Phase Noise
Enrico Rubiola Dept. LPMO, FEMTOST Institute Besanon, France email rubiola@femtost.fr or enrico@rubiola.org
Summary
1. 2. 3. 4. 5. 6. 7. 8. 9. Introduction Spectra Classical variance and Allan variance Properties of phase noise Laboratory practice Calibration Bridge (Interferometric) measurements Advanced methods References www.rubiola.org
you can download this presentation, an e-book on the Leeson effect, and some other documents on noise (amplitude and phase) and on precision electronics from my web page
1 Introduction
V0
V0
v (t)
polar coordinates
Figure 1.1: .
where V0 is the nominal amplitude, and the normalized amplitude uctuation, (t) = V0 cos frequency 0 t + n Cartesianwhich coordinates is adimensional. v The instantaneous isc (t) cos 0 t ns (t) sin 0 t
|nc (t)|
(1.3) that nc (t) ns (t) This of stable signals with (t) = of the form and (1.2), (t )= V and deals |nswith (t)| the measurement V0 0 book 0 main focus on phase, thus frequency and time. This V involves several topics, V0 (t) =
= 2
= 2
Basic problem: how can we measure a low random signal (noise sidebands) close to a strong dazzling carrier?
traditional instruments for phase-noise measurement (saturated mixer) bridge (interferometric) instruments
1. 2. 3. 4.
too wide IF bandwidth noise and instability of the conversion oscillator (VCO) detects both AM and PM noise insufcient dynamic range
Some commercial analyzers provide phase noise measurements, yet limited (at least) by the oscillator stability
The Schottky-diode double-balanced mixer sa The Schottky-diode double-balanced mixer saturated at both at both inputs is the most used phase detec inputs is the most used phase detector
s(t) =
r(t) =
The AM noise is rejected by saturation Saturation also account for the phase-to-voltage gain k
2 Spectra
Spectra denitions
Sv (f ) = F v (t)
This is possible (Wiener-Khinchin theorem) with ergodic processes In many real-life cases, processes are ergodic and stationary Ergodicity: ensemble and time-domain statistics can be interchanged. This is the formalization of the reproducibility of an experiment Stationarity: the statistics is independent of the origin of time. This is the formalization of the repeatability of an experiment
Spectra meaning
Sv (f0 ) P = R0
Sv (f0 ) R0
Spectra meaning
The power spectrum density extends the concept of root-mean-square value to the frequency domain
DSB
B
P0 B N
v0
v0 +f
v0 f
v0
p
R0 N B R0 P0
v0 +f
R0 N B/2 R0 P0
rms =
1 p = 2
NB 2P0
rms =
1 p = 2
NB P0
dBc/Hz
N L= 2P0
3 dB
N S = P0
dBrad2/Hz
(f) (re)defined
The rst denition of (f) was (f) = ( SSB power in 1Hz bandwidth ) / ( carrier power ) The problem with this denition is that it does not divide AM noise from PM noise, which yields to ambiguous results Engineers (manufacturers even more) like (f)
Useful quantities
phase time
1 x(t) = (t) 20
x(t) is the phase noise converted into time uctuation physical dimension: time (seconds)
y(t) is the fractional frequency uctuation -0 normalized to the nominal frequency 0 0 y (t) = (dimensionless) 0
Spectra jitter
Jitter Jitter
The phase fluctuation can be described in terms of a single The phase fluctuation be described in terms of a single parameter, either phasecan jitter or time jitter parameter, either phase jitter or time jitter
The phase noise must be integrated over over the the bandwidth of the The phase noise must be integrated bandwidthBB of (which may bebe difficult to identify) thesystem system (which may difficult to identify)
&rms ! x rms
S & ' f ( df % $
B
radians
1 ! 2 " #0
S & ' f ( df % $
B
seconds
The jitter is useful in digital circuits because the bandwidth B is known The jitter is useful in digital circuits because the bandwidth is known
lower limit: the inverse propagation time through the system - lower limit: the inverse propagationdivergent time through the system
this excludes the low-frequency processes) this excludes the low-frequency divergent processes)
upper limit: ~ the inverse switching speed - upper limit: the inverse witching speed
Victor Reinhardt (invited), A Review of Time Jitter and Digital Systems, Proc. 2005 FCS-PTTI joint meeting
All these noise types are generally present at the output of oscillators, while Typical phase noise of some devices and oscillators two-port devices show white phase and icker phase noise only. For reference, Fig. 1 reports the typical phase noise of some oscillators and devices.
10 GH zS ap p GH
b2 f b3 f 3 b4 f 4
Spectra examples
-40 -60 -80 -100 -120 -140 -160 -180 -200 0.1 1
10
zD
iel
ec
10
GH
hir eR
tri
cR
10 0M
zS
es
es
RO
on
on
Hz
qu
wi th
ato r
ato
Os
rO
art
no
cil
sc
zo
ise
lat
ill
micro
wave a
sci l
lat
de ge ne rat
or
ato
(S R
O)
ion
micro
wave fe
rrite i
solato
rs
10
100
1k
10k
100k
3 Variances
1 y ! 1 y ! "0 "0
Variances classical
normalized reading of a counter that normalized reading of over a counter that measures (averages) a time T measures (averages) over a time T
) )
N N
%%
2 N classical variance, N y 2 classical variance, j file of N counter readings j !1 y j file of N counter readings j !1average of the N readings
) )
&&
For a given process, the classical variance For a given process, the classical variance For a given process, the classical variance depends of N depends of N depends of N Even worse, if the spectrum f-1 or steeper, Even worse, if the spectrum is f-1is or steeper, -1 or steeper, Even worse, if the spectrum is f the classical variance diverges the classical variance diverges the classical variance diverges The filter associated to measure the measure The filter associated to the The associated to the measure takes in the dc component takes infilter the dc component takes in the dc component
Enrico 21 21 Enrico Rubiola Noise Phase Noise Enrico Rubiola Rubiola Phase Phase Noise 21
Variances Allan
Zero dead-time two-sample variance Zero dead-time two-sample variance Zero dead-time two-sample variance (Allan variance) (Allan variance) (Allan variance)
1 2 2 1 2 2 !y " y % y !y " y % y 2 22 12 1
Denition Definition Definition (Let N == 2, and (Let N average) (Let N2, = and 2,average) and average)
( $(
$ &
&
Estimated Allan variance, Estimated Allan variance, Estimated Allan variance, le of readings file of mm counter readings file of counter m counter readings
The to difference The filter associated to the difference Thefilter lterassociated associated tothe the difference of measures is oftwo two contiguous measures isa aband-pass band-pass ofcontiguous two contiguous measures is a band-pass (about one octave) (about one octave) The estimate converges to the variance The estimate converges to the variance The estimate converges to the variance
Variances Allan
Variances Allan
b0
h2 f
b1 f 1 b2 f b3 f
2 3
h1 f h0 h1 f
h1 h2
b3 = 2 0 b4 = 2 0
random walk b4 f 4 FM
h2 f 2
frequency drift y = Dy
fH is the high cuto frequency, needed for the noise power to be nite.
Frequency synthesis
Frequency synthesis
Ideal synthesizer
Ideal synthesizer - -noise-free noise-free - -zero time zero delay delay time
timetranslation: translation: time output jitter jitter = jitter output =input input jitter phase time time x ==x xi i oo phase x linearity of the integral and the linearity of the integral and the derivative operators: derivative operators: o = (n/d)i => o = (n/d)vi !o = (n/d)!i => "o = (n/d)"i spectra spectra
n S! o " f # $ d
"#
S! i " f #
Carrier collapse
Simple physical meaning, complex mathematics. Easy to understand in the case of sinusoidal phase modulation
Filtering <=> Phase Locked Loop (PLL) Filtering <=> Phase Locked Loop (PLL) Filtering <=> Phase Locked Loop (PLL)
k ! " V ( rad #
k o " rad ( s ( V #
The FFT analyzer (n The FFT analyzer (not b The FFT analyzer (not can needed here) needed here) can be used needed here) can be us to measure S !(f) to measure S(f)S!(f) to measure tracks 1
tracks 1
2
The PLL low-pass The PLL low-pass The PLL low-pass filters the phase lters the phase filters the phase
Output voltage: the PLL Output voltage: the PLL Output voltage: the PLL is a high-pass filter is aa high-pass lter is high-pass filter
S ! 2""ff## S !2 S ! 1""ff## S
!1
$ $
$ $
%%
% #%
o !
Frequency discriminator
A resonator turns a slow frequency uctuation into a phase uctuation
0
phase
resonator 0 Q
For slow frequency uctuations, a delay-line t is equivalent to a resonator of merit factor
The Leeson effect: The Leeson phase-to-frequency noise Effect conversion in oscillators
Enrico Rubiola Dept. LPMO, FEMTO-ST Institute Besanon, France e-mail rubiola@femto-st.fr or enrico@rubiola.org D. B. Leeson, A simple model for feed back oscillator noise, Proc. IEEE 54(2):329 (Feb 1966) D. B. Leeson, A simple model for feed back oscillator noise, Proc. IEEE 54(2):329 (Feb 1966) out
resonator
S ! " f # $ 1&
oscillator noise
%" # (
'0
2
2Q
S) " f #
ampli noise
oscillator noise
S! " f #
E. Rubiola, The Leeson effect, Tutorial 2A, Proc. 2005 FCS-PTTI (tutorials) some free documents on noise (amplitude and phase) and on E. Rubiola, The Leeson effect,will e-book, (http://arxiv.org/abs/physics/0502143 or rubiola.org) precision electronics be available in this non-commercial site
www.rubiola.org
5 Laboratory practice
RF mixer (5-10) MHz Good operating conditions (10 dBm each input) Low-noise preamplier (1 nV/Hz)
Warning: if only one arm of the power supply is disconnected, the LT1028 may delivers a current from the input (I killed a $2k mixer in this way!) You may duplicate the low-noise amplier designed at the FEMTO-ST Rubiola, Lardet-Vieudrin, Rev. Scientic Instruments 75(5) pp. 1323-1326, May 2004
k o " rad ( s ( V #
k ! " V ( rad # k o " rad ( s ( V #
The PLL low-pass filters the phase Phase: the PLL The PLL low-pass
is a low-pass filters thelter phase
The FFT analyzer (not needed here) can be used to measure S!(f) tracks 1 2
The The signal 2 FFT ana tracks 1 needed here to measure S tracks 1
Output voltage: the PLL Output voltage: the PLL Output voltage: the PLL is a is high-pass lter is a high-pass filter filter a high-pass
2 4 & f 2 k! $ $ 2 2 S1! S! " f1#" f # 4 &2 f 24 '% k o k! "fk #%2 k H " & fH c' o ! c
S vo " f #
f #%2
compare an oscillator under test to a reference low-noise oscillator or compare two equal oscillators and divide the spectrum by 2 (take away 3 dB)
but you have to correct the spectrum for the PLL transfer function
a. power level b. lock condition (there is no beat note at the mixer out) c. zero dc error at the mixer output (a small V can be tolerated)
2. 3. 4.
Choose the appropriate time constant Measure the oscillator noise At end, measure the background noise
At low Fourier frequencies, the synthesizer noise is lower than the oscillator noise At higher Fourier frequencies, the white and icker of phase of the synthesizer may dominate
Due to the lower carrier frequency, the noise of a VHF synthesizer is lower than the noise of a microwave synthesizer. This scheme is useful with narrow tuning-range oscillator, which can not work at the same freq. to prevent injection locking due to microwave leakage
Afterwards, you will appreciate more the commercial instruments: assembly instruction manual computer interface and software
6 Calibration
Calibration general
2 measure the mixer gain k (volts/rad) > next 3 measure the residual noise of the instrument
Calibration general
Make sure that the power and the quadrature are the same during all the calibration process
Calibration measurement of k
tone:
white noise
Some FFTs have a white noise output Dual-channel FFTs calculate the transfer function |H(f)|2=SVm/SVd
Calibration measurement of k
Calibration measurement of k
Wheatstone bridge
Bridge Wheatstone
equilibrium: Vd = 0 > carrier suppression static error Z1 > some residual carrier real Z1 => in-phase residual carrier Vre cos(0t) imaginary Z1 => quadrature residual carrier Vim sin(0t) uctuating error Z1 => noise sidebands real Z1 => AM noise imaginary Z1 => PM noise
Bridge scheme
Interferometer scheme
Synchronous detection
Synchronous in-phase and quadrature detection Synchronous in-phase and quadrature detection
Bridge summary
Bridge summary
Bridge commercial
You will appreciate the computer interface and the software ready for use
8 Advanced Techniques
Low-flicker scheme
Interpolation is necessary
Advanced correlation
Advanced correlation
Advanced matrix
x( t )
CP3
DUT R 0 =50
R0
arbitrary phase
1020dB coupl.
R0
RF
v1
FFT analyz.
w1
B matrix w2
atten RF I u1 IQ modul Q u2
Advanced comparison
er e r ble om ete int e r fer r cor om rel e . sa t. m ter res ix. idu al f lick res idu er, res al f by idu lick ste al f res p in er, idu lick fix ter al f er, ed i fer lick om fix nte ete er, ed i rfer r nte fix o ed rfe mete int r erf omet r ero e me r ter , 4 5
int
erf
ted
mix
nested interferometer
saturated mixer
Fourier frequency, Hz
10 10 2
10 3
104
105
9 References
STANDARDS J. R. Vig, IEEE Standard Denitions of Physical Quantities for Fundamental Frequency and Time Metrology--Random Instabilities, IEEE Standard 1139-1999 ARTICLES J. Rutman, Characterization of Phase and Frequency Instabilities in Precision Frequency Sources: Fifteen Years of
Progress, Proc. IEEE vol.66 no.9 pp.1048-1075, Sept. 1978. Recommanded E. Rubiola, V. Giordano, Advanced Interferometric Phase and Amplitude Noise Measurements,
Rev. of Scientic Instruments vol.73 no.6 pp.2445-2457, June 2002.
Interferometers, low-icker methods, correlation, coordinate transformation, calibration strategies,
advanced experimental techniques BOOKS Chronos, Frequency Measurement and Control, Chapman and Hall, London 1994.
Good and simple reference, although dated W. P. Robins, Phase Noise in Signal Sources, Peter Peregrinus,1984.
Specic on phase noise, but dated. Unusual notation, sometimes difcult to read. Oran E. Brigham, The Fast Fourier Transform and its Applications, Prentice-Hall 1988.
A must on the subject, most PM noise measurements make use of the FFT W. D. Davenport, Jr., W. L."Root, An Introduction to Random Signals and Noise, McGraw Hill 1958.
Reprinted by the IEEE Press, 1987.
One of the best references on electrical noise in general and on its mathematical properties. E. Rubiola, The Leeson effect (e-book, 117 pages, 50 gures) arxiv.org, document arXiv:physics/0502143 E. Rubiola, Phase Noise Metrology, book in preparation ACKNOWLEDGEMENTS I wish to express my gratitude to the colleagues of the FEMTO-ST (formerly LPMO), Besancon, France, rsts of which Vincent Giordano and Jacques Groslambert for a long lasting collaboration that helped me to develop these ideas and to put them in the present form.