IPN Progress Report 42-182 • August 15, 2010

DOT Ground Laser Receiver: Overview and

Major Trades
Kevin M. Birnbaum,* Jeffrey R. Charles,* William H. Farr,* Jonathan Gin,*
Kevin J. Quirk,* William T. Roberts,* Jeffrey A. Stern,† and Yen-Hung Wu ‡

The Deep-space Optical Terminals (DOT) project will provide high-data-rate optical com-
munications from planetary distances. An overview of the architecture of the ground-based
receiver subsystem is presented, along with a conceptual design of the subsystem assem-
blies. Major trades are described along with the selected approach.

I. Introduction

Optical communications has the potential to increase the achievable data rate from
spacecraft at planetary distances by orders of magnitude [1,2,3]. The Deep-space Optical
Terminals (DOT) project is intended to be the irst demonstration of bidirectional optical
communications between Earth and deep space [4]. The top-level requirement is to deliver
more than 10 times the data rate of a link-limited state-of-the-art Ka-band system [5] with
comparable mass and power burden on the host spacecraft.

The DOT system is composed of four major subsystems, as shown in Figure 1. The DOT
Mission Operations Center (MOC) controls DOT operations and performs data analysis and
archiving. The Ground Laser Transmitter (GLT) sends an uplink beam to the spacecraft [6].
This beam is used as a pointing reference (i.e., beacon) at the spacecraft, as well as carry-
ing uplink communication data. The Flight Laser Transceiver (FLT) is the DOT subsystem
mounted on the spacecraft [7]. It receives the uplink beam and transmits a downlink beam.
The Ground Laser Receiver (GLR) receives the downlink light and recovers the communica-
tion data. DOT also supports ranging by measuring the time of light on both the uplink
and downlink beam.

An overview of the DOT requirements is given in [4]. In order to perform all of the required
demonstrations, the DOT system speciies two sets of requirements on the GLR correspond-
ing to two modes of operation: a high-rate mode and a low-rate mode. The high-rate mode,
which is used only for nighttime operations, must provide a net gain of 142 dB to support
the maximum DOT downlink data rate of 267 Mb/s. The low-rate mode is used to demon-
strate link acquisition over a range of conditions, including Sun–Earth–probe (SEP) angles

* Communications Architectures and Research Section.

† Instrument Electronics and Sensors Section.
‡ Optics Section.

The research described in this publication was carried out by the Jet Propulsion Laboratory, California Institute of
Technology, under a contract with the National Aeronautics and Space Administration. © 2010 California Institute
1 of Technology. Government sponsorship acknowledged.
 CMD/TLM
(X-band/Ka-Band)
Deep
FLT
Space
Network Beacon/Uplink Science/Telemetry
(>200 kbs) (>250 Mbs)
Ranging (30 cm) Ranging (30 cm)

CMD/TLM GLT GLR

CMD/Monitor
Data Science/TLM/
CMD/Monitor Ranging Monitor
Data

Spacecraft DOT MOC

Operations Center

Figure 1. The architecture of the DOT system. The GLR accepts the downlink light from the FLT
and sends the decoded data to the DOT MOC.

down to 5 deg. In the low-rate mode, the net gain need only be 124 dB. In our trades,
discussed in more detail in Section II, we considered both using a single set of hardware to
meet all requirements, as well as using different equipment for each mode of operation.

There are four major factors that drive the design of the GLR subsystem. The irst is the
large net gain requirement. This drives the use of a large collecting area, as well as highly
eficient optics and detectors for receiving the faint signal from deep space. The second
factor is the requirement for daytime and low SEP angle operations. This is unusual for
telescopes, which are typically designed to operate only at night (or in the case of solar
telescopes, only while pointing directly at the Sun). The third factor is the low rate of signal
photons. The link operates in the photon-starved regime, which necessitates the use of
eficient modulation and error-correcting codes to maximize the bits per photon [8]. This
ultimately impacts the electronics used to receive and decode the downlink signal. The
fourth factor is the low ratio of signal photons to background photons. The detected rate
of background photons may exceed the rate of signal photons by as much as 18 dB during
low SEP operations. This increases the dificulty of performing spatial and temporal acquisi-
tion of the signal. It also makes it necessary to precisely ilter the incoming light to keep
the background rate as low as possible while minimizing the loss of the signal photons.

The reference architecture of the GLR is shown in Figure 2. There are six assemblies in the
GLR subsystem. The telescope assembly collects and concentrates the downlink light coming
from the FLT. The aft optics assembly takes the light from the telescope assembly and ilters
out the background light. It also provides functions for beam monitoring and alignment.
The downlink light is then coupled to the detector assembly, where the photons are convert-

2
 Telescope Aft Optics Assembly Element
Assembly
Filters out Signal
From Collects light Light
background light
FLT
Light Light
Acq/ Acq/ Detector
Track Track Assembly

Converts
photons into
Controls spatial and Photon electronic signals
temporal acq/track Counts

Element
Electronics
Sl
ot
St
ati
sti
cs
Telescope Aft Optics Assembly Element
Assembly
From Filters out Signal StationElectronics
Station Electronics Atmospheric
Atmospheric
Collects light background light Light Monitoring
Monitoring
FLT
Light Light Decodes and stores data Measures
Measures
Acq/ Acq/ Detector Conditions
Conditions conditions
conditions
Track Track Assembly that
that

Data
Converts affect
affect link
link
s
photons into tic
Controls spatial and tis To MOC
Photon electronic signals S ta
temporal acq/track ot
Counts Sl
Element
Electronics

Figure 2. The architecture of the GLR. The assemblies within the GLR subsystem are represented by green boxes.
Four of the six assemblies are repeated in each element (shown in gray) of the array, whereas the remaining two
are shared by all elements. An array of two elements is shown, but one or more may be used.

ed into an electronic signal. The detector assembly must distinguish between photons strik-
ing different regions of the focal plane, in order to provide the tracking information used to
stabilize the optical line of sight. The electronic signals from the detector assembly are sent
to the element electronics, which synchronizes to the downlink signaling format and esti-
mates the signal and background photo-count rates. It provides the control signals to the
telescope assembly and aft optics assembly, which are used to acquire and track the down-
link light. The slot statistics (which may be represented as the number of photo-counts in
each time interval of the downlink signal structure) are sent from the element electronics to
the station electronics. The station electronics uses the synchronized slot statistics to decode
the information that was transmitted over the downlink; it then stores the resulting data for
eventual relay back to the DOT MOC. The station electronics also relays to the DOT MOC
the atmospheric conditions that affect the link, such as sky radiance, atmospheric attenua-
tion, and turbulence. These parameters are measured by an atmospheric monitoring assembly.
The atmospheric monitoring assembly is based on a suite of previously developed instru-
ments [9–16], many of which are in operation at NASA’s Table Mountain Facility or Gold-
stone Deep Space Communications Complex.

As shown in Figure 2, the telescope assembly, aft optics assembly, detector assembly, and
element electronics may be grouped into an element. Multiple elements may be combined
in an array within the GLR. The net gain of the GLR is approximately the sum of the net
gain of the elements (minus the losses in decoding). Any number of elements may be used.
However, in this architecture each element must be capable of performing signal acquisition
and tracking independently. This imposes a lower limit on the gain of each element, given
the link conditions [17]. Note that the signals are combined electronically after they have

3
 been detected and synchronized, so that complicated beam combiners or pathlength equal-
izers are not needed.

We will now discuss the trades and baseline conceptual design of these assemblies in
more detail. The telescope assembly is treated in Section II; the aft optics assembly is in
Section III; the detector assembly is in Section IV; and the element electronics and station
electronics are in Section V.

II. Telescope Assembly

The telescope assembly points to the FLT and collects the downlink light. The requirement
for a large net gain in the GLR lows down to a requirement for a large collecting area in the
telescope assembly. After allocations are made for optical losses, the requirement of 142 dB
(124 dB) gain in the GLR is used to derive a requirement of 110 m2 (3.8 m2) collecting
area in the high-rate (low-rate) mode, corresponding to an equivalent diameter of 11.8 m
(2.2 m).

There are multiple approaches to the acquisition of the telescope assembly: renting exist-
ing facilities (possibly with some modiication), building a new facility, or a combination of
renting a facility for one mode of operation and building a facility for the other.

First we discuss the options for renting existing facilities. We identiied three candidate
facilities with suficient collecting area for the high-rate mode: the Very Large Telescope
(VLT) when at least three of its four 8.2-m telescopes are used in an array; the Keck Observa-
tory when both of its 10-m telescopes are used; and the Large Binocular Telescope (LBT).
The VLT, located in the Atacama Desert of Chile and operated by the European Southern
Observatory, would pose a number of logistic problems. For example, its location makes it
impossible to demonstrate long link durations with a GLT located in the United States. Of
the other two candidates, the LBT has the advantage of a lower cost per night of operation
and a location in the southwestern continental United States. This location is favorable
since it allows the use of existing facilities for the GLT.

The LBT, located on Mount Graham, Arizona, consists of twin 8.4-m Gregorian telescopes
attached to a single altitude–azimuth mount. Each telescope may be used in one GLR ele-
ment, eliminating the need for optical combining of the two. The LBT facility could be
used by packaging the aft optics assembly and detector assembly into visitor instruments
that are mounted at focal stations of the twin telescopes. The LBT meets all requirements
for use as the telescope assembly in high-rate mode. However, the LBT cannot meet (and
cannot reasonably be modiied to meet) the 5-deg SEP angle requirement of the low-rate
mode.1

Next we consider existing telescope facilities that may be used for the low-rate mode. We
did not identify any existing telescopes with ≥ 2.2-m aperture that currently may operate

1Tom McMahon, “Use of Large Binocular Telescope as Ground Lasercom Receiver Demonstrator,” JPL Subcontract
No. 1400570 (internal document).

4
 while pointing at 5 deg from the Sun. However, some telescope facilities may be modiied
to allow near-Sun operations. The Hale Telescope, a 5-m telescope on Palomar Mountain,
California, has a 42-m dome. With modiication to the dome opening and additional baf-
ling, the Hale Telescope could provide suficient collecting area while pointing near the
Sun [18]. The cost of modiications, however, is higher than other options discussed below.
The Hooker Telescope, a 2.5-m telescope on Mount Wilson, California, is another tele-
scope that may be bafled to allow near-Sun pointing. The optics of the Hooker Telescope,
though, are temperature sensitive and have a long time constant for thermal relaxation.
There is a signiicant risk that the telescope point-spread function would be too large to
allow enough background light rejection when the telescope is operated during the day and
across day–night transitions.

We turn now to the options for building a new facility to serve as the telescope assembly.
A new facility could be constructed to meet both the high-rate and low-rate requirements.
Such a facility would feed forward beyond the DOT demonstrations, as it could be used in
the eventual operations of deep-space optical communications.

Except for the near-Sun pointing capability, the requirements on the telescope assembly are
very relaxed compared to most astronomical telescopes. For example, the image spot size of
the telescope may be up to 20 µrad (depending on the atmospheric seeing at the telescope
site), which is much larger than the diffraction-limited performance commonly demanded
of astronomical instruments. Furthermore, the ield of view of the telescope can be as small
as 50 µrad, as the telescope must only collect light from a single point source. That source
is also monochromatic, whereas astronomical telescopes are typically designed to accom-
modate a wide range of wavelengths. These factors led us to consider approaches to the
telescope design that give a much lower cost compared to an astronomical telescope of the
same diameter.

With these considerations in mind, a number of point designs for low-cost telescopes
were created. Since a number of smaller telescopes may be combined to form an array (as
discussed in Section I), the igure of merit is taken to be the cost of a telescope assembly
divided by its effective area of light collection. Note that the cost of the telescope assem-
bly includes not only the cost of the telescope optics, but also the mounts, gimbal, dome,
site preparation work, etc. The relative igures of merit of the point designs are shown in
Figure 3.

We consider two broad types of designs: those based on a monolithic primary mirror and
those based on a segmented primary mirror. Monolithic designs with primary diameters
between 0.8 m and 8.2 m have been investigated. We have found that there is a knee in
the cost curve for these designs, such that the cost per area is nearly constant for apertures
below 2.2 m. Above this diameter, the cost per area increases nearly proportional to diam-
eter. These results are consistent with the “conventional wisdom” that telescope costs scale
between ∝ D2 and ∝ D3, depending on the design regime [19,20]. Based on this analysis,
if a monolithic primary design is chosen, then the lowest cost build approach that meets
requirements is an array of 2.2-m terminals.

5
 Relative Telescope Cost per Collection Area
2

MONOLITHIC
1

0.5

SEGMENTED

1 10
Aperture Diameter, m

Figure 3. The relative cost of building a telescope assembly per collecting area (in arbitrary units) vs. diameter of
the collecting aperture. Circles indicate point designs, while smooth curves are based on parametric scaling. The
dashed red curve is a it to the point designs based on a monolithic primary mirror. The solid blue curve is based
on the segmented mirror point design, with parametric scaling of each of its major components.

We also investigated an approach to the telescope assembly based on a large segmented pri-
mary.2 The relatively modest requirements on spot size and ield of view allow signiicant
cost savings, comparable to those found in the Hobby-Eberly Telescope [21]. A conceptual
point design was created to meet the GLR requirements with a single telescope. Parametric
scaling of the cost of each major telescope component (e.g., primary mirror, dome, gimbal,
etc.) was used to estimate the cost as a function of telescope diameter. The 12-m telescope
design is near the minimum of the cost curve, though the minimum is shallow. The design
could be scaled up or down by 50 percent in area with negligible impact on the cost per
area.

There are a number of factors to consider when comparing the telescope assembly design
based on the monolithic 2.2-m and the segmented 12-m. An array of 2.2-m telescopes
(29 would be required to meet the total collecting area requirement) has the advantage of
graceful degradation, as failure of any element would only reduce the signal by 3 percent.
Similarly, this design could easily be made more robust by adding additional elements to
introduce redundancy. Finally, there could be programmatic advantages to building the ar-
ray slowly over time, with a limited capability available at intermediate steps along the way.
On the other hand, the single-aperture approach has the advantage that it can acquire sig-
nals at a lower irradiance, enabling future missions at farther ranges or with lower-mass and
lower-power light terminals. Furthermore, the single-aperture approach has much lower
build costs and costs for maintenance and operations. Therefore, of the build options, the
approach based on the single segmented aperture is considered most favorable.

2 Thomas A. Sebring, “A Study to Illustrate the Concept and Likely Cost for a 12-Meter-Aperture Ground-Based Telescope
for Laser Communications,” JPL Subcontract No. 1400382 (internal document).

6
 Finally, we consider the option of using a combination of existing facilities along with new
facilities to meet the DOT requirements. This approach would use the LBT to satisfy the
high-rate requirements, while building a new telescope for the low-rate mode of operation.
This eliminates the costs and risks of attempting to modify an existing telescope to point
near the Sun, while avoiding the high cost of building a large-area telescope. To minimize
costs, the new telescope would have the minimum aperture of 2.2 m. Beyond the DOT
demonstrations, this new telescope would also provide an operational capability for bidi-
rectional communication to spacecraft in geosynchronous orbits, at the Moon, or at the
Lagrange points. The telescope could also be adapted to serve as an uplink station for deep
space.

Based on the DOT project goals, including minimizing cost and risk, as well as providing
feed forward to future capabilities, the baseline approach chosen was that of renting the
LBT and building a new 2.2-m telescope with near-Sun pointing capability. For the initial
operational capability, after successful demonstrations of the DOT terminals, we recom-
mend pursuing the ground receiver approach based on a large segmented primary mirror
telescope.

III. Aft Optics Assembly

The aft optics assembly relays the signal light from the telescope assembly to the detector
assembly while rejecting the background light. It ilters the light by controlling the allowed
polarization, wavelength, and spatial mode.

The primary trade in the design of the aft optics assembly is the choice of technology for
spectral iltering. The spectral ilter needs to have a high eficiency at the desired wave-
length and a narrow bandwidth (noise equivalent bandwidth ≤ 0.17 nm), and accept a large
etendue. The etendue is the product of the area and solid angle of the rays crossing a plane,
and is invariant throughout the optical system. The etendue of the GLR system is 3.9 ×
10–8 m2sr in the high-rate mode.

There are three candidate approaches for the spectral ilter. The irst is an optic with a
multilayer dielectric coating. Optical bandpass ilters are commonly made this way, but this
approach does not work eficiently for the extremely narrow passband required. The second
approach is an etalon ilter, based on two parallel lat relectors. This technique can achieve
the required bandwidth, but has a number of drawbacks, including poor out-of-band rejec-
tion, relatively low peak transmission, high angular selectivity (requiring a large ilter to
meet the etendue requirement), and high sensitivity to thermal and mechanical disturbanc-
es. The third approach is based on a relective volume Bragg grating (VBG) [22,23]. This is
a bulk piece of glass that has been patterned with an index of refraction that varies peri-
odically in the direction of propagation. It is highly relective at the ilter wavelength and
highly transmissive at other wavelengths. The ilter wavelength can be adjusted through
thermal tuning. The VBG was selected for the baseline design because of its narrow band-
width, large acceptance angle (permitting a design with a large etendue with small optics),
and high eficiency.

7
 The architecture of the aft optics assembly is shown in Figure 4. The light from the tele-
scope assembly passes through a coarse ield stop. This stop does not deine the ield of
view, but is simply used to reduce the stray light entering the system. The light is then
collimated and passed through a coarse dual bandpass ilter. One passband is at the signal
wavelength, and the other is at a visible wavelength for alignment light. The light then
relects off a ine steering mirror placed at a pupil plane. This mirror is controlled by the
element electronics to compensate for pointing errors in the telescope assembly. The light
is then refocused. A dichroic splitter is used to divert the alignment light to a separate
camera. This dichroic optic, though highly transmissive at the signal wavelength, can also
be used to couple in light from a iber injection port. The transmitted signal light comes to
a focus on an adjustable ine ield stop, which deines the ield of view of the system. The
light is then collimated and sent through a quarter-wave plate and a half-wave plate, each
on an independent rotation stage. This combination allows any polarization of light to be
converted to any desired linear state. The light continues on to a polarizing beam splitter.
The signal light passes through, while the portion of the background light in the orthogo-
nal polarization is diverted to a power meter via a bafled path. The signal light then passes
through a quarter-wave plate (whose axes are set at 45 deg) and strikes the VBG. The light
outside the spectral band passes through the grating and goes along a bafled path to anoth-
er power meter. The signal light relects off of the grating and returns through the quarter-
wave plate. The linear polarization is rotated through 90 deg by the two passes through the
quarter-wave plate, so the light now relects off of the polarizing beam splitter. It continues
on to a rotatable linear polarizer, used to attenuate the beam during testing and calibration.
The light then passes through a rotatable half-wave plate and strikes another polarizing
beam splitter. This combination makes effectively a splitter with a variable ratio of transmis-

Coarse Dual
Field Bandpass
Stop Filter Fine
Input Steering
Mirror

Collimating
Lens
Focusing
Power Meter Lens

Polarizing Alignment
Beam Camera
Splitter Fine
Collimating Field Dichroic
Lens Stop Splitter
Power
Meter Fold
Narrow Mirror
Band Filter Imaging
λ/4 Rotatable
Fiber Lens
λ/4 and λ/2
Rotatable Injection
Polarizer and λ/2
Output
Polarizing Focusing
Beam Splitter Lens

Monitor
Port

Figure 4. The architecture of the aft optics assembly. The path marked “Input” comes from the telescope assem-
bly, and the path marked “Output” goes to the detector assembly. The optic labeled “Dual Bandpass Filter” trans-
mits light at the signal wavelength and at a visible wavelength used for alignment. The cube marked “Dichroic
Splitter” transmits the signal wavelength and relects the alignment wavelength.

8
 sion to relection. This is used to divert a controlled fraction of the light to a monitor port,
used during testing and calibration. The light inally exits the aft optics assembly and enters
the detector assembly.

This architecture has a number of beneicial features. It provides several ports for checking
the spatial, polarization, and spectral proile of the input light. It has a port for injecting
a known test signal light. It has multiple stages of iltering and can be bafled to minimize
stray light. Additionally, this design can be adapted for use with different telescope assem-
blies simply by changing lenses.

IV. Detector Assembly

The detector assembly takes the light from the aft optics assembly and converts the pho-
tons into an electronic signal. The driving requirements for the detector assembly are detec-
tion eficiency >50 percent; dark count rate <330 kHz; timing jitter <120 ps; and etendue
>3.9 × 10–8 m2sr. The etendue requirement can be converted to a requirement on area given
the maximum angle of incidence for eficient detection. For example, a 274-µm-diameter
circular detector that detects light up to 27 deg (f/1 beam) would meet the requirement.

The detector must have an array format. A minimum of three pixels must be spatially sepa-
rated in the focal plane, so that a feedback signal can be extracted for tracking the line of
sight to the FLT. Furthermore, the high required etendue can only be achieved by arraying
many pixels, given current technological limitations. Finally, spreading the light over many
pixels is required in order to avoid blocking losses due to the saturation of each pixel, given
the high count rates expected in the link. This last factor, however, is less stressing than the
high etendue requirement for the technologies under consideration.

The major trade in the detector assembly is the choice of detector technology. Table 1
shows some of the candidate technologies for detection of light at the downlink wave-
length of 1550 nm. As shown in Table 1, no existing, demonstrated detector system pres-
ently meets the GLR requirements.

The superconducting nanowire single-photon detectors (SNSPDs) offer the best combina-
tion of high timing resolution, high saturation rate, and high detection eficiency [24,25];
individual pixels and small arrays have demonstrated the required performance charac-
teristics. However, the small pixel sizes imply that arrays of several hundred elements are
needed to meet GLR requirements. Thus, technology development and signiicant cryogen-
ic engineering will be needed to create the required large-format arrays.

The InGaAs devices are available in large arrays with suficient collecting area [26]. How-
ever, the performance of each pixel does not meet requirements. In particular, the timing
jitter is too high. Low jitter has been demonstrated in single-pixel InGaAs devices [27], but
only under conditions where the dark count rate was unacceptably high.

The intensiied photodiode (IPD) is a relatively mature technology that meets nearly all
requirements [28]. The only limitation is the low detection eficiency of 30 percent at
1550 nm.

9
 Table 1. The detector technology trade table. Performance is color-coded,
with green most favorable and red least favorable.

Downlink Wavelength: 1550 nm

Geiger Mode NAF InGaAs Photocathode Linear Superconducting
Technology→ Ge InGaAs InGaAs PMT[1] IPD[2] InGaAs HgCdTe Nanowire TES[3]
Detection Eficiency
@ 1550 nm 15% 45% 23% 15% 30% 8% 60% 90%
Dark Rate/
mm2
Timing Resolution
(1 sigma) 270 ps 240 ps 1000 ps 90 ps 50 ps
Pixel
Area >1 mm dia >1 mm dia 100 µm2
Operating 80– 200– 200–
Temperature 180 K ~220 K ~220 K 220 K 300 K 300 K 80K <4 K <0.3 K
Array Size
(To Date) 1 256×256 8×8 16 52 >128×128 >128×128 4×4
TRL[4]
for GLR 2 2 2 2 6 2 2 4 2
Can Meet
Requirements? No No No No No No No Maybe No
[1] PMT = photomultiplier tube
[2] IPD = intensiied photodiode

[3] TES = transition-edge sensor

[4] TRL = technology readiness level

Based on this state of the art, the baseline concept of the GLR is to develop large arrays of
SNSPDs for the GLR detector assembly. The InGaAs photon counting arrays will continue
to be considered as an alternative as technology development improves the timing resolu-
tion. The IPD will be kept as a low-risk backup option; however, meeting GLR requirements
using the IPD-based detector assembly will require a larger collecting area in the telescope
assembly. The backup option would use the Keck Telescopes instead of the LBT to meet the
GLR requirements.

The development of large SNSPD arrays will require several issues to be addressed. The yield
of individual pixels will need to be increased in order to achieve the required high average
detection eficiency. A closed-cycle cryostat operating at ~3 K will need to be engineered;
it must support eficient free-space coupling, and must be able to feed the hundreds of
high-bandwidth signals to room temperature. This will likely require cryogenic front-end
processing (pixel combination) to reduce the number of wires, and hence thermal con-
ductance, from room temperature to 3 K. Another area of development is bias control and
monitoring for the large arrays. Relative propagation delays across the array will also need
to be addressed. The array will also require a precisely aligned microlens array to create a
high equivalent ill factor.

V. Electronics Assemblies

The element electronics accepts the electronic signal from each detected region of the focal
plane. It determines the number of photons received in each temporal slot in each region.
It synchronizes to the downlink signal, and estimates the rate of signal and background
photons. The element electronics also controls the acquisition and tracking of the down-
link. The station electronics decodes the telemetry data based on the synchronized slot

10
 statistics from the element electronics. The electronics assemblies also provide monitor and
control functions for the GLR.

The electronics are allocated 1.5 dB implementation loss, in addition to the 1.2 dB gap to
capacity for the serially concatenated pulse-position modulation (SCPPM) code [8]. The
electronics must meet this requirement across the full range of signaling parameters [4],
including all operating points from 13 kb/s to 267 Mb/s. The signaling format has vari-
able PPM orders, code rates, slot widths, and symbol repetitions to cover the wide range of
signal and background photon rates. The electronics must be lexible enough to reconigure
to any operating point within 5 min.

The major trade in the electronics is the choice of whether to use commercial off-the-shelf
(COTS) hardware or custom hardware. COTS products were the basis for the planned Mars
Laser Communications Demonstration (MLCD) Project [29,30], which used similar signal-
ing but at much lower data rates. The COTS platform, however, still required development
of custom interfaces. It was also found to be more expensive to maintain than an internally
developed platform because of the short life cycle of the commercial products. A custom
hardware platform was developed under a NASA Interplanetary Network Directorate (IND)
technology program to ill the needs for reception of high-rate optical communications
signals from deep space. These products, which have been successfully demonstrated in
emulated links [31], are the best match to the DOT GLR requirements and were selected for
the baseline approach.

The architecture of the GLR electronics is shown in Figure 6. This architecture is scalable to
data rates over 1 Gb/s, and accommodates both a single-element or arrayed architecture for
the GLR. There are six major subassemblies shown in the igure: the programmable oscilla-
tor, receiver, channel combiner, and signal acq/track controller within the element elec-
tronics; and the channel combiner/de-interleaver and decoder in the station electronics.

The programmable oscillator is a frequency synthesizer with an output phase that tracks an
arbitrary function. It is used to compensate for the time-base distortion between the space-
craft and the ground station, based on the predicted Doppler proile. It accepts reference
frequency and time from the station and sends the synthesized frequency to the receivers.

The receiver samples the output of the photon-counting detector, synchronizes to it, and
performs estimation of the signal and background rates. It uses a custom mixed-signal (RF/
digital) offset phase-locked-loop circuit to control the phase of the clock that samples the
photon-counting detector output. Precise placement of the position of the sample times
facilitates synchronization and detection without the use of multiple samples per slot, en-
abling slot rates up to 6.4 GHz. Each receiver has 10 Gb/s of input/output (I/O) bandwidth
over standard iber optics, which it uses to send the synchronized slot statistics and esti-
mates to the channel combiner.

The channel combiner is used to combine the synchronized receiver outputs prior to in-
formation decoding. Each channel combiner has 20 Gb/s I/O over iber optic connections.
Channel combiners may be replicated and connected in a tree to combine an arbitrary
number of receivers.

11
 Completed Hardware

Receiver

Te
Triad Decoder

le
Channel
Receiver

sc
Detector Combiner

o pe
Receiver Decoder

Signal Acq/ Programmable

Track Oscillator
Channel
Combiner/
De-
Interleaver
Receiver

Decoder
Te

Triad
le

Channel
Receiver
sc

Detector Combiner
op
e

Decoder
Receiver

Signal Acq/ Programmable

Track Oscillator

Element Electronics Station Electronics

Figure 6. The architecture of the GLR electronics. The subassemblies in the dashed box have been built and are
shown on the right. Two elements of a GLR array are shown. The subassemblies belonging to one instance of
the element electronics are shown in the dotted box on the lower left; those belonging to the station electronics
are shown in the dotted box in the center. Three receivers are shown for each detector, assuming the detector is
partitioned into three parts, with each part connected to one receiver. The ratio of signal intensity on each part of
the detector is used by the signal acq/track controller to feed back to the telescope pointing. The decoders are
connected in a chain, with a length that can be increased in order to accommodate a higher data rate.

The channel combiner on each element of the GLR array passes the signal photon lux
estimates to the signal acq/track controller. The latter subassembly controls the acquisition
and tracking of the signal light by adjusting the pointing of the telescope assembly and aft
optics assembly. This architecture is based on a detector that is divided into at least three
regions in the focal plane, with each detector region connected to a receiver. The receiver
estimates the signal lux detected in that region, and the signal acq/track controller uses the
map of signal lux across the regions to compute a centroid. The signal acq/track controller
also controls the spectral ilter wavelength, ield stop size, and state of polarization in the
aft optics assembly, which it adjusts to maximize the signal margin in the link.

The inal channel combiner in the tree also functions as a de-interleaver. Interleaving is
used in the signaling format to mitigate the effect of fading in the link. The de-interleaver
will be implemented on a daughter card with a large memory (>2 GB), which will plug in to
the existing channel combiner board.

The de-interleaved data are then sent to a chain of decoders. The decoders have a scalable
architecture, allowing an arbitrary number of decoder cores to be implemented by extend-
ing the chain. Each decoder accepts encoded data if it is idle, or passes it down the chain if

12
 it is busy. The decoded bits from each decoder are packetized, aggregated across a network,
and restored to their original ordering. Using the current SCPPM decoder core with four
iterations per codeword, each decoder can process ~150 Mb/s. The decoder has 20 Gb/s I/O
over iber. A centralized server accepts the decoded data frames over gigabit Ethernet via
User Datagram Protocol/Internet Protocol (UDP/IP) and performs the reordering.

The electronics subassemblies accept standard interfaces for time and frequency distribu-
tion. The master time for the station is set by a commercial reference source, such as a
GPS-disciplined crystal oscillator. Commercial time-code generators and translators are
used to transfer time from the station to each element. Precise comparison of the received
clock on the downlink signal to the station time is used in computation of the range to the
spacecraft.

Each subassembly also supports monitor and control through a gigabit Ethernet interface.
A custom graphical user interface (GUI) has been created for monitor and control, data
logging, and system characterization, with modules created for the receiver and decoder.
Additional modules will be developed and integrated to provide a uniied interface.

VI. Conclusion

Trades were performed for each assembly within the GLR, and the selections led to a base-
line concept. This concept uses the LBT to support the high-rate mode of operation, and
builds a new 2.2-m telescope with near-Sun pointing capability for the low-rate mode. The
concept uses a common design in the optics behind the telescope, using a narrow spec-
tral ilter based on VBG technology and standard components. The downlink signal light
is detected by SNSPDs, which will require new technology development. The electronics,
irmware, and software are based on products previously developed under a NASA technol-
ogy program. Except for the photon-counting detector array of the required size, all of the
technology is currently available. A backup option using only existing technology (based
on the IPD detector and the Keck Telescopes) was also identiied. All DOT requirements can
be met, and a plan and schedule have been drafted for completing the GLR to support a
light system launching in 2018.

While addressing the requirements for DOT, we have also identiied a path forward to an
operational capability. The GLR concept is scalable to greater ranges or higher data rates.
The architecture allows higher gain to be achieved by arraying receiver elements. Large seg-
mented telescopes have been identiied as a low-cost approach to the large collecting area
needed for deep-space receivers. We have baselined technology development of large arrays
of superconducting photon counting detectors, which have demonstrated high eficiency
in single devices. The electronics are based on a scalable architecture to support higher data
rates and larger arrays. Together, these concepts will enable unprecedented data rates from
planetary distances, which will allow more scientiic data to be returned, enhancing the
value of each mission that uses optical communications.

13
 References

[1] H. Hemmati, ed., Deep Space Optical Communications, JPL Deep Space Communications
and Navigation Series, New Jersey: Wiley & Sons, Inc., 2006.

[2] A. Biswas, D. Boroson, and B. Edwards, “Mars Laser Communication Demonstration:

What It Would Have Been,” Proceedings of SPIE, edited by G. S. Mecherle, vol. 6105,
p. 610502, 2006. http://dx.doi.org/10.1117/12.669551

[3] H. Hemmati, A. Biswas, and D. Boroson, “Prospects for Improvement of Interplanetary

Laser Communication Data Rates by 30 dB,” Proceedings of the IEEE, vol. 95, pp. 2082–
2092, October 2007. http://dx.doi.org/10.1109/JPROC.2007.905057

[4] A. Biswas, H. Hammatti, S. Piazolla, B. Moision, K. Birnbaum, and K. Quirk, “Deep-

space Optical Terminals (DOT) Systems Engineering,” to be published in The Interplan-
etary Network Progress Report, vol. 42-183, Jet Propulsion Laboratory, Pasadena, Califor-
nia, November 15, 2010.

[5] J. Taylor, D. K. Lee, and S. Shambayati, “Mars Reconnaissance Orbiter Telecommunica-

tions,” DESCANSO Design and Performance Summary Series 12, Jet Propulsion Laboratory,
Pasadena, California, September 2006.
http://descanso.jpl.nasa.gov/DPSummary/MRO_092106.pdf

[6] W. T. Roberts and M. W. Wright, “Deep-space Optical Terminals (DOT) Ground La-
ser Transmitter (GLT): Trades and Conceptual Point Design,” to be published in The
Interplanetary Network Progress Report, vol. 42-183, Jet Propulsion Laboratory, Pasadena,
California, November 15, 2010.

[7] W. H. Farr, M. W. Regher, M. W. Wright, A. Sahasrabudhe, J. W. Gin, and D. H. Nguy-

en, “Overview and Trades for DOT Flight Laser Transceiver,” to be published in The
Interplanetary Network Progress Report, vol. 42-184, Jet Propulsion Laboratory, Pasadena,
California, February 15, 2011.

[8] B. Moision and J. Hamkins, “Coded Modulation for the Deep-Space Optical Channel:
Serially Concatenated Pulse-Position Modulation,” The Interplanetary Network Progress
Report, vol. 42-161, Jet Propulsion Laboratory, Pasadena, California, pp. 1–25, May 15,
2005. http://ipnpr.jpl.nasa.gov/progress_report/42-161/161T.pdf

[9] P. W. Nugent, J. A. Shaw, and S. Piazzolla, “Infrared Cloud Imaging in Support of Earth–
Space Optical Communication,” Optics Express, vol. 17, no. 10, pp. 7862–7872, 2009.
http://dx.doi.org/10.1364/OE.17.007862

[10] G. E. Shaw, “Sun Photometry,” Bulletin of the American Meteorological Society, vol. 64,
no. 1, pp. 4–10, 1983.
http://dx.doi.org/10.1175/1520-0477(1983)064%3C0004:SP%3E2.0.CO;2

[11] J. M. Beckers, “A Seeing Monitor for Solar and Other Extended Object Observations,”
Experimental Astronomy, vol. 12, pp. 1–20, 2001.
http://dx.doi.org/10.1023/A:1015712720291

14
 [12] A. Tokovinin and V. Kornilov, “Accurate Seeing Measurements with MASS and DIMM,”
Monthly Notices of the Royal Astronomical Society, vol. 381, no. 3, pp. 1179–1189, 2007.
http://dx.doi.org/10.1111/j.1365-2966.2007.12307.x

[13] J. C. Wyngaard, Y. Izumi, and S. A. Collins, Jr., “Behavior of the Refractive-Index-

Structure Parameter Near the Ground,” Journal of the Optical Society of America, vol. 61,
no. 12, pp. 1646–1650, 1971. http://dx.doi.org/10.1364/JOSA.61.001646

[14] F. Hill and M. Collados, “Inverting Scintillometer Array Data to Estimate Cn2 for the
ATST Site Survey,” Bulletin of the American Astronomical Society, vol. 35, p. 848, May
2003.

[15] S. Vogt and P. Thomas, “SODAR — A Useful Remote Sounder to Measure Wind and
Turbulence,” Journal of Wind Engineering and Industrial Aerodynamics, vol. 54–55,
pp. 163–172, 1995. http://dx.doi.org/10.1016/0167-6105(94)00039-G

[16] E. J. Seykora, “Solar Scintillation and the Monitoring of Solar Seeing,” Solar Physics,
vol. 145, pp. 389–397, 1993. http://dx.doi.org/10.1007/BF00690664

[17] K. Quirk and J. Gin, “Optical PPM Combining Loss for Photon Counting Receivers,”
IEEE Military Communications Conference, pp. 1–4, October 2006.
http://dx.doi.org/10.1109/MILCOM.2006.301986

[18] W. T. Roberts, H. L. Petrie, A. J. Pickles, R. P. Thicksten, and C. Echols, “Feasibility of

Utilizing the 200-inch Hale telescope as a Deep-Space Optical Receiver,” Proceedings of
SPIE, edited by J. C. Ricklin and D. G. Voelz, vol. 5550, pp. 326–335, 2004.
http://dx.doi.org/10.1117/12.563324

[19] J. R. Lesh and D. L. Robinson, “A Cost-Performance Model for Ground-Based Optical

Communications Receiving Telescopes,” The Telecommunications and Data Acquisition
Progress Report, vol. 42-87, pp. 56–64, November 15, 1986, issue dated July–September
1986. http://ipnpr.jpl.nasa.gov/progress_report/42-87/87G.PDF

[20] T. Schmidt-Kaler and P. Rucks, “Telescope Costs and Cost Reduction,” Proceedings of
SPIE, edited by A. L. Ardeberg, vol. 2871, pp. 635–640, 1997.
http://dx.doi.org/10.1117/12.269092

[21] V. L. Krabbendam, T. A. Sebring, F. B. Ray, and J. R. Fowler, “Development and Perfor-

mance of Hobby-Eberly Telescope 11-m Segmented Mirror,” Proceedings of SPIE, edited
by L. M. Stepp, vol. 3352, pp. 436–445, 1998. http://dx.doi.org/10.1117/12.319265

[22] J. Lumeau, L. B. Glebov, and V. Smirnov, “Tunable Narrowband Filter Based on a

Combination of Fabry-Perot Etalon and Volume Bragg Grating,” Optics Letters, vol. 31,
no. 16, pp. 2417–2419, 2006. http://dx.doi.org/10.1364/OL.31.002417

[23] D. Caplan, P. Chapnik, J. Carney, M. Stevens, M. Willis, and M. Glynn, “Narrow-Band

Low-Loss Multi-Mode Spectral Filtering for Free-Space Optical Receivers,” IEEE Lasers
and Electro-Optics Society, pp. 163–164, November 2008.
http://dx.doi.org/10.1109/LEOS.2008.4688539

[24] J. Stern and W. Farr, “Fabrication and Characterization of Superconducting NbN

Nanowire Single Photon Detectors,” IEEE Transactions on Applied Superconductivity,
vol. 17, pp. 306–309, June 2007. http://dx.doi.org/10.1109/TASC.2007.898060

15
 [25] X. Hu, T. Zhong, J. E. White, E. A. Dauler, F. Najai, C. H. Herder, F. N. C. Wong, and
K. K. Berggren, “Fiber-Coupled Nanowire Photon Counter at 1550 nm with 24% Sys-
tem Detection Eficiency,” Optics Letters, vol. 34, pp. 3607–3609, December 2009.
http://dx.doi.org/10.1364/OL.34.003607

[26] M. A. Itzler, M. Entwistle, M. Owens, K. Patel, X. Jiang, K. Slomkowski, S. Rangwala,

P. F. Zalud, T. Senko, J. Tower, and J. Ferraro, “Geiger-Mode Avalanche Photodiode Focal
Plane Arrays for Three-Dimensional Imaging LADAR,” Proceedings of SPIE, edited by
M. Strojnik and G. Paez, vol. 7808, p. 78080C, 2010.
http://dx.doi.org/10.1117/12.861600

[27] M. A. Itzler, R. Ben-Michael, C.-F. Hsu, K. Slomkowski, A. Tosi, S. Cova, F. Zappa, and
R. Ispasoiu, “Single Photon Avalanche Diodes (SPADs) for 1.5 µm Photon Counting Ap-
plications,” Journal of Modern Optics, vol. 54, pp. 283–304, 2007.
http://dx.doi.org/10.1080/09500340600792291

[28] R. La Rue, G. Davis, D. Pudvay, K. Costello, and V. Aebi, “Photon Counting 1060-nm
Hybrid Photomultiplier with High Quantum Eficiency,” IEEE Electron Device Letters,
vol. 20, pp. 126–128, March 1999. http://dx.doi.org/10.1109/55.748909

[29] J. W. Gin, D. H. Nguyen, and K. J. Quirk, “High Data Rate Receiver for Optical PPM
Communications,” NASA Science Technology Conference, 2007.
http://esto.nasa.gov/conferences/nstc2007/papers/Gin_Jonathan_D6P2_NSTC-07-0140.
pdf

[30] J. Mendenhall, L. Candell, P. Hopman, G. Zogbi, D. Boroson, D. Caplan, C. Digenis,

D. Hearn, and R. Shoup, “Design of an Optical Photon Counting Array Receiver System
for Deep-Space Communications,” Proceedings of the IEEE, vol. 95, pp. 2059–2069, Oc-
tober 2007. http://dx.doi.org/10.1109/JPROC.2007.905098

[31] K. Birnbaum, W. Farr, J. Gin, B. Moision, K. Quirk, and M. Wright, “Demonstration of

a High-Eficiency Free-Space Optical Communications Link,” Proceedings of SPIE, edited
by H. Hemmati, vol. 7199, p. 71990A, 2009. http://dx.doi.org/10.1117/12.809720

16