1.

INTRODUCTION

Satellite Constellations are launched in space mainly at Low Earth Orbit (LEO). In 2015, three main groups were competing to provide the internet from satellites through optical communications; Google, Facebook, SpaceX-Airbus, with a plan to have a global 5000-6000 satellites in a few constellations. Now they are about a dozen major players planning to populate the space with about 55,000 to 60,000 satellites by 2029 [1].

In parallel, the studies of intermediate liaisons between satellites and Earth such as balloons, airplanes, or drones are now assembled within a group called HAP(S) High Altitude Platforms or High-Altitude Pseudo-Satellites. Satellite primes are testing airplanes high at about 25-35 km in space (e.g. Zephyr from Airbus) as HAP. The market for optical inter-satellite communications has evolved based on the growth in demand for bandwidth capacity.

RF communications is still a bottleneck for moving data at Internet speeds in space. The centimeter-long wavelengths of microwaves in high-frequency RF transmission bands result in widely diffracting beams that spread over hundreds of kilometers on the Earth when transmitted over 39,000 km from a satellite above in geosynchronous orbit (GEO). Increasing data transmission to speeds in excess of 100 Gb/s requires a more concentrated signal at the target and higher transmit powers to deliver the same energy per bit to the receiver. Satellite services based on RF, offer limited bandwidth and have high latency, and they require licensing of the extremely crowded frequency spectrum.

Fiber optic cables provide the capacity and latency. However, the installation of fiber optic infrastructure is often impractical, inaccessible, and/or uneconomical in underserved or remote geographical regions. It is also not an option for transport vehicles such as aircraft and ships where the demand for internet access and associated services is also growing. Therefore, optical inter-satellite links offer the potential for “affordable terrestrial fiber quality service with global coverage” for:

The growing internet market is driving the interest in optical satellite telecom, following the plans of Google, Facebook, SpaceX, Airbus, Telesat, and others, to invest in the development of large constellations using optical telecom. There are tremendous potential commercial applications, in particular, to connect a considerable percentage of the Earth’s population to the Internet.

The first generation of satellite constellations using optical communications is planned for deployment in 2022 with service in 2025. While these constellations will offer Gbit capacity, there is already a desire to see terabit/s capacity, in order to seamlessly integrate with the existing terrestrial infrastructure.

It is expected that the 10 W optical amplifier based on Erbium and Erbium-Ytterbium Doped Fiber (EDFA) will permit more than 100 Gbps data transfer. The EDFAs are key devices in the optical communications in space with the following examples of applications:

2.

OPTICAL INTERSATELLITE LINK BUDGET

To validate the functionality of a transmitter for a fast data transfer, one main item to find is the number of photons getting in the unit time, e.g. 10^-11 second for the 100 Gbps, is high enough to have an acceptable small Bit Error Rate (BER). This leads to the estimation of the link budget between the transmitter and the receiver. A simplified schematic of the inter-satellite link is presented in Figure 1.

Figure 1.

Simplified schematic of an Optical Intersatellite Link (OISL)

It starts with a transceiver sending a small signal in the high gain amplifier called usually a “booster”. For the Optical Intersatellite link and Optical to ground link the booster output power is between 1 and 10W (1550 nm). The signal is reduced by a large factor due to the dispersion in long distances in free space called the “Isotropic Space Loss” which is about -260 dB for LEO links (1300 km) and -290 dB for a GEO Link (39770 km) [2] Although these values are very large the signal is getting at the receiver point is a very small signal (down to -48 dBm). A Low Noise Amplifier (called pre-amplifier) regenerates the signal and sends it in the input of the transceiver to repeat the transmission.

Various parameters need to be known in o evaluate the link budget we consider various involved parameters:

To note, the term “GigaBit Per Second” or Gbps is used, the important parameter for the data transfer is the “Giga Symbol Per Second”. The Bit/Symbol depends on the used modulation technique. Table 1 gives the number of Bit/Symbol for some frequently used modulation techniques. For example, most of the current commercial fast transceivers claiming the capability of 100 Gbps rate, are in fact at 28 Gbps, with a DP-QPSK modulation transferring at 112 Giga-Symbol per second.

Table 1:

Number of Bit/Symbol for some frequently used modulation techniques

The link budget relates the power at the Receiver is related to the Transmitter power, its evaluation is needed to verify the required power at the transmitter level, to determine the number of photons per unit time (bit), arriving at the entrance of the receiver (pre-amplifier). The following equation is commonly used by the satellite primes to evaluate the required transmitter power, the size of the transmitter and receiver telescopes, and the requirement of the preamplifier:

P_R is the received optical power, detected at distance L

P_T is the transmitted average optical power with wavelength λ

G_T = (πD_T/λ)² and G_R = (πD_R/λ)² are the transmitter and receiver gains, respectively

(D_T and D_R are the diameters of the transmitter and receiver telescope, respectively)

η_T, η_R, and η_C are the transmitter, receiver, and fiber coupling efficiencies, respectively,

η_ATM is the atmospheric attenuation in the case of Satellite Ground or Ground-Satellite;

L_FS = (λ/4πL)² is the free-space loss, L is the distance Transmitter-Receiver

L_p = exp(G_Tθ²_BW)corresponds to the pointing errors, θ_BW is the pointing jitter

L_SI is the scintillation-induced errors when the light is transmitted through the Earth Atmosphere

For a fixed distance, to increase the signal at the receiver (P_R) we can increase the transmitter power, (P_T), the diameter of the transmitter telescope (D_T), and that of the receiver (D_R). In most cases, they are the other parameters related to the efficiencies and pointing errors

In 2019 and 2020 ESA accorded two contracts in parallel, “High Throughput Optical Network (Hydron), Artes ScyLight” within Hydron subprogram (ESA AO/1-9814/19/UK/AB); to study the feasibility of 100 Gbps with 10 wavelengths each at 10 Gbps. The final objective is to compare the advantages and disadvantages of the two methods:

3.

OPTION-1-ONE TRANSMITTER (10W, 1λx100Gbps)

3.2

Development of the 10 W Amplifiers in the 1550 nm range at MPB

For the method with one Transmitter at 10W (1λx100Gbps), a simple diagram of the demonstrator with the 100 Gbps transceiver is:

The demonstrator with the 100 Gbps transceiver is being built in collaboration with the supplier of such a commercial product. This product is still in the prototype stage, MPB will perform its space qualification during the collaboration.

MPB developed a family of EDFA amplifiers (1550 nm) at 1-3W output power, presented at ICSO-2018 [9]. Recently a preliminary space-qualified 10 W Polarization Maintaining Amplifier based on Erbium-Ytterbium Doped Fiber. The amplifier characterization (Table 3, Figure 2) is presented in a paper at ICSO 2020 [10]. At the 10 W level, the thermal dissipation of the heat generated by the electronics and the laser diode pumps at 940 nm becomes the main challenge for a space-qualified amplifier, with limited surfaces, volume, and mass. To have the amplifier functional between -20°C and 50°C MPB had to use an Innovative Proprietary Method of the spool design for the active fibers. Work is progressing to have a larger range, at least between -35°C and + 65°C.

Figure 2.

Wall-plug Power conversion efficiency of the (1552nm) output power. Blue points are electrical input and red points are wall-plug efficiency

The optical loss signal in space of the order of 60-72 dB, estimated from the Link budget (previous section) can be represented by Variable Optical Attenuators (VOA).

Table 3:

Experimental Efficiency 10 W-PM EDFA

Loss Power (W)	Loss Power (W)	Last (2020)
Nominal output Power(W)	-	10.4
Er:Yb-doped fiber efficiency	25	32%
Combiner efficiency signal	0.07	93%
Combiner pump transmission	1.5	95%
Isolator efficiency	1	89%
Laser Diode Pump power	35.5	51.5%
Driver and control board efficiency	15	83%
Total electrical power needed(W)	-	87
Total Loss (W)	77	89.6%
Wall-plug efficiency	-	11.4 %

4.

OPTION-2- 10 TRANSMITTERS (1W,10λx10Gbps)

4.2

Present Development of 10λx10Gbps at MPB

For the method with 10 transmitters: MPB completed a detailed optical design of Reconfigurable Optical Add-Drop Multiplexer (ROADM) for Space Optical Cross Connection Links (OXC), based on DWDM, having 10-ITU channels for a total 100Gbps.

The ROADM is connecting in bi-direction, with Add/Drop. We consider 4 satellites each in one direction noted East (E), South (S), West (W), and North (N), with a fifth option to the Add/Drop to Ground (Figure 3). The OXC design permits switching signals from one direction to any other for two channels having the same ITU wavelengths or two different wavelengths. The Add/Drop signals – can add one or two ground data signals or drop one or two ROADM data signals into and from any direction and any DWDM (wavelength) channel among the ten λ-channels, respectively (Figure 4).

Figure 3.

Space Reconfigurable Optical Add-Drop Multiplexer (ROADM) Design

Figure 4.

Space ROADM Design-Connection of the channels (wavelengths)

The optical signal power flatness between the λ-channels is controlled by power monitors, including a 2-5% tap photodetector (PIN-photodiode) and VOAs. For 10λx10Gbps, various suppliers were contacted.

Although all the optical components are available as commercial products, the price of optical components (Optical switches, MUX/DEMUX, couplers, VOAs, channel-power-monitor, DFB-transmitters, and photodetectors) is of the order of 200 K€, without the electronics that monitor these components.

MPB reduced the number of components (e.g. wavelengths) to build a simplified prototype significantly representative of ROADM-OXC at an acceptable cost. This exercise permits MPB to keep a minimum of know-how to stay within the best international companies providing the ROADM-OXCs.

Figure 5 illustrates the demonstrator to build, it was reduced to 2x2 Cross Connection in all direction between two satellites e.g. satellite East (E) and Satellite West (W).

Figure 5.

Space ROADM Design Summary of Electronics and Optical components

For the method with 10 Transmitters at 1W (10λx10Gbps), MPB will develop a 2x2 demonstrator to verify the technology and compare the experimental results to available satellite data at close rate e.g. those from EDRS, and works developed at the NASA or ESA.

A simple diagram of the demonstrator with 2x2 Cross Connection, and 10 Gbps Transceivers is:

MPB is in the process of building the 1W (2λx10Gbps), demonstrator using MPB amplifiers of 1W at the transmitters and the Low Noise Amplifier at the receivers

5.

COMPARISONS OF TRANSMITTERS: 1X(10W,1λX100GBPS) VS 10X(1W,10λX10GBPS)

In the next step, we will compare the performance of the two methods (1λx 10Wx100Gbps) and (10λ x 1Wx 10Gbps). A preliminary comparison is given in Table 5. The 10W version seems very well needed unless the telescope diameters at the receiver and transmitters are increased to > 20 cm each.

Table 5:

Comparison of the two methods 1λx 10Wx100Gbps with 10λ x 1Wx 10Gbps