🔧 Project

Op-Amp Characterization in LTspice

A full SPICE characterization of an operational amplifier against its datasheet — DC offsets and bias currents, differential and common-mode gain, CMRR and PSRR, unity-gain bandwidth, and step-response slew rate — with every parameter measured, plotted, and compared to spec.

LTspiceAnalogOp-AmpCharacterizationMixed-Signal

Tool

LTspice

CMRR

≈ 98 dB

PSRR

≈ 91 dB

Diff. gain

≈ 198,000 V/V

Unity-gain

≈ 4.6 MHz

Input offset

≈ 1.1 mV

The goal: treat a datasheet as a spec to verify

A datasheet makes a lot of promises — an offset voltage here, a CMRR there, a unity-gain bandwidth on the last page. The point of this project was to take those promises and verify every one of them in simulation, the same way you'd characterize a real part on the bench: build the right test setup for each parameter, measure it, and put the result next to the specified value.

It's a compact version of exactly the mixed-signal characterization mindset I use professionally — decide what "good" means, build a repeatable measurement for it, and let the numbers, not intuition, tell you whether the part meets spec.

Results at a glance

Every parameter, expected vs. simulated, in one table. This is the deliverable the whole project builds toward — a single sheet you could hand to someone and say "here's how this part actually behaves."

Summary table of op-amp parameters: expected vs simulated values for offset, bias, gain, CMRR, PSRR, bandwidth, slew rate. — Full characterization summary — DC, AC and step-response parameters, expected value beside my simulated result.

DC parameters — offsets, bias, and rejection

The DC tests pin down the amplifier's static, real-world imperfections — the things an ideal op-amp wouldn't have:

Parameter	What it measures	My result
Input offset voltage	The small differential voltage needed to drive the output to zero — mismatch in the input stage.	≈ 1.12 mV
Input bias / offset current	The tiny current the inputs actually draw, and the mismatch between them.	≈ 25 pA / 1.9 pA
Differential-mode gain (A_vd)	Open-loop gain on a differential input — how hard the amp amplifies the signal you want.	≈ 198,000 V/V
CMRR	How well it rejects a signal common to both inputs vs. amplifying the difference.	≈ 98 dB
PSRR	How well it rejects noise/ripple riding on its supply rails.	≈ 91 dB

CMRR plot versus frequency from LTspice. — Common-mode rejection ratio — high at DC and rolling off with frequency, as expected.

Differential-mode open-loop gain plot. — Open-loop differential gain — the ≈198,000 V/V (≈106 dB) figure the closed-loop behaviour is built on.

Table of input offset voltage, offset current and bias current. — Input offset voltage, offset current and bias current, extracted from DC sweeps.

Power-supply rejection ratio table. — PSRR measured by perturbing the supply rails and reading the referred output shift.

AC parameters — gain across frequency

An AC sweep turns the amplifier's single open-loop gain number into a curve, and two points on that curve matter most: the 3-dB bandwidth (≈ 43.8 kHz here, where gain has fallen 3 dB) and the unity-gain frequency (≈ 4.6 MHz, where the gain finally crosses 0 dB). Their product is roughly constant — the classic gain-bandwidth trade — so knowing one tells you what closed-loop bandwidth you can buy at a given gain.

Unity-gain frequency plot showing 0 dB crossing. — Unity-gain frequency — the 0 dB crossing at ≈ 4.6 MHz, close to the 5 MHz datasheet figure.

Transient test setup schematic in LTspice. — One of the test benches — each parameter needed its own setup and stimulus.

Step response — slew rate and settling

The transient tests push a step into the amplifier and watch how fast it can actually move. Slew rate is the maximum output rate of change (V/µs) — a large-signal limit that a small-signal bandwidth number won't warn you about — and settling time is how long the output takes to land within a tolerance band of its final value (≈ 178 ns to 0.01% here). Rise time came out at ≈ 41 ns.

Slew-rate measurement from a step response. — Slew-rate measurement — the slope of the output during a large step.

What I took away from it

Building a separate, correct test setup per parameter — DC operating-point sweeps for offsets and rejection, AC for gain/bandwidth, transient for slew and settling.
Reading an amplifier the way a validation engineer does: CMRR, PSRR, offset, gain-bandwidth and slew as the handful of numbers that decide whether it works in a real signal chain.
Comparing measured behaviour against a datasheet spec and being able to explain the deltas (e.g. simulated ≈ 98 dB CMRR vs. an 86 dB typical).
The same discipline — define the metric, build the measurement, compare to spec — that post-silicon characterization runs on, just at the board/SPICE level.