Documente Academic
Documente Profesional
Documente Cultură
LW source O/O LW
receiver
O/
E
Electrical
receiver/
O display
E/
E/E Amp
RF source Device
under
test
O/O, E/O, O/E, or E/E
Table of Contents Page
Introduction 3
General measurement techniques and considerations 3
2
Introduction
General measurement
techniques and Device
considerations under
test
The concept of lightwave com-
ponent analysis is straightfor- Input Modulator Output
ward. Measurements are made
of the small-signal linear trans-
mission and reflection charac-
teristics of a variety of light-
E/O measurement =
wave components. A lightwave
component analyzer consists of Pmod, out
Laser
a microwave network analyzer Imod, in
with an optical test set attached
to it. A precise electrical (signal
generator) or optical (transmit-
ter) source is used to stimulate
the component under test and
a very accurate and calibrated Photo diode
optical or electrical receiver O/E measurement =
measures the transmitted (or Imod, out
reflected) signal. Since charac-
Pmod, in
terization over a range of modu-
lation frequencies is required,
the frequency of modulation is
Fiber
normally swept over the band-
width of interest. O/O measurement =
Pmod, out
Measurements are typically Pmod, in
comprised of the appropriate
ratio of microwave modulation
current (or power) and light-
wave modulation power (see Figure 2. Measurement signals
Figure 2). While Figure 1 dem-
onstrates the basic concepts of
lightwave component analysis,
the specific measurement pro-
cesses are illustrated later. An
analysis of how various signals
are used in the measurement
process is found in Appendix 1,
“Signal Relationships in Opto-
Electronic Devices.”
3
E/O measurements O/E measurements O/O measurements
(modulators, lasers, LED’s) (photo diodes and
receivers) Characteristics of purely
The measurement of an E/O optical devices can also be
transducer is a combination The measurement process for measured. In this case, both
of input modulating current O/E devices is similar to E/O the stimulus and response
or voltage and output optical devices. The measurement are modulated light. The ratio
modulation power. For current consists of the ratio of output measurement is simply one of
converting devices the slope electrical modulation current to gain or loss versus modulation
responsivity is used to describe input optical modulation power. frequency.
how a change in input current Slope responsivity for O/E
produces a change in optical devices describes how a change The LCA Family
power. Graphically this is shown in optical power produces a
in Figure 3. For modulators the change in electrical current. There are several instruments
behavior is shown in Figure 4. Graphically this is shown in in the LCA family. Their char-
Figure 5. The LCA measures the acteristics are summarized
An LCA measures input modu- input optical modulation power below.
lating current and output modu- and output modulation current
lation power and displays the and displays the ratio of the two
ratio of the two in Watts/Amp, in Amps/Watt.
either linearly or in decibels.
4
Pout
(mW)
Responsivity RS (W/A) = ∆Pout/∆Iin
RS (dB) = 20 log10 [RS(W/A)/1(W/A)]
Iin(MA)
Pout
Responsivity Rs (W/A) = ∆Pout/∆I (mW)
in
Rs (dB) = 20 log10 [Rs(W/A)/1(W/A)]
Iout
Responsivity Rr (A/W) = ∆Iout/∆Pin (mA)
Pin (mW)
5
Figure 6. N4375B
6
A guided setup (“calibration Path de-embedding The characterization results
wizard”) procedure will guide need to be in the form of a
you through the necessary The path de-embedding feature 2-port S-parameter data file
steps. With ECAL, an automatic, of the LCA allows to remove any (also known as Touchstone
full, two-port calibration can be additional paths between the .s2p data files). All the
accomplished with a single con- calibration reference planes and S-parameters in the supplied
nection to the ECAL module and the device under test, i.e., RF file are used. According to the
minimal operator interaction. cables, adapters, or probes and convention of the LCA (where
optical fibres, optical attenua- any port can be an input or
This establishes the electrical tors. This feature is particularly an output), these values are
calibration reference planes for helpful when path elements directional. This means port
the measurements. Whenever need to be excluded from the 1 of the connector is always
there are additional electrical measurement results, like addi- connected to the network
adapters or elements required tional propagation delay intro- analyzer and port 2 is al-
for connecting the device under duced by fiber leads, or need to ways connected to the DUT.
test these addional electrical be changed after the electrical The AdapterChar macro of
paths can be accounted for by calibration, like setting different the network analyzer in-
“RF path de-embedding” feature levels of optical attenuation. It cludes this directionality in
(see “RF Path De-Embedding”). is also needed when elements its characterization.
cannot be addressed directly by
the electronic calibration at the Optical path de-embedding
Configuring the
LCA measurement RF cable ends, like when work-
ing with wafer probes. For optical path de-embedding
the de-embedding information
The next stage is the LCA RF path de-embedding can be introduced via param-
measurement configuration. To eters, like fiber length, refrac-
simplify the process of making For RF path de-embedding tive index, and attenuation, or
measurements, LCAs have a the RF path needs to be via an adapter file. The adapter
built in “confguration screen” characterized prior to starting file needs to be in the form of
feature. the LCA measurement. For a 2-port S-parameter data file
characterization, we recommend (also known as Touchstone
On this screen, the user can set to use the AdapterChar macro .s2p data files). All the
all LCA measurement parame- supplied with the network S-Parameters in the supplied
ters. An advanced configuration analyzer. Please refer to the file are used. The optical path
screen allows the inclusion of online help on the network can be characterized via an
electrical adapters (e.g., mating O/O measurement.
analyzer for further details.
connectors or cables) and opti-
cal adapters (e.g., optical cables
or attenuators) via parameter
settings or measurement files.
7
When chosing the OO configu-
ration the measurement of a
patch cord is a simple verifica-
tion of system performance.
Accuracy considerations
8
Lightwave Transmitter
Measurements (E/O)
The LCA is used to characterize
the transmission and reflection
parameters of electro-optic
modulators, laser and LED
sources with respect to modula-
tion frequency. The transmission
measurements to be discussed
include:
9
Figure 9. E/O modulation bandwidth measurement
The display units are Watts The other is how efficiently an The relaxation oscillation phe-
per Amp (the vertical axis). In electrical signal is delivered to nomenon creates a resonance
this case, the display is in a the laser. The electrical return in the frequency response,
logarithmic format where 0 dB loss is an indicator for the noise, and distortion of the
represents 1 watt per amp. The quality of impedance match- laser. Figure 10 is a composite
horizontal axis is modulation ing. For the particular laser the of a bandwidth measurement
frequency, indicating that the matching is degraded at high made at four different bias/
measurement is being made frequencies. (For details on the output power levels. Note in
over a wide range of frequen- impact of transmitter impedance the two lower traces that the
cies, in this case from 10 MHz matching see “Transmitter Input response tends to peak before
to 20 GHz. As stated, this Impedance.”) rolling off. This is the region
measurement indicates how of relaxation oscillation. Care
fast the laser can be modulated. The effects of bias on laser must be taken when modulating
This particular laser shows the performance a laser in this region, because
typical relaxation resonance this is where noise and distor-
peak at about 16.5 GHz. Beyond The frequency response of tion properties are often at their
this frequency, the conversion a laser is also dependent on worst. As bias is increased,
efficiency is gradually degraded. biasing conditions. As the DC both peaking and bandwidth
There are two significant com- bias of the laser is increased, increase. For this laser, as bias
ponents that limit the modu- the bandwidth will generally reaches a certain point, the
lation bandwidth. One is the increase. This is typically due peaking begins to saturate. (See
actual construction of a laser to the “relaxation oscillation” Agilent Application Note 371,
including the physical dimen- characteristics that vary with “Measuring Modulated Light.”)
sions and fabrication process. bias.
10
DFB laser 1550 nm
-4 dBm
S21 (dB(W/A))
-1 dBm
2 dBm
5 dBm
Frequency (Hz)
13
To display the modulator re- 50 GHz. The bad modulator has For instance, unfortunately the
sponsivity at a reference output a sharp resonance resulting in input impedance of an ac-
power level use the power very large S11 and consequently tive laser is much lower than
level measured with the inter- a deep notch in the S21 response, the transmission system used
nal power meter (converted to making the component unus- to drive it. Two problems oc-
Watt) and convert the respon- able in many applications. cur when such an impedance
sivity using the equation feature mismatch exists. First, a signifi-
with equation 1 below. Transmitter input cant amount of energy will be
impedance reflected at the transmission
Modulator analysis
line/laser interface. This re-
Figure 15 is a measurement of flected energy may eventually
The conversion efficiency of a
the relative conversion effi- be re-reflected and distort the
transmitter is dependent not
ciency (S21) and electrical return desired data signal. The second
only on the inherent properties
loss (S11) of two wide bandwidth problem is that the reflected
of the transmitter, but also on
(40G) external modulators as energy is “wasted” since it is
how efficiently the electrical
a function of modulation fre- never effectively used to modu-
modulation signal is delivered
quency. The unusual response late the laser. Thus, the overall
to the laser or modulator chip.
of the “bad” modulator in Figure conversion efficiency of the
High speed modulation signals
15 in the 55GHz frequency range laser is degraded.
are generally transmitted to the
is due imperfections of the
chip over transmission lines
electrical impedance matching
with a 50 or 75 Ω characteristic
/RF coupling in the modulator
impedance. Maximum power
package. The measurement
transfer will occur if the input
examples clearly demonstrate
impedance of the chip is the
issues in the electrical perfor-
same as the transmission line.
mance of the modulator above
S21 @ Pref [dBm] = S21 @ Optical DC Power [dBm] * 10^(0.1*(Pref [dBm]-Optical DC Power[dBm]))
Equation 1. Modulator responsivity
40 GHz modulator
S21 (dB)
S11 (dB)
Frequency (Hz)
Figure 15. Modulator bandwidth S21 and electrical return loss S11 of two different
40 GHz modulators
14
Measurement procedure Measurement interpretation Implications of impedance
Figure 15 shows the return The return loss over the whole mismatch on measurement
loss of modulators with match- 67 GHz range varies from a accuracy
ing circuit as measured on the best case of below –18 dB to
lightwave component analyzer. a worst case of –2 dB. It is not When the input impedance of
The measurement is made by unusual for the reflection level the E/O device under test is far
sending a swept RF signal to to get worse as the modulation from 50 Ω, a significant portion
the modulator under test and frequency is increased. Return of the electrical energy sent to
measuring the energy that loss is the ratio of reflected to the device will be reflected. This
reflects back. The setup and incident energy (10 Log (P refl/ reflected energy is accounted
calibration procedure will de- P inc)). The larger the return for in the full two-port measure-
pend on the model of LCA used. loss magnitude, the smaller the ment and thus does not degrade
In all cases, a one port (or full reflected signal and the better measurement accuracy as in
two-port) calibration, preferably the impedance match. previous LCA systems. However,
with an electronic calibration kit the intrinsic device response is
with integrated known electrical A different display for return distorted by ripples in frequency
reflection standards is required loss plot is in the Smith chart response measurements result-
for highly accurate reflection format. A Smith chart is a form ing from reflection.
measurements. of an impedance map (e.g., see
Figure 19). The display shows
the input impedance as a func-
tion of frequency. For the modu-
lator the impedance is near
50 Ω over the 30 GHz range, as
the response would not devi-
ate much from the center of the
chart. The Smith chart data pre-
sentation is selected under the
“Format” key menu. For more
information on network analyzer
display formats see Application
Note AN1287-1.
15
Lightwave Receiver
Measurements (O/E)
The measurements that the
LCA makes on lightwave receiv-
ers are in many ways similar
to those made on lightwave
sources. In this case, the stimu-
lus will be modulated light and
the response will be “demodu-
lated” electrical signals. Mea-
surements include:
16
After completion of the calibra-
tion, one might expect to see a
flat response at 0 dB indicating
the test system response has
been removed. When using an
Agilent N4373B, the display
seen upon completion of the
user calibration process will not
necessarily be a flat line. The
O/E receiver used in the calibra-
tion, which is still in the mea-
surement path, has become the
DUT. Thus its response is now
displayed.
17
Photo diode output impedance The energy will then be re-
reflected in the forward direc-
Once the photo diode has con- tion and potentially interfere
verted the modulated light to a with primary signals. Thus,
proportional electrical current, reflections can lead to commu-
the task is then to efficiently nication degradation.
transmit the demodulated signal
Measurement procedure and
to any following electrical com- interpretation
ponents. High-speed systems The setup and measurement
usually require this transfer of photo diode return loss are
over 50 or 75 Ω transmission identical to the procedure used
lines. The output impedance of in characterizing laser return
a photo diode is usually much loss. See “Laser input imped-
higher than 50 Ω (or 75 Ω). ance”. Figure 19 shows the
This leads to the possibility of return loss of an optical receiver
signal reflections and degraded measured with the component
conversion efficiency. If the analyzer, displayed on a Smith
signal transmitted from the chart. A Smith chart is a form of
photo diode encounters another an impedance map. The display
impedance mismatch along the shows the output impedance as
transmission path, energy will a function of frequency.
be reflected back towards the
photo diode.
18
For this receiver, an electri- Using the time domain feature of
cal amplifier follows the photo the LCA can help to determine
diode, so the measured imped- the locations of any discontinui-
ance is essentially that of the ties in the electrical path of the
amplifier. Over the 20 GHz mea- photo diode assembly.
surement range, the impedance
strongly deviates from 50 Ωs Figure 20 is a time/distance
(the center of the Smith chart). representation looking back into
The ideal case would be for the a ROSA (receiver optical sub-
impedance to be a constant assembly) device mounted in
50 Ω (or 75 Ω). The Smith chart a test fixture. The reflection at
data presentation is selected marker 3 stems from a surface
under the “Format” key menu. mount blocking capacitor. At
marker 1 one sees reflections at
The data can also be displayed the connection of ROSA con-
simply as return loss, the ratio nection pins to test fixture and
of reflected to incident power at marker 2 one sees the key
(10 Log (P refl/ P inc) as dis- reflection from ROSA output
played in the yellow trace of chip.
Figure 17.
19
Balanced Device network analyzer of the LCA For OE measurements the single
is capable of measuring both ended (SE) port 1 is assigned
Measurements individual single-ended trans- to network analyzer port 1 and
Optical transceivers for 10 GE mission characteristics from the balanced (BAL) port 2 is
applications or new (complex) every electrical input port to assigned to nework analyzer
modulation formats require the optical output port and vice ports 2 and 3 (see Figure 22).
balanced devices, often having versa as well as the balanced For EO measurement the single
one optical port and two electri- characteristic from the differen- ended (SE) port 1 is assigned
cal ports. Examples are ROSA tial electrical ports to the single- to network analyzer port 4 and
(receive optical sub-assembly) ended optical port. the balanced (BAL) port 2 is
which is a photodetector inte- assigned to network analyzer
grated with a transimpendance Port configuration ports 2 and 3 (see Figure 23).
amplifier with dual electrical
outputs and dual drive LiNbO3 For balanced measurement, it The network analyzer now al-
modulators for 43 Gb/s low is required to assign network lows to display traces of inter-
drive voltage modulation. analyzer (physical) ports to est for balanced devices like
logical ports (see Figures 22 differential responsivity S, com-
Usually these balanced devices and 23). The logical port con- mon mode rejection or port im-
are 3-port devices. These elec- figuration is as follows: (logical) balance. Since opto-electronic
trical ports operate in push-pull port 1 is assigned to the single components are uni-directional,
operation, namely with 180° ended, unbalanced port and non-reciprocal devices only the
phase offset. The multiport (logical) port 2 is assigned to measurement types encircled in
the balanced, differential port.
20
Figure 22. Logical port mapping for OE measurement
Differential mode
Single-ended Common mode
Port 2
Port 1
(balanced, BAL)
(unbalanced, SE)
Balanced OE
21
Optical receiver
measurement
Photo 3.3 V
1 2
3 4 current power
meter supply
A B
L R1
ROSA
Optical fixture
attenuator
22
Measurement procedure and The trans-impedance is defined
interpretation in Equation 2. The photo detec-
Figure 26 shows a typical tor quantum efficiency is mea-
measurement screen for the sured with the current meter.
differential efficiency, elec- The equation editor of the PNA
trical port return loss, trans- allows to directly calculate and
impedance of each individual display the transimpedance of
port and normalized common each individual OE path. The
mode rejection. encircled section on the lower Figure 26. Attenuation level setting
screen of Figure 25 demon-
strates an imbalance between
50 Ω the transimpedance gains of
T= η SOE
DC port A and B, resulting in strong
reduction of common mode
Equation 2. Trans-impedance definition rejection.
Differential efficiency
[dB(A/W)]
S22 S33
Transimpedance
[dB(Ω)] output A, B
Imbalance
23
< -15 dBm
-10 dBm
Trans-impedance [dB(Ohm)]
-5 dBm
Frequency (Hz)
Figure 28. ROSA trans-impedance for various optical input power levels
Electrical Component
Measurements (E/E)
Lightwave component analyzers
have the capability to operate as
RF and microwave network ana-
lyzers. They can then be used
to characterize the electrical
components used in lightwave
systems including amplifiers,
filters, couplers etc.
24
Appendix 1. Signal Relationships in Opto-Electronic Devices
P1 ∆P P2
E/O O/E
50 Ω ∆I1 ∆I2
PO
50 Ω 50 Ω
50 Ω ∆PO
RS ( W ) Rr ( A )
A W
t
Figure 29. Signal definitions
25
For the model, the following S-parameters are defined:
b1
S11 = (a = 0)
a1 2
b2
S22 = (a = 0)
a2 1
Where:
∆V1
a1 = ► Incident on E/O device
√¯¯¯
Z 0
= ∆I1 • √¯¯¯
Z0
∆V1
b1 = ► Reflected from E/O device
√¯¯¯
Z 0
∆V2
a2 = ► Incident on O/E device
√¯¯¯
Z 0
∆V2
b2 = √¯¯¯ ► Transmitted from O/E device
Z0
= ∆I2 • √¯¯¯
Z0
b2
S21 = (a = 0)
a1 2
26
∆P0
RS = = E/O source responsivity
∆Ii
and
∆I2
Rr = = O/E receiver responsivity
∆P0
Using the above relationships, we can rewrite S21 in terms of the
transducer responsivities RS and RR:
b2
S21 =
a1
∆I2
=
∆I1
(Rr • ∆P)
=
(∆P/RS )
= (RS • Rr)
27
The responsivities Rs and Rr need to be related to some value in order to have
meaning as individual quantities expressed logarithmically, just as 0 dB rep-
resents an S21 of unity or gain of 1. Consequently source responsivity will be
expressed in Watts per Amp, which in decibels will be related to a conversion
efficiency of 1 W/A. Similarly, receiver conversion efficiency will be relative to
1 A/W.
Rs (W/A) Rr (A/W)
20 log10 |Rs • Rr | = 20 log10 • 1 (A/W)
1 (W/A)
Rs (W/A)
Rs (dB) = 20 log10
1 (W/A)
Rs (W/A)
Rs (dB) = 20 log10
1 (W/A)
28
Appendix 2 Operation in the Time Domain
Basic considerations
29
Range and resolution
Measurement range
Measurement resolution
(Time span)
MR = (N-1)
MR = (
(0.003 • Phase uncertainty (deg)
Aperature (Hz) )
where the aperture is the measurement frequency range. Phase
uncertainty will vary depending upon the type of measurement
made, but typically is better than 2 degrees, which leads to sub-
pico-second time uncertainties and sub-millimeter distance
uncertainties.
30
Response resolution
31
Agilent Email Updates Remove all doubt www.agilent.com
www.agilent.com/find/emailupdates Our repair and calibration services For more information on Agilent Technologies’
Get the latest information on the will get your equipment back to you, products, applications or services, please contact
performing like new, when prom- your local Agilent office. The complete list is
products and applications you select.
ised. You will get full value out of available at:
your Agilent equipment through- www.agilent.com/find/contactus
out its lifetime. Your equipment
Agilent Direct will be serviced by Agilent-trained Americas
technicians using the latest factory Canada (877) 894-4414
www.agilent.com/find/agilentdirect
calibration procedures, automated Latin America 305 269 7500
Quickly choose and use your test
repair diagnostics and genuine parts. United States (800) 829-4444
equipment solutions with confidence.
You will always have the utmost
confidence in your measurements. Asia Pacific
Australia 1 800 629 485
Agilent Agilent offers a wide range of ad- China 800 810 0189
Open Hong Kong 800 938 693
ditional expert test and measure-
www.agilent.com/find/open ment services for your equipment, India 1 800 112 929
Agilent Open simplifies the process including initial start-up assistance, Japan 0120 (421) 345
of connecting and programming onsite education and training, as Korea 080 769 0800
test systems to help engineers well as design, system integration, Malaysia 1 800 888 848
design, validate and manufacture and project management. Singapore 1 800 375 8100
electronic products. Agilent offers Taiwan 0800 047 866
open connectivity for a broad range For more information on repair and Thailand 1 800 226 008
of system-ready instruments, open calibration services, go to:
industry software, PC-standard I/O Europe & Middle East
and global support, which are www.agilent.com/find/removealldoubt Austria 0820 87 44 11
combined to more easily integrate Belgium 32 (0) 2 404 93 40
test system development. Denmark 45 70 13 15 15
Finland 358 (0) 10 855 2100
France 0825 010 700*
*0.125 € fixed network rates
Germany 01805 24 6333**
www.lxistandard.org **0.14€/minute
LXI is the LAN-based successor to Ireland 1890 924 204
GPIB, providing faster, more efficient Israel 972-3-9288-504/544
connectivity. Agilent is a founding Italy 39 02 92 60 8484
member of the LXI consortium. Netherlands 31 (0) 20 547 2111
Spain 34 (91) 631 3300
Sweden 0200-88 22 55
Lightwave component analyzer Switzerland (French) 41 (21) 8113811(Opt 2)
Switzerland (German) 0800 80 53 53 (Opt 1)
online information: United Kingdom 44 (0) 118 9276201
Other European Countries:
www.agilent.com/find/lca www.agilent.com/find/contactus
Revised: October 24, 2007