In short: An inverting op-amp forces its minus input to sit at a "virtual ground," which pins the gain to the pure resistor ratio −Rf/Rin, inverts the signal by 180°, and makes the input impedance equal to Rin.
The inverting amplifier is the first circuit most analog engineers draw from memory, and for good reason: it distills the entire feedback-amplifier mindset into three components and one powerful idea. Master it and the summing amplifier, the current-to-voltage converter, the integrator, and the active filter all fall out as variations. This article derives the gain rigorously from the virtual-ground concept, then confronts the two trade-offs interviewers love to probe: the 180° phase inversion and the finite, resistor-set input impedance.
The circuit and the two golden rules
Picture an ideal op-amp. The non-inverting input (+) is tied to ground. The signal source drives the inverting input (−) through a resistor Rin, and a feedback resistor Rf connects the output back to that same − node. Two idealizations carry the whole analysis:
- Rule 1 — no input current. An ideal op-amp has infinite input impedance, so zero current flows into either input terminal.
- Rule 2 — the inputs track. With negative feedback and infinite open-loop gain, the op-amp drives its output until the voltage difference between its inputs is essentially zero (V+ = V−).
Why the minus node is a "virtual ground"
Because V+ is physically tied to 0 V, Rule 2 forces V− to 0 V as well. The − node behaves as if it were grounded — it holds 0 V — yet no wire connects it to ground and no current is dumped into a real ground there. That is the meaning of "virtual ground." It is one of the most useful fictions in analog design: it lets you treat the summing node as a fixed-potential anchor without it actually sinking current to ground.
Key fact: The virtual ground exists only while the loop is closed and the output is unsaturated. Drive the output into the rail, and V− is no longer held at 0 V — a common bench surprise when a gain-of-100 stage clips on a 50 mV offset.
Deriving the gain
Apply Kirchhoff's current law at the virtual-ground node. By Rule 1 no current enters the op-amp, so the current through Rin must continue entirely through Rf:
Current in through Rin: Iin = (Vin − 0) / Rin = Vin / Rin
Current out through Rf: If = (0 − Vout) / Rf = −Vout / Rf
Setting Iin = If and solving gives the closed-loop gain:
Av = Vout / Vin = −Rf / Rin
The minus sign: 180° phase inversion
That negative sign is not cosmetic. A positive-going input produces a negative-going output: a sine wave emerges exactly 180° out of phase. For DC and low frequencies this is a clean, flat 180°. The design consequence is subtle but real — the inversion is part of the feedback loop's phase budget. Above the op-amp's dominant pole, the amplifier adds its own 90° lag on top of the ideal 180°, which is why stability analysis (phase margin) always treats the inverting and non-inverting configurations with the same loop-gain Bode plot even though their signal phases differ.
Input impedance: the price of the virtual ground
Here is the trade-off. Because the source drives into a node pinned at 0 V, the impedance the source sees is simply Rin. The op-amp's own enormous input impedance is hidden behind the virtual ground and never helps the source.
Key fact: Zin = Rin. A gain of −100 built as Rf = 1 MΩ / Rin = 10 kΩ loads the source with only 10 kΩ — fine for a buffered driver, a problem for a 1 MΩ sensor.
This is the single biggest practical difference from the non-inverting amplifier, whose input goes straight to the high-impedance + terminal and presents megohms to the source. Choosing between the two topologies is often an impedance decision, not a gain decision.
| Rin | Rf | Gain (V/V) | Zin |
|---|---|---|---|
| 10 kΩ | 10 kΩ | −1 (unity invert) | 10 kΩ |
| 10 kΩ | 100 kΩ | −10 | 10 kΩ |
| 1 kΩ | 47 kΩ | −47 | 1 kΩ |
Where the ideal model breaks
Real parts add constraints worth knowing for both the bench and the interview. Finite open-loop gain AOL means the virtual ground is really at Vout/AOL, not exactly 0 V; with AOL of 105 this error is microvolts. Gain–bandwidth product sets the usable frequency: a 1 MHz GBW part at gain −100 gives only ~10 kHz of bandwidth. Input bias current flowing through Rf creates an offset, which is why a bias-compensation resistor Rin∥Rf often sits between the + input and ground. When you simulate this stage in Cadence Virtuoso (Spectre), Synopsys HSPICE, or Siemens EDA's Eldo/AFS, an .ac sweep exposes the GBW roll-off and an .op point confirms the virtual-ground voltage is a few microvolts, not a clean zero.
For a hands-on build and PCB-level design review, the engineers at Avecas analog design services walk through these second-order effects on real silicon.
Want to derive these circuits yourself, then verify them in a real SPICE deck? The CourseTron Analog Design track takes you from the virtual-ground intuition to a simulated, stability-checked amplifier.
Related building blocks
Once the inverting amplifier clicks, extend it: tie several input resistors to the same virtual ground and you have a summing amplifier; swap Rf for a capacitor and you have an integrator. And always compare against the non-inverting topology when input impedance matters.
Frequently asked questions
Why is it called a "virtual" ground and not a real one?
The node sits at 0 V like a ground, but no conductor connects it to the ground rail and it does not sink current to ground. Feedback actively holds the potential there; remove the loop and the "ground" disappears.
Does the inverting amplifier really present low input impedance?
Yes. The source sees Rin because the far end of Rin is pinned at 0 V. The op-amp's high input impedance is behind that node and does not benefit the source.
Can I get a gain of exactly −1?
Set Rf = Rin. This "unity-gain inverter" is common for generating a differential pair or flipping a signal's polarity before a summing stage.
How does the phase inversion affect stability?
The ideal 180° signal inversion is separate from the loop-gain phase. Stability depends on how much extra phase the op-amp's poles add to the loop; you still check phase margin on the open-loop Bode plot exactly as for a non-inverting stage.