Skip to content
ELECTRONICS ENGINEERINGelectronics-engineering.com
Telecom Electronics

Solved oscillator exercises: Colpitts, Clapp, phase noise and crystal

Six exercises with solutions: frequency and start-up of a Colpitts, the Clapp variant, phase noise with Leeson's model, reciprocal mixing in a receiver, resonances and pulling of a crystal, frequency drift.

https://www.electronics-engineering.com/en/oscillator-exercises-colpitts-phase-noise-solved/

No ratings yet: be the first to rate this page.

Exercises on oscillators of the kind set in Telecommunication Electronics and Applied Electronics courses, at three levels: basic, for the direct use of the formulas; intermediate, on phase noise and its effects in receivers; exam, on crystals and stability. The result is given at once; the full solution opens with a click. The proofs are in the lesson and in the in-depth page linked at the top.

Basic level

Exercise 1. The frequency of a Colpitts

A Colpitts oscillator has L = 4.7 µH, C_1 = 470 pF and C_2 = 1 nF. Compute the oscillation frequency and the minimum gain g_mR for start-up.

Result: f_0 \approx 4.1 MHz; g_mR \ge 0.47.

Solution

The capacitance seen by the inductor is the series of the two capacitors:

    \[ C_s = \frac{C_1C_2}{C_1 + C_2} = \frac{470\cdot 1000}{1470}\ \mathrm{pF} \approx 320\ \mathrm{pF} \]

    \[ f_0 = \frac{1}{2\pi\sqrt{LC_s}} = \frac{1}{2\pi\sqrt{4.7\cdot 10^{-6}\cdot 3.2\cdot 10^{-10}}} \approx 4.10\ \mathrm{MHz} \]

The start-up condition is g_mR \ge C_1/C_2 = 0.47.

Exercise 2. The Clapp variant

A capacitor C_3 = 47 pF is added to the same circuit, in series with the inductor. What is the new frequency? How much does it change if the transistor capacitance adds 20 pF to C_1?

Result: f_0 \approx 11.5 MHz; with 20 pF more on C_1 the frequency drops by only about 0.2%.

Solution

    \[ \frac{1}{C_{eq}} = \frac{1}{470} + \frac{1}{1000} + \frac{1}{47}\ \mathrm{pF^{-1}}\quad\Rightarrow\quad C_{eq} \approx 41.0\ \mathrm{pF} \]

    \[ f_0 = \frac{1}{2\pi\sqrt{4.7\cdot 10^{-6}\cdot 4.10\cdot 10^{-11}}} \approx 11.5\ \mathrm{MHz} \]

With C_1 = 490 pF: C_{eq} \approx 41.1 pF, 0.4% more, and the frequency drops by about 0.2%. In the Colpitts without C_3 the same 20 pF would have moved C_s from 320 to 329 pF, almost 3%, and the frequency by 1.4%: this is the advantage of the Clapp.

Intermediate level

Exercise 3. Phase noise of an LC oscillator

A 100 MHz oscillator has a resonator with Q = 50, an amplifier with a 10 dB noise figure (F = 10), 1 mW in the resonator, f_c = 1 kHz. Estimate \mathcal{L} at 10 kHz from the carrier. (kT = 4\cdot 10^{-21} J.)

Result: \mathcal{L}(10\ \mathrm{kHz}) \approx -121 dBc/Hz.

Solution

The floor: 2FkT/P_s = 2\cdot 10\cdot 4\cdot 10^{-21}/10^{-3} = 8\cdot 10^{-17}, that is −161 dBc/Hz.

The resonator frequency: f_0/(2Q) = 10^8/100 = 1 MHz. At \Delta f = 10 kHz the second factor is 1 + (10^6/10^4)^2 \approx 10^4, that is +40 dB. The flicker factor is 1 + 10^3/10^4 = 1.1, about +0.4 dB.

    \[ \mathcal{L}(10\ \mathrm{kHz}) \approx -161 + 40 + 0.4 \approx -121\ \mathrm{dBc/Hz} \]

With a crystal resonator of Q = 10^5 at the same frequency, f_0/(2Q) = 500 Hz: at 10 kHz we would stay on the floor, at −161 dBc/Hz. This is why the reference of every synthesizer is a crystal.

Exercise 4. Reciprocal mixing

A receiver with 10 kHz wide channels must receive a weak signal while a transmitter 60 dB stronger is on the channel 25 kHz away. The local oscillator has −120 dBc/Hz at 25 kHz. What is the maximum signal-to-noise ratio?

Result: The signal-to-noise ratio cannot exceed 20 dB.

Solution

The strong signal mixes with the “skirt” of the oscillator and brings into the wanted channel a power equal to its own, attenuated by the phase noise integrated over the channel bandwidth:

    \[ -120\ \mathrm{dBc/Hz} + 10\log_{10}(10^4\ \mathrm{Hz}) = -80\ \mathrm{dBc} \]

The disturbance in the channel is therefore 80 dB below the strong signal, that is 60 - 80 = -20 dB relative to the wanted signal: the signal-to-noise ratio cannot exceed 20 dB, whatever the noise figure of the receiver. A modulation that needs 25 dB requires an oscillator at least 5 dB better.

Exam level

Exercise 5. The two resonances and crystal “pulling”

A 10 MHz crystal has C_m = 20 fF, C_0 = 5 pF, R_m = 20 Ω. Compute L_m, the quality factor, the distance between f_s and f_p and the frequency shift when it works with an 18 pF load capacitance C_L in series.

Result: L_m \approx 12.7 mH, Q \approx 4\cdot 10^4, f_p - f_s \approx 20 kHz, with an 18 pF load the frequency rises by about 4.3 kHz.

Solution

L_m = 1/(\omega_s^2C_m) = 1/((2\pi\cdot 10^7)^2\cdot 2\cdot 10^{-14}) \approx 12.7 mH: a huge inductance, impossible to build as a coil with such low losses. The quality factor is Q = \omega_sL_m/R_m = 2\pi\cdot 10^7\cdot 0.0127/20 \approx 4\cdot 10^4.

    \[ f_p - f_s \approx f_s\frac{C_m}{2C_0} = 10^7\cdot\frac{2\cdot 10^{-14}}{10^{-11}} = 20\ \mathrm{kHz} \]

With the load C_L in series, the resonance moves to the frequency where the inductive reactance of the crystal cancels C_L: f_L \approx f_s\left(1 + \frac{C_m}{2(C_0 + C_L)}\right), that is f_L - f_s \approx 10^7\cdot 2\cdot 10^{-14}/(46\cdot 10^{-12}) \approx 4.3 kHz, 435 ppm. Changing C_L by 1 pF shifts the frequency by about f_sC_m/(2(C_0 + C_L)^2) \approx 190 Hz, 19 ppm: this is the pulling used in VCXOs to fine-tune the frequency with a varactor.

Exercise 6. The effect of a phase shift

In a 433 MHz oscillator with an LC resonator of Q = 30, the transistor capacitance changes with temperature and produces a 0.02 rad phase shift. How much does the frequency move? Is it enough to stay within a 25 kHz channel?

Result: \Delta f \approx 144 kHz: no, it is about six channels.

Solution

    \[ \frac{\Delta f}{f_0} = \frac{\Delta\varphi}{2Q} = \frac{0.02}{60} \approx 3.3\cdot 10^{-4}\quad\Rightarrow\quad \Delta f \approx 433\cdot 10^6\cdot 3.3\cdot 10^{-4} \approx 144\ \mathrm{kHz} \]

Six channels of drift: unacceptable. The solutions are a SAW resonator (surface acoustic wave, with a Q of a few thousand) or a PLL locked to a crystal, which brings the stability back to a few ppm (about 1 kHz).

Subscribe
Notify of
guest
0 Comments
Oldest
Newest Most Voted
0
Would love your thoughts, please comment.x
()
x