Upload
others
View
2
Download
0
Embed Size (px)
Citation preview
MICROWAVE OVER FIBERApplications and Performance
March 31, 2010
Rochester IEEE JCM 20101
John A. MacDonald
Vice President of Engineering
Linear Photonics, LLC
Linear Space Technology, LLCHamilton, NJ 08619
609-584-5747
OUTLINE
1. MICROWAVE LINKS: ANALOG / DIGITAL
2. INTENSITY MODULATION & DETECTION
3. PRACTICAL LINK STRUCTURES
– DIRECT MODULATION
– EXTERNAL MODULATION
– PHOTORECEIVERS
– TRANSMISSION MEDIUM
4. PERFORMANCE LIMITATIONS
– LINEAR EFFECTS
– NONLINEAR EFFECTS
5. NONLINEAR PERFORMANCE IMPROVEMENT
6. SYSTEM TRADE-OFFS
7. ANALOG / DIGITAL REPRISE
2Rochester IEEE JCM 2010
Analog / Digital
• The bulk of fiber optic communications uses digital
modulation
– Fast switching and low pulse distortion determine link fidelity
• Certain applications not suited to digital:
– If bandwidth too high to be effectively digitized
– If system complexity is better suited toward simpler modulation (size, – If system complexity is better suited toward simpler modulation (size,
weight, power constraints)
– Examples to follow…
• Primary distinction between digital and analog is linearity
– Analog/Microwave links depend upon low distortion to achieve high
fidelity
• I will show later why everything is really analog
3Rochester IEEE JCM 2010
Microwave Link Applications
• Phased Array Radar/Comms
– Narrowband RF feeds directly to antenna arrays
– Lower weight, lower complexity
– True Time Delay beamsteering
• Antenna and Signal Remoting
– Direct-RF over longer distances (many km)
– Reliable alternative to wireless in fixed services– Reliable alternative to wireless in fixed services
• Electronic Warfare / SIGINT / ELINT
– Secure Comms (EMI hard)
• Connection to passive sensors (listening)
• Remoting personnel from active sensors (protection)
– Towed Decoys provide very high bandwidth
• Space-based
– Mass
– Deployed fiber: can be radiation hard; less thermally sensitive
• Time and Frequency Distribution
4Rochester IEEE JCM 2010
Microwave Link
Characterized by an RF input and an RF output
– Necessarily has a method of modulating and
demodulating an optical carrier
– Today’s Microwave Links dominated by
• Intensity modulation of semiconductor lasers
• Envelope detection using PIN or APD photodiodes
5Rochester IEEE JCM 2010
Intensity Modulation
6Rochester IEEE JCM 2010
Intensity Detection
Detection of Intensity-modulated optical signal can be performed using semiconductor photodiode
(photoelectric effect).
A photon (power) recombines to generate an electron (current)
For a P-I-N diode, the best case is one e- for every p
For an Avalanche Photodiode (APD), generally > 1 e- per p due to secondary recombinations: Gain (and Noise)
7
In either case, the ensemble average is called the Responsivity (amps/watt)
The electric current:
η = detector quantum efficiency
q = electron charge
The complex RF output spectrum: hω = photon energy
• The input square-law is reversed: RF output is linear with RF input
• Optical Phase information is lost
))](Re(21[)( tmRti +⋅∝
)()( tmFM ∝ω
DC term (ave power)
ωη
hq
R =
Rochester IEEE JCM 2010
Shot Noise
One side-effect of photoelectric effect is quantum shot noise
In particle terms, photons arrive randomly with Poisson distribution; each generates a shot of current. This gives
rise to a mean-square shot noise current density in W/Hz:
q = electron charge
R = photodiode responsivity
qRPi 22 =
8
R = photodiode responsivity
P = average optical power
z = load impedance
This signal is Gaussian (white) over all microwave frequencies and may limit the performance of the link.
Rochester IEEE JCM 2010
Other Modulation Formats
• Modulation need not be “Intensity”
– If m(t) is imaginary then “Phase Modulation” of the optical carrier
• Constant Intensity: p(t) = 1
• Difficulty in efficient/stable phase detection
– Coherent Detection
• Intensity modulated signal mixed with optical LO at receiver• Intensity modulated signal mixed with optical LO at receiver
– Optical Amplitude and Phase are preserved
– Requires additional filtering, higher laser stability
– Up/Down Converting Links
• Modulate twice:
– Sum and Difference signals
• A Multitude of approaches
– Focus here on Intensity Modulation
9
)()(Re21)( 21 tmtmtp +∝
Rochester IEEE JCM 2010
OUTLINE
1. MICROWAVE LINKS: ANALOG / DIGITAL
2. INTENSITY MODULATION & DETECTION
3. PRACTICAL LINK STRUCTURES
– DIRECT MODULATION
– EXTERNAL MODULATION
– PHOTORECEIVERS
– TRANSMISSION MEDIUM
4. PERFORMANCE LIMITATIONS
– LINEAR EFFECTS
– NONLINEAR EFFECTS
5. NONLINEAR PERFORMANCE IMPROVEMENT
6. SYSTEM TRADE-OFFS
7. ANALOG / DIGITAL REPRISE
10Rochester IEEE JCM 2010
Practical Intensity Modulation
• Direct
– Laser diode is modulated directly
• External
– Laser source drives a separate optical – Laser source drives a separate optical
component
11Rochester IEEE JCM 2010
Direct Modulation
Bias
ModulatingSignal
SemiconductorLaser
0
2
4
6
8
10
12
0 20 40 60 80 100 120 140 160
Bias Current (mA)
Optical Power
(mW)
Slope EfficiencyηL
IL
im
12
ITH
)()( mTHLLm iIIiP +−=η
IL
Rochester IEEE JCM 2010
• A CW optical signal is intensity modulated
via a field-dependent optical medium
• µwave / mm-wave modulation speeds can
be achieved with 2 major methods:
External Modulation
be achieved with 2 major methods:
– Electro-Optic Modulation
• Field-dependent change in optical index (electo-
optic effect)
– Electro-Absorption Modulation
• Field-dependent change in optical attenuation
13Rochester IEEE JCM 2010
• Mach-Zehnder Interferometer
– Index along propagation axis is dependent on
applied field (modulating signal)– Electro-Optic effect can be realized in Lithium Niobate
(LiNbO3), InP, and other crystal structures, i.e. KDP (KH2PO4)
Electro-Optic Modulation
(LiNbO3), InP, and other crystal structures, i.e. KDP (KH2PO4)
14
Vm (RF + Bias)
OpticalPropagation
• Propagation constant of the beam in the lower leg is retarded due to applied electric field – experiences less phase shift.
• Optical intensity (power) is modulated by the applied RF signal
Diffused Optical Waveguide on LiNbO3 substrate
Rochester IEEE JCM 2010
0
0.5
1
-15 -10 -5 0 5 10 15
Vm
Inte
nsi
ty
Vπ
π
φπ
V
VPVP m
mm2
cos)( 2 +=
ex. Vπ = 5 V
Optical Output Power:
Pm = max output intensity
typically 3 to 4 dB below input
Vm = bias voltage
Mach-Zehnder
15
Vmex. Vπ = 5 V
φ = 0
Vm = bias voltage
φ = phase offset (shift from origin)
[ ]tOMIV
tVm ωπ cos12
)( ⋅+=
Quadrature Analog CW:
OMI = Optical Modulation Index
• Optical Output power follows cos2 function o Vector phase summation
• Vππππ is DC voltage that causes 180° phase rotationo Depends on crystal physics and electrode length o Corresponds to “min” and “max” output powero Digital Modulation: variation between min and max
• Analog Modulation: Bias at Quadrature (shown)o Results in linear intensity modulationo Slope = 1 at quadrature pointo Even order distortions are balanced (zero)
Rochester IEEE JCM 2010
• Absorption of optical signal dependent on applied bias
– Transmission follows exponential relationship with applied field
• Exponential f(Vm) is dependent on device length, carrier confinement, and instantaneous change in absorption
• Generally not as linear as MZM
Electro-Absorption
16
0
0.25
0.5
0.75
1
0 0.5 1 1.5 2 2.5 3
Reverse Bias (V)
Relative Transmission
)()( Vmf
m eVP−=
N
P
ia
Vm
Pi
n
Pout
Rochester IEEE JCM 2010
• P-I-N diodes are most common
Photodetection
iout
VDC
Iave
PRPI ⋅=)(
R = responsivity (amps/watt)= ηq / hω
– Intrinsic bandwidth limited by diode capacitance
– Package and launch considerations may also limit
performance
• ~30 GHz bandwidth from lateral PIN
• ~100 GHz from waveguide PIN
17
= ηq / hω
Rochester IEEE JCM 2010
Transmission Medium
• Fiber is Optical Waveguide
• Fiber has very low loss
• High-fidelity Microwave Links
require Single-Mode fiber
cladcore nn >
require Single-Mode fiber
– The “single” mode is a hybrid:
• Nonzero E and M fields in direction
of propagation
• Two polarization modes
• Gives rise to mode dispersion
18
image from Wikipedia
Rochester IEEE JCM 2010
Singlemode
Link Gain
22
G DLηη∝
Transmitter Photo-Receiver
Optical Attenuation(OA)
Direct Modulation: 2OA
G ∝
22
22
OAV
PG DL
π
η∝
Direct Modulation:
External (MZM): PL = Laser Carrier Power
19Rochester IEEE JCM 2010
Direct Mod Intrinsic Gain
Single-frequency match at ωo:
Zin RL
Cj
Rs
Zout
HL(w) HD(w)
m(w)ηL
ηD
OA
D
L
WA
AW
=
=
)/(9.0
)/(25.0
η
ηTypical:
Z
Rochester IEEE JCM 2010 20
222
22
4 OARRCG
LSjo
DL
o ω
ηηω
=
Ls
j
D
RR
pFC
WA
=Ω=
=
=
)(5
)(2.0
)/(9.0η
out
soD
L
outoL
Z
RH
R
ZH
=
=
)(
)(
ω
ω
Then Link Gain at ωo is: Short Link (OA << 1):
)(64
)(5.1644
10
1
dBG
dBG
GHz
GHz
==
==
OUTLINE
1. MICROWAVE LINKS: ANALOG / DIGITAL
2. INTENSITY MODULATION & DETECTION
3. PRACTICAL LINK STRUCTURES
– DIRECT MODULATION
– EXTERNAL MODULATION
– PHOTORECEIVERS
– TRANSMISSION MEDIUM
4. PERFORMANCE LIMITATIONS
– LINEAR EFFECTS
– NONLINEAR EFFECTS
5. NONLINEAR PERFORMANCE IMPROVEMENT
6. SYSTEM TRADE-OFFS
7. ANALOG / DIGITAL REPRISE
21Rochester IEEE JCM 2010
• Linear Factors– Microwave Launch
– Inherent Bandwidth• Limited by device and package reactances
– Noise
– Fiber Medium
Performance Limits
– Fiber Medium• Absorptive Loss
• Dispersion (Chromatic, Polarization Mode)
• Nonlinear Factors– Distortion (Harmonic, Intermodulation, …)
– Fiber Medium• Stimulated Brillouin Scattering
• Raman Scattering
• Four-wave Mixing
22Rochester IEEE JCM 2010
Linear Impacts: Microwave Launch
Primary source of microwave loss is input/output matching
DC
10 GHz
RL
RS
C
RS
C
• Narrow band links can be reactively tuned• Limited Bandwdith
• Fano’s rule: (BW)(Reflection Coefficient) < c
• However: Many or Most links require Broadband• Must use lossy matching – affects link gain
RLCL
Forward Biased Laser Diode
• Small real resistance RL
• Junction Capacitance and Ohmic Resistance
Reverse Biased Photo Diode
• Large reactive impedance dominated by junction
capacitance
• Ohmic Resistance adds dissipative loss
CJ
23Rochester IEEE JCM 2010
Linear Impacts: Noise
• Laser Relative Intensity Noise (RIN)
– Caused by spontaneous emission (laser not a perfect
oscillator) RIN = dB/Hz w/r to Optical Carrier
– Receiver detects as microwave noise
– Noise Power follows detection square-law: No ~ I2·RIN– Noise Power follows detection square-law: No ~ I2·RIN
• Shot Noise
– Noise power follows Optical power: No ~ I
• Thermal Noise
– Ubiquitous Johnson Noise: No ~ constant
24Rochester IEEE JCM 2010
Output Noise of F/O Link
-180
-175
-170
-165
-160
-155
-150
-10 -8 -6 -4 -2 0 2 4 6 8 10
Optical Receive Power (dBmo)
No
ise
Po
we
r D
en
sit
y (
dB
m/H
z)
Thermal
Shot
RIN
Total
Gain of F/O Link
-60
-50
-40
-30
-20
-10
0
-10 -8 -6 -4 -2 0 2 4 6 8 10
Optical Receive Power (dBmo)
Ga
in (
dB
)
Noise Figure of F/O Link
0
10
20
30
40
50
60
-10 -8 -6 -4 -2 0 2 4 6 8 10
Optical Receive Power (dBmo)
No
ise
Fig
ure
(d
B)
Noise Figure
• Total output noise is • Gain maintains 2:1 • Resulting link noise
25
• Total output noise is related to optical received power (ORP):
2:1 in RIN region1:1 in shot region0 in thermal region
• Gain maintains 2:1 relationship with ORP
• Resulting link noise figure best at RIN limit. • Minimum value depends on laser RIN noise
Link Noise Figure and Dynamic Range vary with optical power
– defined in conjunction with the operational system
Rochester IEEE JCM 2010
Linear Impacts: Fiber Medium
• SM Fiber Attenuation
– Due primarily to Rayleigh (elastic) Scattering
– 0.25 dB/km (1550 nm)
– 0.5 dB/km (1310 nm)
• Chromatic Dispersion
– Wavelength dependent propagation velocity
– Sidebands arrive out-of-phase: gain nulling– Sidebands arrive out-of-phase: gain nulling
• Polarization mode Dispersion
– Orthogonal Polarization Modes different
propagation velocities
– Non-uniform through fiber length
• concentricity defects
• mechanical and thermal perturbations
• laser spontaneous emission
– Nondeterministic
– Affects pulsed (digital) systems
– Affects Time/Freq Distribution systems
26
Chromatic Dispersion
Polarization Mode Dispersion
FAST AXIS
SLOW AXIS
Rochester IEEE JCM 2010
Nonlinear Impacts:
Modulator Transfer Function
0
2
4
6
8
10
12
0 20 40 60 80 100 120 140 160
Bias Current
Optical Power Direct Modulation
• Gain Compression when drive level peaks reach threshold
• Even and Odd order amplitude distortion
• Phase distortion due to laser chirp (FM to PM)• Laser wavelength = function(drive level, temperature)
Bias Current
0
0.5
1
-15 -10 -5 0 5 10 15
Vm
Inte
nsi
ty
MZM External Modulation
• Gain Compression when as drive level approaches peaks
• Primarily Odd order amplitude distortion
• Little phase distortion
27Rochester IEEE JCM 2010
Linearity Measures
• Third-Order Intercept (IP3)
– (Imaginary) point where two-tone third-order intermodulation products (IMD3) are equal to the fundamental
• Spur-Free Dynamic Range
– Range of power over which the fundamental is above the noise floor and the IMD3 are below the noise floor
Pout
[ ])()(3)(1743
2)( 3
2
3 dBBWdBmIIPdBNFHzdBSFDR ++−=⋅
28
Pin
Output Noise Floor
IIP3SFDR-3
Rochester IEEE JCM 2010
Nonlinear Impacts: Fiber
• Stimulated Brilluoin Scattering (SBS)
– Vibrational/Acoustic oscillations generated by high energy photons
– Forward wave energy is converted to acoustic backward wave (phonons)
– Threshold Effect
– Reduced forward gain; Increased noise floor
• Stimulated Raman Scattering (SRS)
– Inelastic photon scattering Nonlinear fiber effects must be – Inelastic photon scattering
– Wavelength translation
– Reduction in gain
• Self-Phase Modulation
– AM-PM conversion of a single signal
• Cross-Phase Modulation
– AM-PM Transfer from one signal to another (WDM systems)
• 4-Wave Mixing
– Intermodulation Distortion (WDM systems)
29
Nonlinear fiber effects must be
considered during link design phase.
Rochester IEEE JCM 2010
Modulation Summary
TYPE COMPLEXITY SIZE
WEIGHT
POWER
PRACTICAL
MODULATION
FREQUENCY
LINEARITY COST
DIRECT Low: one optical
component
(laser)
Lowest < 12 GHz Poor 2nd and
3rd-order
performance
Lowest
ELECTRO- Moderate: Similar to 40+ GHz Poorest Higher,
30
ELECTRO-
ABSORPTION
Moderate:
requires separate
source laser and
small modulator
Similar to
direct mod
40+ GHz Poorest
Linearity, but
may have 2nd-
and 3rd-order
null operating
points
Higher,
comparable
to EOM
ELECTRO-
OPTIC (MZM)
Highest: requires
source laser,
large modulator,
plus optical and
electrical
controls for bias
locking
Highest 40+ GHz Well-defined
(sin curve).
Operation at
quadrature
provides 2nd-
order null.
Highest
Rochester IEEE JCM 2010
Closing the Link
LINK• Directly Modulated Laser / PIN Receiver / Postamplifier• Reactively Tuned 3.25 – 4.00 GHz
LINK• E-O Modulator (MZM) / Waveguide PIN Receiver• Broadband Response to 40 GHz
PHOTORECEIVERBroadband O/E Response to 20 GHzand Output Return Loss
31
Link Type
Centerband
Gain Input IP3 Noise Fig SFDR3dB dBm dB dB Hz^2/3
4 GHz
Direct Mod-20 32 28 118
20 GHz
MZM-25 31 38 111
30 GHz
EAM-30 18 40 101
40 GHz
MZM-30 25 43 104
“Typical” Link
Gain, Noise, Dynamic Range
Rochester IEEE JCM 2010
OUTLINE
1. MICROWAVE LINKS: ANALOG / DIGITAL
2. INTENSITY MODULATION & DETECTION
3. PRACTICAL LINK STRUCTURES
– DIRECT MODULATION
– EXTERNAL MODULATION
– PHOTORECEIVERS
– TRANSMISSION MEDIUM
4. PERFORMANCE LIMITATIONS
– LINEAR EFFECTS
– NONLINEAR EFFECTS
5. NONLINEAR PERFORMANCE IMPROVEMENT
6. SYSTEM TRADE-OFFS
7. ANALOG / DIGITAL REPRISE
32Rochester IEEE JCM 2010
• Linearization Techniques
– Linearization improves nonlinear distortion, increases dynamic range
– Major techniques under study include optical, electrical, and combinatorial approaches
Performance Improvement
electrical, and combinatorial approaches
• Electrical: aim is to cancel distortion products
– Feedforward/Feedback: Inject out-of-phase distortion products to cancel
– Predistortion: Nonlinear circuits with opposing distortion characteristics
• Optical: generally more complex
– Operates in optical domain – inherently wide-band
33Rochester IEEE JCM 2010
• Predistortion Linearization has long history in
broadcast power amplifiers; SSPAs, TWTAs,
klystrons, space and ground station equipment
– Generally much less complexity than optical or
combinatorial systems
Electrical Predistortion
combinatorial systems
• Does not rely on sampled waveforms
• Bandwidth is the major challenge
– The aim is to compensate for the gain and phase
compression of the nonlinear system by providing a
cascaded element function that has the opposite gain and
phase characteristic: gain and phase expansion
34Rochester IEEE JCM 2010
Predistortion Linearization
-5
-4
-3
-2
-1
0
1
2
3
-5 -4 -3 -2 -1 0 1
1
2
3
4
Input Power
Output Power
Phase
35
• Nonlinear Device exhibits Gain and Phase Compression
-2
-1
0
1
-5 -4 -3 -2 -1 0 1
Input Power
Rochester IEEE JCM 2010
Predistortion Linearization
-5
-4
-3
-2
-1
0
1
2
3
-5 -4 -3 -2 -1 0 1
1
2
3
4
-5
-4
-3
-2
-1
0
1
2
3
4
5
-5 -4 -3 -2 -1 0 1
1
2
3
4
Input Power Input Power
Output Power Output Power
Phase Phase
36
• Nonlinear Device exhibits Gain and Phase Compression
• Precede it with another nonlinear device that exhibits gain and phase expansion, in
conjugate with the device to be linearized (the linearizer)
-2
-1
0
1
-5 -4 -3 -2 -1 0 1
-2
-1
0
1
-5 -4 -3 -2 -1 0 1
Input Power Input Power
Rochester IEEE JCM 2010
Predistortion Linearization
-5
-4
-3
-2
-1
0
1
2
3
-5 -4 -3 -2 -1 0 1
1
2
3
4
-5
-4
-3
-2
-1
0
1
2
3
4
5
-5 -4 -3 -2 -1 0 1
1
2
3
4
-5
-4
-3
-2
-1
0
1
2
-5 -4 -3 -2 -1 0 1
-1
0
1
2
3
Input Power Input Power Input Power
Output Power Output Power Output Power
Phase Phase Phase
37
-2
-1
0
1
-5 -4 -3 -2 -1 0 1
-2
-1
0
1
-5 -4 -3 -2 -1 0 1
-5
-4
-3
-2
-1
-5 -4 -3 -2 -1 0 1Input Power Input Power Input Power
• Nonlinear Device exhibits Gain and Phase Compression
• Precede it with another nonlinear device that exhibits gain and phase expansion, in
conjugate with the device to be linearized (the linearizer)
• The desired outcome is an ideal limiter
– The linearity of an ideal limiter cannot be improvedRochester IEEE JCM 2010
MZM Linearized Link
(Predistortion)
5.7 dB1.5 dB
38
Non linearized @ 8 GHzP1 dB is 5.7 dB from saturationPhase compression rapidly above sat
Linearized @ 8 GHzP1 dB is 1.5 dB from saturationPhase nonlinearity < 1°
A measure of linearity is the distance between P1dB and Psat
(How close are we to an ideal limiter?)
Rochester IEEE JCM 2010
MZM Linearized Link
(Predistortion)MZM Microwave Link
Linearized IM D Products
60
80
100
120T
hir
d-O
rde
r IM
D (
dB
c)
Non Linearized
Linearized
39
0
20
40
0.01 0.10 1.00
Optical Modulation Index
Th
ird
-Ord
er
IMD
(d
Bc
)
MZM Microwave Link two-tone IMD measured at 8 GHz
> 20 dB improvement in IMD
Rochester IEEE JCM 2010
OUTLINE
1. MICROWAVE LINKS: ANALOG / DIGITAL
2. INTENSITY MODULATION & DETECTION
3. PRACTICAL LINK STRUCTURES
– DIRECT MODULATION
– EXTERNAL MODULATION
– PHOTORECEIVERS
– TRANSMISSION MEDIUM
4. PERFORMANCE LIMITATIONS
– LINEAR EFFECTS
– NONLINEAR EFFECTS
5. NONLINEAR PERFORMANCE IMPROVEMENT
6. SYSTEM TRADE-OFFS
7. ANALOG / DIGITAL REPRISE
40Rochester IEEE JCM 2010
Trade-Offs, or
Why Use Fiber?
• What are the System Engineer’s Tradeoffs?
• Should I consider fiber?
– Tradeoff performance, cost, weight vs. other options
• If fiber, then what do I need to know?
– Direct Mod vs. Ex-Mod (Decision Needed)– Direct Mod vs. Ex-Mod (Decision Needed)
• Direct Mod (usually < 12 GHz)
– Weight, Cost, Loss, Dynamic Range, DC Power
– Often Comparing directly to Coax
• Ex-Mod
– Must evaluate system impact for linearity, G/T, etc
41Rochester IEEE JCM 2010
Direct Mod Comparison to Coax
• At Direct Mod frequencies, choice of Fiber vs.
Copper often made on basis of performance as a
function of link length
• Primary System Trades include:
– Attenuation– Attenuation
– Noise Figure
– Weight
– Cost
42Rochester IEEE JCM 2010
Direct Mod Comparison to Coax
Attenuation
0
5
10
15
20
25
30
35
40
45
50
0 100 200 300 400 500 600 700 800 900 1000
Distance (meters)
Att
enuation (dB
)
RG8 100 MHz
RG8 1 GHz
RG8 2 GHz
LDF6 100 MHz
LDF6 1 GHz
LDF6 2 GHz
Direct Mod Link
Noise Figure
0
5
10
15
20
25
30
35
40
45
50
0 100 200 300 400 500 600 700 800 900 1000
Distance (meters)
Nois
e F
igure
(dB
)
RG8 100 MHz
RG8 1 GHz
RG8 2 GHz
LDF6 100 MHz
LDF6 1 GHz
LDF6 2 GHz
Direct Mod Link
Distance (meters)
Weight
0
50
100
150
200
250
300
350
0 10 20 30 40 50 60 70 80 90 100
Distance (meters)
Weig
ht (p
ounds)
RG8
LDF6
Direct Mod Link
Cost
$0
$1,000
$2,000
$3,000
$4,000
$5,000
$6,000
$7,000
$8,000
$9,000
$10,000
0 100 200 300 400 500 600 700 800 900 1000
Distance (meters)
Cost ($
) RG8
LDF6
Uncooled Direct Mod
Cooled Direct Mod
Distance (meters)
43Rochester IEEE JCM 2010
• Direct Mod Link attractiveness depends on system requirements.
– Crossover Length: when DM performance exceeds that of coax in various categories
Attenuation Noise Figure Weight Cost
RG8 110 m 180 m 30 cm 1800 m
Direct Mod Comparison to Coax
• Other factors may influence choice of Fiber vs. Copper:
– Linearity (link dynamic range is less than coax)
– EMI immunity (may trump performance)
– Safety, Reliability, etc.
RG8 110 m 180 m 30 cm 1800 m
Heliax LDF6 560 m 900 m 50 cm 100 m
Example: Crossover Lengths at 2 GHz
44Rochester IEEE JCM 2010
Some Comments on Digital Comms
• Modulation Formats: same hardware as analog
– Direct
• Laser driven between threshold (off) and Imax (on)
– External
• MZM Vmin (off) to Vmax (on), range = Vπ• MZM Vmin (off) to Vmax (on), range = Vπ
• EAM 0 (on) to Vmax (off)
• Receiver Formats
– Emphasis on limited (square wave) output
• PIN & Transimpedance Amp (TIA) / Limiter
• Sensitivity: dBmo ave optical power for 10-12 BER
45Rochester IEEE JCM 2010
It’s Really Analog
Power Spectral Density of 10 GB/s Random Binary Waveform (NRZ)
2
)sin()(
=
fT
fTTfSxx
π
π
0
20
Sxx f( )
1000.1 f
0.1 1 10 10020
15
10
5
0
GHz
Sxx(f)
(dB)
• High data rates require microwave design techniques
– Random binary waveform = Broadband microwave spectrum
– Requires broadband flat link response
– Gain Flatness, Noise, Linearity will impact Bit Error Rate
– Packaging, RF launch
– Ultra broadband waveforms
– Device capacitance and parasitics
GHz
46Rochester IEEE JCM 2010
• Applications for fiber optic links employing microwave
analog modulation steadily increasing
– Improvements in semiconductor laser manufacturing
– Availability of low cost optical passives and fiber optic cable
• Microwave links useful alternative to coaxial or wireless in
Summary
• Microwave links useful alternative to coaxial or wireless in
many applications
– Performance
– Size, Weight, Cost
– EMI
• Practical intensity modulation links were presented
– Typical Performance, Limitations, and Methods of Improvement
Rochester IEEE JCM 2010 47