Text preview for : a-204.pdf part of HP a-204 HP Publikacje a-204.pdf
Back to : a-204.pdf | Home
LOW NOISE
OSCILLATOR DESIGN
ROGER MUAT
and ART UPHAM
RF & Microwave
Measurement
Symposium
and
Exhibition .
www.HPARCHIVE.com
LOW NOISE OSCILLATOR DESIGN
I. Spectral Purity
A. What is spectral purity in oscillators?
B. What determines spectral purity?
II. Low Noise Oscillator Design
A. Establish objectives
B. Select a resonator
1. Measure: Qu, l/f noise, AM-FM conversion
2. Analyze phase noise
C. Select a circuit topology
D. Select an active device
E. Select matching networks
F. Measure:
1. Open loop gain/phase
2. Closed loop gain/phase
III. Oscillator Computer Analysis
A. Open loop
B. Closed loop
IV. Other Noise Mechanisms
A. Spurious modes
B. Upconverted noise
3
www.HPARCHIVE.com
_ _ _ _ Spectral Purity _
LOW NOISE
OSCILLATOR DESIGN
RF & Microwave
Measurement
Symposium
and
Exhibition
961
LOW NOISE OSCILLRTOR DESIGN
SPECTRRL PURITY
A. What is :;pectraJ purity in OSCillators?
B. What determine:; spedral purityl
LOYV NOISE OSCILLRrOR DESIGN
A. EstabW:;h objectives
B. Select a resonator
C. SelM:! a Circuit topology
O. Seled an active <:l.,yice
E. Select malchill9 networK
F. fAeasUte
OSCILLRTOR COMPUTER RNRLYSIS
A. Open loop
B. Close<:lloop
OTHER NOISE MECHRNISMS
A. SpuriOUS modes
B. Upconverted noise
962
Spectral Purity describes the degree of degrada-
tion from a perfect impulse in the frequency
SPECTRAL PURITY DEFINITIONS domain:
Vet) = cos W
o t
VO) = cos Wo t
V(W} = o V(f):: [1 +~R(t)] cos [w ot+q>(f)]
I I L----l
RM Noise PM NOise Real signals have some noise associated
SSB NOiSe Powet in a Vet) = [1 + eA(t)] cos [wot + (t)]
1Hz Bw f m Hz hom Cattiet
t.(fm)~ Total Si9nal Powet
We will focus primarily on phase noise (PM)
t.(fm) = S(2$ 4 power in a 1 Hz B.W. due to phase noise, offset
fm Hz, from the carrier, to the total signal power.
4 964 Ref. 2, 14, 15
www.HPARCHIVE.com
-This is a typical phase noise measurement. This
oscillator uses a 832.5 MHz surface acoustic
wave resonator as the resonator. How good is
this performance? Could it be better? What are
the limits to the noise performance of this oscil-
lator? What are the significant contributors to its
noise?
The previous phase noise plot was measured on
IIF~i.
Hewlett-Packard's 3047 phase noise measure-
ment system. This system has the capability of
measuring noise as low as --170 dEc. It covers
the frequency range 10 MHz to 18 GHz .