AM vs FM Comparison
Two modules, two complete schemes, and now every number is on the table. Five axes — and the surprise is how cleanly they split. AM takes two. FM takes three. None of the five is close.
Nobody Lost Everything
- Bandwidth — AM
- Noise immunity — FM
- Power efficiency — FM
- Demodulator complexity — AM
- Fidelity — FM
A scheme that lost every axis would have disappeared. Neither did.
FM Spends the Extra 2Δf
AM's bandwidth is exactly twice the message frequency. Carson's rule adds the deviation to it — and note that FM's formula reduces to AM's when Δf goes to zero. That extra 2Δf is what FM spends, and the next axis is what it buys.
Twenty Stations, or One
Broadcast FM: 2(75 + 15) = 180 kHz in a 200 kHz channel. Broadcast AM: 5 kHz of audio in a 10 kHz channel. Compare voice with voice and the gap shrinks — NBFM 16 kHz against AM's 6.8 — but AM is narrower every time.
FM Wins by 18.75 dB
M4-L2 gives the improvement as 3β² — at equal received power, with the noise referred to the 2fm message band. Broadcast FM has β = 75/15 = 5, so 3 × 25 = 75. M2-L4 converts it without a calculator: ×100 is 20 dB, and 75 is three quarters of 100, which costs 1.25 dB. So 20 − 1.25 = 18.75.
Buy Quality With Spectrum
β² Against β
Here is the whole reason FM exists. The advantage grows as β²; the bandwidth grows as roughly β. Double the deviation and you double the spectrum but quadruple the noise advantage — +6 dB per doubling of bandwidth. That is the knob M3-L4 said AM does not have.
Limiter and Capture
The limiter clips the incoming wave to constant height before demodulation and throws amplitude noise away wholesale — the circuit M3-L4 showed has no AM equivalent. And the capture effect means the stronger of two co-channel FM signals suppresses the weaker almost completely, where two AM signals simply add and you hear both.
The Threshold Cliff
Below a carrier-to-noise ratio of roughly 10 dB, individual noise spikes start winning control of the instantaneous frequency and FM collapses — a cliff, not a slope. AM at the same point is noisy but still intelligible. On a genuinely marginal link, AM degrades more gracefully. That is a real argument, not a footnote.
FM Wins This Twice
First: AM peaks at 1/3 and typically manages 4.3%, while FM's constant envelope means the whole waveform is the information. Second, and separately: with no envelope to preserve, FM's amplifier can be run hard in class C at about 70%. AM's must stay linear.
The Axis That Inverted
AM: a diode, a capacitor, a resistor. FM: a limiter plus a discriminator or a PLL. In 1935 that difference decided markets — an FM set needed hand-aligned tuned transformers and a mains supply. Today a PLL is a fraction of a cent of silicon and the hand-aligned transformer is the expensive part.
15 kHz Against 5
- A 200 kHz channel has room for the full audio band; AM's 10 kHz, halved between two sidebands, does not
- The 18.75 dB of Axis 2 is what makes that bandwidth usable rather than hissy
- Pre-emphasis shapes against FM's rising noise — and only a constant envelope can survive the boost
AM keeps one thing here: no cliff. A distant AM station fades gracefully; a distant FM station just stops.
The Scoreboard
- Bandwidth: 10 kHz vs 200 kHz → AM, by 20×
- Noise: 3β² = 75 = 18.75 dB → FM, above threshold
- Power: 1/3 at best vs all of it, and class C → FM, twice
- Complexity: diode+RC vs limiter+PLL → AM on parts, a draw on cost
- Fidelity: 5 kHz vs 15 kHz → FM
Which Would You Choose?
- Bandwidth dominates? SSB, then AM — HF voice, 3 kHz
- Power dominates? FM — a battery buys hours from class C
- Receiver cost dominates? AM — a million receivers per transmitter
- Fidelity dominates? Wide FM — push β and cash in β²
And a fifth question overrides all four: what does failure look like? Aviation chose AM because two overlapping transmissions beat one silently disappearing.
Key Takeaways
- 2fm against 2(Δf + fm) — AM narrower at every message bandwidth
- 3β² = 75 at β = 5, i.e. 18.75 dB, and it grows as β² while bandwidth grows as β
- FM wins power twice: no idle carrier, and a class-C amplifier
- AM's simplicity was decisive in 1935 and is nearly free in silicon now
- Two caveats: FM's threshold cliff, and AM's missing capture as a safety feature
- Choose by the constraint that dominates — then ask what failure looks like
Module 4 Complete!
You have built frequency modulation from the equation to Carson's rule to the discriminator, and now weighed it against AM on every axis that matters. Armstrong's 1933 insight — that bandwidth can be traded for noise immunity — is the most important trade in communications, and it holds far beyond FM.