Frequently
Asked Questions
Polarization Controllers
1. How many fiber loops are needed in each waveplate?
2. What options are available regarding the values of each
waveplate?
3. Why does the FPC1/2/3 and MPC1 have 3 waveplates?
4. Is full coverage of the Poincaré sphere achieved using only
the ¼ and ½waveplate combination?
5. Each of the controller's waveplates are described as
'approximate' ¼ and ½waveplates. Textbook proofs are show with exact ¼ and
½ waveplates. How do these nonexact waveplates affect the coverage of the
Poincaré sphere?
6. Is there a significant difference between these waveplate
configurations? In other words, why chose one configuration over another?
7. What is the resolution on the surface of the Poincaré sphere?
8. What extinctionratio can be achieved with the MPC1?
9. How does FiberControl's 2channel MPC1, having a total of 6
independently controllable quarterwaveplates with a resolution of 0.15 deg each,
compare to Agilent's 11896 (4Qwaveplates design with 0.18 deg each)?
Polarization Scramblers
Polarization Scrabmlers
1. What are the primary differences between the PS700 and
the RCPS600/B?
2. Do either of these scramblers have GPIB control?
3. What version of LabView are the drivers?
4. I have two wavelengths one at 1300nm and the other at
1550nm. Can I use one PS700 scrambler?
5. Can the PS700 be used at 980nm?
6. Can the PS700 be used to perform PDL measurements?
7. Do you have any reliability data on your fiber coils wound
on the PZT cylinder? Or more specifically, if your have FIT data from your
polarization scrambler that is related to the fiber coil/PZT assembly based
on the units in the field.
8. What is the PMD and PDL for the PS700?
9. What is meant by the AM Modulation specification which says
< 3 percent?
10. What is the DOP dependence on wavelength for a given
configuration of the scrambler?
· Polarization Controllers:
Q1.) How many fiber loops are needed in each waveplate?
A1.) The number of fiber loops depends on the desired amount of
birefringence. See the following curves that correspond to your device and
wavelength.
Q2.) What options are available regarding the values of
each waveplate?
A2.) With the FPC1/2/3, the enduser can configure the waveplates
approximately as ¼:¼:¼ or ¼:½:¼. The standard waveplate configurations for
the MPC1 are either an approximate ¼:¼:¼wave or ¼:½:¼wave.
Q3.) Why does the FPC1/2/3 and MPC1 have 3 waveplates?
A3.) Three waveplates are necessary to transform any arbitrary input
statesofpolarization (SOP) into any other arbitrary output SOP.
Q4.) Is full coverage of the Poincaré sphere achieved using
only the ¼ and ½waveplate combination?
A4.) Yes, but only for certain well defined input SOPs. If the input SOP is
arbitrary, and a transformation to any other SOP is desired, then a third
waveplate is required (i.e., another ¼waveplate). Since three identical
¼waveplates is also an acceptable solution, all controllers have the option
of being configured as either ¼:¼:¼ or ¼:½:¼ waveplates.
Q5.) Each of the controller's waveplates are described as
'approximate' ¼ and ½waveplates. Textbook proofs are show with exact ¼ and
½ waveplates. How do these nonexact waveplates affect the coverage of the
Poincaré sphere?
A5.) Since the trigonometric relationships describing the cascade of
waveplates greatly simplify when the plates are ideally ¼ and ½wave,
elegant proofs can be obtained. However, for a set of three identical
waveplates, it can be shown that complete coverage of the Poincaré sphere is
obtained, for any arbitrary input SOP, provided the plates reside within the
following range: 0.171l < waveplate < 0.342l, mod l/2. Since the
mathematics for nonexact ¼ and ½waveplates is quite protracted, the
proof's pedagogical intent would be obscured.
Q6.) Is there a significant difference between these
waveplate configurations? In other words, why chose one configuration over
another?
A6.) In some cases, it may simply be a matter of personal preference. In
others, such as in maximizing SOP resolution, the ¼:¼:¼wave configuration
would be preferred. For the case of maximizing uniform SOP coverage around
the Poincaré sphere in the shortest period of time (e.g., using the MPC1's
scramble mode), the ¼:½:¼waveplates would be favored.
Q7.) What is the resolution on the surface of the Poincaré
sphere?
A7.) Because of the inherent factoroftwo relationship between the rotation of
an electricfield vector and the rotation of a waveplate, the necessity for
fine waveplate rotational resolution is paramount to access regions of
Stokes' space. See A9, for a more detailed comparison with a competitor's
controller.
Q8.) What extinctionratio can be achieved with the MPC1?
A8.) Customers have reported extinctionratios in excess of 70 dB for the
polarized portion of the signal.
Q9.) How does FiberControl's 2channel MPC1, having a
total of 6 independently controllable quarterwaveplates with a resolution of
0.15 deg each, compare to Agilent's 11896 (4Qwaveplates design with 0.18 deg
each)?
A9.) In comparing the angular resolution of a polarization controller based
on rotating waveplates or their analogs, there are two important issues that
must be considered: (1) worstcase behavior vs. averagecase behavior, and
(2) the likelihood of occurrence of nearworstcase behavior.
If Agilent is claiming 0.03 degrees resolution for their unit, based on
4Qwaveplates with 0.18 degrees/step mechanical resolution, then that claim
cannot be related to worstcase performance. Worstcase performance must be
about 8 times worse than that, by some simple geometric reasoning. Worstcase
optical resolution is always somewhat greater than the minimal mechanical
resolution. Since the inherent mechanical resolution of the Agilent is 20%
worse than the FiberControl MPC1, its worstcase resolution must also be 20%
worse. The precise effective resolution for either unit depends on the input
SOP as can be seen by the nonuniform transversal in Stokes' 3space around
the figureeight or tear shape.
Secondly, the likelihood of having nearworstcase behavior is inversely
proportional to the number of Qwaveplates in use for adjustment. For
Agilent's 4Qwaveplates, the probability of hitting one of these pathological
regions in control space is quite low. But with the MPC1 2 channel's six
quarterwave plates in play, the chances of being near worstcase control
regions simultaneously is reduced by 100x and there is more than enough
control redundancy to move away from such a region without changing the
output SOP.
The poorer results observed so far with the first demo MPC1 were due to
two factors:
a. each single channel of the MPC1 2 channel, operating without GPIB, had
only three waveplates compared to the Agilent's unit.
b. the demo MPC1 was supplied with the default QHQ configuration rather
than QQQ and had only two quarterwaveplates in play, rather than three.
Thus, it was two orders of magnitude more likely to be operating near a
pathological control region than was the Agilent device.
In summary, while a singlechannel QQQ MPC1 unit will be comparable to
the Agilent QQQQ unit in average resolution, given the MPC1's inherently
finer mechanical precision, it will encounter pathological control regions
more frequently, due to the lack of a fourth waveplate. Therefore, to exceed
every performance metric of the Agilent unit in static adjustment
applications, FiberControl recommends a dualchannel MPC1, with each of the
two channels configured as QQQ.
As a side note, the QHQ configuration is preferred for dynamic tracking or
generation of polarization states, while a QQQ configuration gives
uniformly finer control when performing static adjustments over small regions
of Stokes' 3space.
· Polarization Scramblers:
Q1.) What are the primary differences between the PS700 and
the RCPS600/B?
A1.) The PS700 is an industrial grade optical polarization scrambler and is
suitable for the lab and manufacturing environment  all that need be done
for optical scrambling is: plug it in, turn it on, and connect the optical
signals. On the other hand, the RCPS600/B is an inexpensive grade where time
and ease of operation is not critical since it requires the simultaneous
control and adjustment of an r.f.signal drive level/frequency and the input
SOP.
Q2.) Do either of these scramblers have GPIB control?
A2.) Yes, the PS700 has an IEEE 488.2 GPIB. Its default mode is with
scrambling on. By utilizing FiberControl's LabView drivers, simple GPIB
commands can disengage the polarization scrambling. The RCPS600/B can be
controlled via GPIB only if the function generator has a GPIB (i.e.,
HP33120).
Q3.) What version of LabView are the drivers?
A3.) FiberControl's supplied drivers are built in LabView version 6.0i and
7.0.
Q4.) I have two wavelengths one at 1300nm and the other at
1550nm. Can I use one PS700 scrambler?
A4.) Yes. Both wavelengths can be scrambled when the PS700 is configured for
the telecom wavelength. In fact, it's possible to scramble both wavelengths
simultaneously.
Q5.) Can the PS700 be used at 980nm?
A5.) Yes, provided the PS700 was manufactured with fiber of this wavelength
range. As above, multiple wavelengths can also be scrambled simultaneously.
Q6.) Can the PS700 be used to perform PDL measurements?
A6.) Yes. Its inherent ultralow insertionloss (< 1.5dB) and PDL (<
0.002dB) are extremely attractive for this application.
Q7.) Do you have any reliability data on your fiber coils
wound on the PZT cylinder? Or more specifically, if your have FIT data from
your polarization scrambler that is related to the fiber coil/PZT assembly
based on the units in the field.
A7.) While FiberControl maintains no official FIT data, we have sold hundreds
of PS700 scramblers since September 2000 and have had only one unit with any
reported fiber breaks.
Q8.) What is the PMD and PDL for the PS700?
A8.) When the PS700 is off, the PMD of the unit is extremely low…approaching
that of a 100meter length of fiber by itself. When the unit is on and
scrambling, the PMD is high. The PDL (< 0.002dB) is the same when on or
off.
Q9.) What is meant by the AM Modulation specification which
says < 3 percent?
A9.) The 3% residual AM modulation refers to change in optical intensity in
the 500kHz to 700kHz range.
Q10.) What is the DOP dependence on wavelength for a given
configuration of the scrambler?
A10.) Low DOPs can be obtained over the wavelength range of the scrambler 
from 1310 nm to 1600 nm.
