The Core Gain Formula for Differential Amplifier Circuits
A standard op-amp differential amplifier (often called a difference amplifier) subtracts two input voltages and amplifies the difference. The most common topology uses a single operational amplifier and four resistors. For the circuit to function as a true differential amplifier with high common-mode rejection, the resistor ratios must be perfectly matched: $R_1 = R_3$ and $R_2 = R_4$.
The foundational gain formula for differential amplifier circuits, assuming ideal resistor matching and an ideal op-amp, is:
$A_d = \frac{R_2}{R_1}$
$V_{out} = A_d \times (V_2 - V_1)$
Every symbol in these equations carries strict physical and electrical boundaries. Below is the definitive reference table for this topology.
| Symbol | Parameter | Standard Unit | Definition & Practical Constraints |
|---|---|---|---|
| $A_d$ | Differential Gain | V/V (Dimensionless) | The ratio of output voltage change to differential input voltage change. Realistically constrained between 1 and 50 for discrete builds. |
| $R_1, R_3$ | Input Resistors | Ohms ($\Omega$) | Resistors connecting the input signals to the op-amp terminals. They set the circuit's input impedance ($Z_{in} \approx R_1 + R_3$). |
| $R_2, R_4$ | Feedback & Ground Resistors | Ohms ($\Omega$) | $R_2$ is the negative feedback resistor; $R_4$ ties the non-inverting node to ground. Must match $R_1$ and $R_3$ ratios exactly. |
| $V_{out}$ | Output Voltage | Volts (V) | The amplified difference. Constrained by the op-amp's supply rails (e.g., cannot exceed $V_{CC} - 1.5V$ on a standard LM358). |
| $V_1$ | Inverting Input Voltage | Volts (V) | Signal applied to the inverting (-) terminal via $R_1$. |
| $V_2$ | Non-Inverting Input Voltage | Volts (V) | Signal applied to the non-inverting (+) terminal via $R_3$. |
When the Formula Applies (and Its Assumptions)
This formula assumes an ideal operational amplifier: infinite open-loop gain, infinite input impedance, and zero output impedance. In reality, op-amps like the TL072 or LM358 have finite open-loop gains (typically 100 dB to 120 dB). The formula holds true for closed-loop gains up to about 100 V/V. Beyond that, the finite open-loop gain introduces significant calculation errors.
Furthermore, the formula assumes perfect resistor matching. If $R_2/R_1 \neq R_4/R_3$, the circuit will amplify the common-mode voltage (noise present on both inputs), destroying the Common-Mode Rejection Ratio (CMRR). For precision DC measurements, 0.1% thin-film resistors are mandatory; standard 5% carbon film resistors will render the differential gain formula practically useless in noisy environments.
Real-World Resistor Selection and Gain Magnitudes
What does a realistic answer magnitude look like? While the math allows you to plug in $R_2 = 1\text{ M}\Omega$ and $R_1 = 1\text{ k}\Omega$ for a gain of 1000 V/V, doing so in a single-stage discrete differential amplifier is a design trap. High resistor ratios severely degrade CMRR and make the circuit hyper-sensitive to parasitic capacitance, killing your high-frequency bandwidth.
According to Analog Devices MT-044 tutorial guidelines, keeping the gain between 1 and 50 V/V preserves stability and CMRR. If you need a gain of 500, you should cascade two stages (e.g., $10 \times 50$) or switch to a dedicated instrumentation amplifier IC like the TI INA128.
The table below maps realistic resistor pairs to their resulting differential gain and the expected CMRR based on resistor tolerance. This data is critical for bench design.
| $R_1$ & $R_3$ (Input) | $R_2$ & $R_4$ (Feedback) | Gain ($A_d$) | CMRR (1% Resistors) | CMRR (0.1% Resistors) | Typical Application |
|---|---|---|---|---|---|
| 10.0 k$\Omega$ | 10.0 k$\Omega$ | 1 V/V | ~54 dB | ~74 dB | Level shifting, unity-gain buffers |
| 10.0 k$\Omega$ | 49.9 k$\Omega$ | ~5 V/V | ~40 dB | ~60 dB | High-side current shunt sensing |
| 4.70 k$\Omega$ | 100 k$\Omega$ | ~21 V/V | ~30 dB | ~50 dB | Thermocouple amplification |
| 1.00 k$\Omega$ | 100 k$\Omega$ | 100 V/V | ~20 dB (Poor) | ~40 dB (Marginal) | Strain gauges (Use In-Amp instead) |
Note: As gain increases via resistor ratio, the CMRR achievable with discrete resistors plummets. This is why Texas Instruments integrates laser-trimmed resistor networks inside their difference amplifier ICs to guarantee >80 dB CMRR at higher gains.
Rearranged Forms and Unit Traps
On the bench, you rarely know all variables upfront. You usually have a target output voltage for an ADC and a known sensor output, meaning you need to solve for the resistors. Here are the rearranged forms of the gain formula for differential amplifier design:
- Solving for Feedback Resistor ($R_2$): $R_2 = A_d \times R_1$
- Solving for Input Resistor ($R_1$): $R_1 = \frac{R_2}{A_d}$
- Solving for Required Input Differential ($\Delta V_{in}$): $\Delta V_{in} = \frac{V_{out}}{A_d}$
- Solving for Output Voltage ($V_{out}$): $V_{out} = \left(\frac{R_2}{R_1}\right) \times (V_2 - V_1)$
Which Unit Mistakes Break the Math?
When calculating by hand or writing a quick Python script to validate bench measurements, three specific unit traps will yield catastrophic design errors:
- The Millivolt/Volt Mixing Trap: Sensor outputs (like shunt resistors) are often in millivolts (mV), while ADC references are in volts (V). If $V_2 - V_1 = 40\text{ mV}$ and your target $V_{out}$ is $2.0\text{ V}$, calculating $A_d = 2.0 / 40 = 0.05$ is wrong. You must convert to base units first: $A_d = 2.0\text{ V} / 0.040\text{ V} = 50\text{ V/V}$.
- The Dimensionless Gain Trap: Gain ($A_d$) is a ratio (V/V). It does not have units of Ohms. If you calculate $R_2 = 10 \times 4.7\text{ k}\Omega$, the gain is $10$, not $10\text{ }\Omega$. The resulting $R_2$ is $47\text{ k}\Omega$.
- The Rail Clipping Oversight: The formula assumes $V_{out}$ can swing infinitely. If you calculate $V_{out} = 4.5\text{ V}$ but your op-amp is powered by a single 5V supply, a standard LM358 will saturate around 3.5V. You must verify your calculated $V_{out}$ against the op-amp's datasheet output swing limits.
Worked Examples: From Sensor Millivolts to ADC Voltages
Let's apply the formula to two common real-world scenarios, tracking every unit to ensure precision.
Problem 1: Low-Side Current Shunt Sensing
Scenario: You are measuring the current draw of a 12V DC motor using a $0.1\text{ }\Omega$ shunt resistor on the low side (ground path). At peak load, the motor draws $750\text{ mA}$. You want to feed this signal into a 3.3V microcontroller ADC, targeting a maximum readout of $1.65\text{ V}$ (mid-scale) to allow for transient spikes. You have chosen $R_1 = R_3 = 10\text{ k}\Omega$ to keep input impedance reasonable. Find the required $R_2$ and $R_4$.
Step 1: Calculate the differential input voltage ($\Delta V_{in}$).
Using Ohm's Law ($V = I \times R$):
$\Delta V_{in} = 750\text{ mA} \times 0.1\text{ }\Omega$
$\Delta V_{in} = 0.750\text{ A} \times 0.1\text{ }\Omega = 0.075\text{ V}$ (or $75\text{ mV}$)
Step 2: Determine the required Differential Gain ($A_d$).
Using the rearranged formula $A_d = V_{out} / \Delta V_{in}$:
$A_d = 1.65\text{ V} / 0.075\text{ V}$
$A_d = 22\text{ V/V}$
Step 3: Solve for the Feedback Resistors ($R_2$ and $R_4$).
Using $R_2 = A_d \times R_1$:
$R_2 = 22 \times 10\text{ k}\Omega$
$R_2 = 220\text{ k}\Omega$
Verification: A gain of 22 V/V is well within the realistic magnitude limit (under 50). Using standard 1% E96 series resistors, $220\text{ k}\Omega$ and $10.0\text{ k}\Omega$ are exact, off-the-shelf values. The CMRR will be approximately 38 dB, which is acceptable for low-side sensing where common-mode voltage is near 0V.
Problem 2: Wheatstone Bridge Strain Gauge (The High-Gain Trap)
Scenario: A load cell outputs a differential voltage of $2.4\text{ mV}$ at full scale. You need to amplify this to $2.4\text{ V}$ for a data acquisition system. You decide to use a standard TL072 op-amp and attempt to build a single-stage differential amplifier. You select $R_1 = R_3 = 1\text{ k}\Omega$ to minimize Johnson-Nyquist thermal noise. Calculate $R_2$ and evaluate if this design will actually work on the bench.
Step 1: Convert units and find Gain ($A_d$).
$\Delta V_{in} = 2.4\text{ mV} = 0.0024\text{ V}$
$V_{out} = 2.4\text{ V}$
$A_d = 2.4\text{ V} / 0.0024\text{ V} = 1000\text{ V/V}$
Step 2: Solve for $R_2$.
$R_2 = A_d \times R_1$
$R_2 = 1000 \times 1\text{ k}\Omega = 1,000,000\text{ }\Omega$ ($1\text{ M}\Omega$)
Step 3: Bench Evaluation (The Reality Check).
While the math yields $R_2 = 1\text{ M}\Omega$, this design will fail in practice. As noted in the All About Circuits semiconductor textbook, a 1000:1 resistor ratio in a discrete difference amplifier guarantees terrible CMRR (likely below 15 dB). Any 50/60Hz mains hum common to both bridge wires will be amplified and completely drown out the $2.4\text{ mV}$ signal. Furthermore, $1\text{ M}\Omega$ feedback resistors generate significant thermal noise and are highly susceptible to stray capacitive coupling.
The Fix: Abandon the single-stage discrete diff amp for this application. Instead, use a dedicated Instrumentation Amplifier like the INA128, which uses a 3-op-amp topology to achieve a gain of 1000 V/V via a single external resistor ($R_G$) while maintaining >100 dB of CMRR and high input impedance.
