If you are asking what is an op amp, the direct answer is this: an operational amplifier is a high-gain, differential DC-coupled voltage amplifier with a single-ended output. In practical bench terms, it measures the voltage difference between its two input pins, multiplies that difference by a massive internal open-loop gain (typically 100,000 to 1,000,000), and drives its output pin to match the result—constrained only by its power supply rails. We use external feedback resistors to tame this massive gain into precise, predictable circuits like amplifiers, filters, and comparators.
The Anatomy: Symbol, Pinout, and Internal Stages
On a schematic, an op amp is drawn as a triangle pointing to the right. The top input is marked with a minus sign (inverting input, V-), and the bottom input is marked with a plus sign (non-inverting input, V+). The point of the triangle is the output (Vout). Two additional pins, often omitted in simplified schematics but mandatory on the bench, are the positive power supply (VCC or V+) and negative power supply/ground (VEE or V-).
Most hobbyist and prototype op amps come in an 8-pin DIP (Dual In-line Package) containing two independent amplifiers (a 'dual' op amp). Here is the industry-standard pinout for a dual op amp like the LM358 or TL072:
- Pin 1: Output A
- Pin 2: Inverting Input A (-)
- Pin 3: Non-Inverting Input A (+)
- Pin 4: V- (Ground or Negative Rail)
- Pin 5: Non-Inverting Input B (+)
- Pin 6: Inverting Input B (-)
- Pin 7: Output B
- Pin 8: V+ (Positive Rail)
Selecting and Biasing: Safe Defaults and Operating Regions
Choosing the right op amp depends entirely on your power supply and signal type. Biasing an op amp means setting its DC operating point so the AC signal has room to swing without hitting the power rails. If you use a dual supply (e.g., +9V and -9V), the inputs and outputs naturally sit at 0V. If you use a single supply (e.g., +5V and GND), you must create a 'virtual ground' at half the supply voltage (2.5V) using a resistor voltage divider, biasing the non-inverting input to this midpoint.
Stop guessing and stock these three. They cover 95% of DIY and prototype needs:
1. LM358 (~$0.15): The single-supply workhorse. Inputs can swing down to ground, but the output cannot reach the positive rail.
2. TL072 (~$0.60): The audio and dual-supply standard. JFET inputs offer low noise and high slew rate, but it requires at least ±5V dual rails.
3. MCP6002 (~$0.40): The modern 3.3V/5V microcontroller interface. True Rail-to-Rail Input/Output (RRIO), perfect for ESP32 and Arduino ADC buffering.
Understanding how the op amp behaves across different voltage thresholds is critical for debugging. Here is how the internal stages react in the real world:
| Operation Region | Condition | Typical Bench Behavior | Output State |
|---|---|---|---|
| Linear (Active) | V+ ≈ V- (via feedback) | Output precisely follows the gain equation. Inputs draw nanoamps to picoamps. | Between V- rail and V+ rail |
| Positive Saturation | V+ > V- (open loop) | Internal PNP/NPN transistors fully saturate. Output hits the positive ceiling. | VCC - 1.5V (Standard) or VCC - 0.05V (RRIO) |
| Negative Saturation | V+ < V- (open loop) | Output stage pulls hard to the negative rail. Acts as a comparator. | VEE + 0.2V (Standard) or VEE + 0.01V (RRIO) |
| Slew Rate Limiting | Fast transient input | Internal compensation capacitor limits charge rate. Output ramps linearly. | Distorted/Triangular waveform |
Application Circuit: Building a Non-Inverting Amplifier
Let's build a circuit to amplify a 0.5V analog sensor signal to 2.5V so it can be read safely by an ESP32 ADC (which maxes out around 2.5V on its ADC pins). We need a gain of exactly 5. We will use an MCP6002 on a single 3.3V supply.
The gain formula for a non-inverting amplifier is: Gain = 1 + (Rf / Ri). To get a gain of 5, we need Rf to be 4 times larger than Ri. We will use standard 1% tolerance resistors: Rf = 40.2kΩ and Ri = 10.0kΩ.
- Power the Chip: Connect Pin 8 (V+) to your 3.3V rail and Pin 4 (V-) to GND. Place a 100nF ceramic decoupling capacitor directly across Pin 8 and Pin 4 to prevent high-frequency oscillation.
- Wire the Input: Connect your 0.5V sensor signal to Pin 3 (Non-inverting input).
- Set the Feedback Network: Connect the 10.0kΩ resistor (Ri) from Pin 2 (Inverting input) to GND.
- Close the Loop: Connect the 40.2kΩ resistor (Rf) from Pin 2 (Inverting input) to Pin 1 (Output). This negative feedback forces the op amp to adjust its output until Pin 2 matches the voltage at Pin 3.
- Verify: Measure Pin 1 with your multimeter. With 0.5V at Pin 3, Pin 1 should read exactly 2.51V.
Bench War Story: When the Output Rails and How to Fix It
Theory is clean; the bench is messy. Here is a classic trap that burns hobbyists trying to interface 5V sensors with op amps.
The Setup: I needed to buffer a 0V to 4.8V analog signal from a pressure transducer before feeding it to a microcontroller. I grabbed an LM358 because it's cheap and handles single supplies. I powered it with a single 5.0V USB rail, wired it as a unity-gain buffer (output tied directly to the inverting input), and fed the 4.8V signal into the non-inverting input.
The Numbers: Expected output: 4.8V. Measured output with the DMM: 3.6V.
What Went Wrong: The LM358 is famous for its ability to swing its output all the way down to the negative rail (0V). However, it is not a rail-to-rail output device on the high side. The datasheet specifies the High-Level Output Voltage (VOH) as typically VCC - 1.5V. With a 5.0V supply, the absolute maximum output the LM358 can physically produce is 3.5V to 3.6V. The op amp was hitting positive saturation and clipping the top off my signal.
The Fix: I had two choices. Option A: Swap the LM358 for an MCP6002, which features a true rail-to-rail output stage that can swing within 20mV of the 5V rail. Option B: Keep the LM358 but power it from a 9V battery, giving the output stage enough headroom to reach 4.8V (since 9V - 1.5V = 7.5V max). I chose Option A to keep the BOM simple.
Troubleshooting: How Op Amps Fail and How to Test Them
Op amps rarely die of old age; they die from abuse. The most common failure modes are exceeding the common-mode input voltage range (which blows the input differential pair), shorting the output to a voltage rail (which melts the internal output transistors), and ESD strikes during breadboarding.
If your circuit is misbehaving, do not just rip the chip out. Use your multimeter to diagnose it in-circuit.
The Power-Off Diode Test
Disconnect power from the breadboard. Set your DMM to Diode Test mode. This tests the internal ESD protection diodes that connect the input pins to the power rails.
- Place the red probe on the inverting input (Pin 2) and the black probe on V- (Pin 4). You should read a forward voltage drop of roughly 0.6V to 0.8V.
- Reverse the probes (black on input, red on V-). It should read 'OL' (Open Loop).
- If it reads 0.00V, the input protection diode is shorted—the chip is dead. If it reads 'OL' in both directions, the internal bond wire is blown.
The Power-On DC Bias Check
If the diode test passes, power the circuit and switch your DMM to DC Voltage.
- Check the Rails: Measure Pin 8 to Pin 4. Ensure you have clean, stable power. If it's noisy, your decoupling capacitor is missing or too far away.
- Check the Virtual Ground: If using a single supply with a bias network, measure the non-inverting input. It must sit exactly at your bias voltage (e.g., 2.5V). If it's floating or at 0V, your bias resistor divider is broken.
- Check the Differential: Measure the voltage between Pin 3 and Pin 2. In a properly functioning linear circuit with negative feedback, this voltage must be virtually zero (typically < 1mV). If you measure a significant voltage difference (e.g., 0.5V) between the inputs while the op amp is powered, the internal open-loop gain has collapsed. The chip is fried.
For deeper study on input stage architectures and how they affect your DC errors, refer to the Analog Devices MT-035 Tutorial on Op Amp Inputs. Additionally, the Texas Instruments Precision Labs Op Amp video series provides exceptional visual breakdowns of slew rate and bandwidth limitations that text alone cannot convey.
