV of zero, confirming the bridge topology is balanced and the wiring introduces no significant asymmetry.
For further reading on interfacing this analog front-end with modern ADCs, review the bridge sensor conditioning guidelines published by Analog Devices, which detail how to map this differential millivolt output into the input range of a 24-bit sigma-delta ADC without losing the common-mode rejection benefits of the bridge topology.
