Question 6 of 6: FM Modulator Output, Carson Bandwidth and Demodulation
Nivaar worked solution (AI-drafted; not reviewed by a licensed engineer)
Notes on this paper
Paper format: Professional Engineers Ontario / EGBC national examination
07-Elec-A3 — Signals and Communications, May 2016. Closed book (a standard
non-programmable calculator is the only aid permitted), 3 hours. Six questions of equal value
(20 marks each); any five constitute a complete paper. All six are solved below, because the
set is intended as a study resource.
Reference texts.
B. P. Lathi and Z. Ding, Modern Digital and Analog Communication Systems, 4th ed.,
Oxford University Press — Ch. 3 (signal transmission), Ch. 4 (amplitude modulation),
Ch. 5 (angle modulation), Ch. 6 (sampling and PCM).
S. Haykin and M. Moher, Communication Systems, 5th ed., Wiley — Ch. 3
(amplitude modulation), Ch. 4 (angle modulation), Ch. 5 (pulse modulation and PCM).
A. V. Oppenheim and A. S. Willsky, Signals and Systems, 2nd ed., Prentice Hall
— Ch. 4 (continuous-time Fourier transform), Ch. 10 (the z-transform).
J. G. Proakis and D. G. Manolakis, Digital Signal Processing, 4th ed., Pearson
— Ch. 3 (z-transform), Ch. 7 (structures for discrete-time systems).
Question 6: FM Modulator Output, Carson Bandwidth and Demodulation (20 marks)
Find. (a) the modulator output $s(t)$ with every parameter named, (b) an
approximate transmission bandwidth, and (c) a non-PLL demodulator.
Approach. Extract the frequency sensitivity from the dc calibration, integrate
the message to obtain the instantaneous phase, then find the true peak of the two-tone message
(which is not the sum of the tone amplitudes) before applying Carson's rule.
Frequency sensitivity from the dc calibration. The modulator is specified by
a constant $k_f$ in Hz per volt such that the instantaneous frequency is
$f_i(t)=f_c+k_f\,m(t)$. A steady 2 V shifts the frequency by 5 kHz, so
$$k_f=\frac{\Delta f_{dc}}{V_{dc}}=\frac{5\ \text{kHz}}{2\ \text{V}}
=\boxed{\;2.5\ \text{kHz/V}\;}$$
Integrate the message to get the phase. An FM signal is
$s(t)=A_c\cos\!\left[2\pi f_ct+2\pi k_f\!\int_0^{t}m(\tau)\,d\tau\right]$. Integrating the two
tones term by term,
$$\int_0^{t}m\,d\tau=\frac{\sin(2\pi f_1t)}{2\pi f_1}-\frac{3\cos(4\pi f_1t)}{4\pi f_1}
+\frac{3}{4\pi f_1},$$
so, absorbing the constant into an arbitrary phase reference, the phase deviation is
$$\theta(t)=\frac{k_f}{f_1}\sin(2\pi f_1t)-\frac{3k_f}{2f_1}\cos(4\pi f_1t)
=\beta_1\sin(2\pi f_1t)-\beta_2\cos(4\pi f_1t).$$
Evaluate the two modulation indices. Each tone contributes its own index,
equal to its own peak frequency deviation divided by its own frequency:
$$\beta_1=\frac{k_f\times 1\ \text{V}}{f_1}=\frac{2500}{2000}=1.25\ \text{rad},\qquad
\beta_2=\frac{k_f\times 3\ \text{V}}{2f_1}=\frac{7500}{4000}=1.875\ \text{rad}.$$
Notice that the larger-amplitude tone sits at twice the frequency, so its index is not simply
three times the first — the frequency in the denominator partly offsets the amplitude.
State the modulator output (part a). Collecting everything,
$$\boxed{\;s(t)=A_c\cos\!\Bigl[2\pi(10^{6}\!\ \text{-class }f_c)t
+1.25\sin\!\left(2\pi\,2000\,t\right)-1.875\cos\!\left(2\pi\,4000\,t\right)\Bigr]\;}$$
with the parameters: carrier amplitude $A_c$ (not specified by the question, so it is carried
symbolically), nominal carrier frequency $f_c$ (the grounded-input frequency),
$k_f=2.5$ kHz/V, $\beta_1=1.25$ rad at 2 kHz and $\beta_2=1.875$ rad at 4 kHz. The instantaneous
frequency is $f_i(t)=f_c+2500\left[\cos(2\pi 2000t)+3\sin(2\pi 4000t)\right]$ Hz.
Find the true peak of the message. The peak frequency deviation is
$\Delta f=k_f\max|m(t)|$, and because the two tones are harmonically related the peak is
not $1+3=4$ V. Writing $x=2\pi f_1t$ and using $\sin 2x=2\sin x\cos x$,
$$m=\cos x+3\sin 2x=\cos x+6\sin x\cos x.$$
Setting the derivative to zero, $-\sin x+6\cos 2x=0$, and substituting
$\cos 2x=1-2\sin^{2}x$ gives the quadratic $12\sin^{2}x+\sin x-6=0$, whose physical root is
$\sin x=2/3$ (hence $\cos x=\sqrt5/3$). Substituting back,
$$\max|m|=\frac{\sqrt5}{3}+6\cdot\frac{2}{3}\cdot\frac{\sqrt5}{3}=\frac{5\sqrt5}{3}
=\boxed{\;3.727\ \text{V}\;}$$
— some 7 % below the naive sum of the amplitudes, because the two tones never reach their
peaks at the same instant.
Peak deviation and Carson bandwidth (part b). Hence
$$\Delta f=k_f\max|m|=2500\times 3.727=\boxed{\;9.32\ \text{kHz}\;}$$
and with the highest message frequency $W=2f_1=4$ kHz, Carson's rule gives
$$B_T=2\left(\Delta f+W\right)=2\left(9.32+4\right)\ \text{kHz}
=\boxed{\;26.6\ \text{kHz}\;}$$
The deviation ratio is $D=\Delta f/W=2.33$, comfortably in the wide-band FM regime, so
Carson's rule (rather than the narrow-band approximation $B_T\approx 2W$) is the right tool. The
output spectrum consists of lines spaced 2 kHz apart — every combination
$n f_1+m(2f_1)$ — whose significant members fall inside that 26.6 kHz band.
Figure Q6.2 — illustrative FM line spectrum: sidebands spaced 2 kHz, significant content inside the 26.6 kHz Carson band.
Non-PLL demodulator (part c). The classical answer is a
frequency discriminator: convert frequency variations into amplitude variations, then
detect the envelope. In cascade: a limiter removes any incidental amplitude
modulation picked up in the channel (FM carries no information in its envelope, so hard limiting
costs nothing and rejects noise); a band-pass filter removes the harmonics the
limiter creates; a differentiator (or a slope circuit / detuned resonant
network, as in a balanced Travis or ratio detector) multiplies the signal by its instantaneous
frequency, producing $A_c\,\omega_i(t)$ as an amplitude; and an envelope detector with a
DC block recovers $m(t)$ up to a scale factor. Differentiating
$s(t)=A_c\cos[\omega_ct+\theta(t)]$ gives
$$\dot s(t)=-A_c\left[\omega_c+2\pi k_f m(t)\right]\sin\left[\omega_ct+\theta(t)\right],$$
whose envelope $A_c\left[\omega_c+2\pi k_fm(t)\right]$ is an exact affine image of the message.
This satisfies the “no phase-locked loop” restriction, and a quadrature detector or
zero-crossing counter would be equally acceptable answers.
Figure Q6.1 — frequency-discriminator FM demodulator (no phase-locked loop).
Check: the carrier amplitude $A_c$ and the numerical value of $f_c$ are not
given in the question (only that $f_c$ is the grounded-input frequency), so both are carried
symbolically; every other parameter is fixed by the data. Carson's rule is an engineering
approximation that captures roughly 98 % of the transmitted power — an exact Bessel-series
count of significant sidebands (those exceeding 1 % of the unmodulated carrier) would give a
slightly different figure, which is why the question asks for an approximate
bandwidth.