Small-satellite networks need fast links between spacecraft without exceeding tight size, mass and power (SMaP) budgets. LED SPAD optical communication offers a potential alternative to conventional radio-frequency links and more complex laser systems: a gallium nitride micro-LED sends digitally modulated light, while a single-photon avalanche diode (SPAD) array detects it with high sensitivity. Experimental results demonstrate data rates up to 100 Mb/s, while simulations examine how optics and beam divergence affect range and pointing tolerance.
Why consider LED SPAD optical communication?
Optical inter-satellite links can support high data rates, but established laser approaches often rely on precise pointing and comparatively complex hardware. Those requirements can be difficult to accommodate on a CubeSat, where available space, electrical power and system resources are constrained.
LEDs offer a different design trade-off. Their emitted light is divergent rather than tightly collimated, which can provide coverage across an angular region and ease alignment demands. The approach may also reduce transmitter complexity relative to laser systems. The trade-off is that spreading light reduces the power reaching the receiver as distance increases, limiting achievable range or data rate unless the optical system and link budget are designed accordingly.
How the micro-LED and SPAD link works
The experimental transmitter used one 99 × 99 µm GaN micro-LED pixel, selected from a 16 × 16 array bonded to CMOS control electronics. The device emits light centered at 450 nm. Its integrated format combines optical emitters with electronics that independently control each pixel.
At the receiving end, a 64 × 64 silicon SPAD array detects incoming photons. SPADs operate in a way that makes them highly sensitive to low light levels. In this setup, digital signals from the detector pixels are combined using XOR trees and ripple counters, so the array functions as a digital silicon photomultiplier.
In intensity modulation, information is conveyed through changes in the light’s intensity rather than its phase. This provides a comparatively direct signaling approach for the LED transmitter and SPAD receiver. The measured sensitivity was -60.5 dBm at 50 Mb/s and -55.2 dBm at 100 Mb/s. These are experimental link results, not evidence of a completed long-range satellite-to-satellite demonstration.
What the range and coverage simulations show
Ray-tracing simulations assessed a point-to-point link using a 1 mW Lambertian LED source and a 10 cm diameter receiver collection area. Three transmitter configurations were compared: no lens, a C240 lens, and an ACL25416 lens. Estimated data rates were derived from the simulated received intensity and an approximation based on measured receiver sensitivity.
The modeling illustrates a basic optical-link trade-off. As a divergent beam spreads with distance, less light reaches the receiver, reducing the data rate that can be supported. The study describes potentially useful rates over ranges extending to hundreds of kilometres in modeled scenarios. These figures are projections from the assumed optical setup, not demonstrated operating distances; actual performance depends on system parameters and deployment conditions.
Beam divergence also affects angular coverage. Without a lens, the broad Lambertian emission provides wider coverage: the simulations indicate that a 100 Mb/s link could be maintained at almost 60 degrees over 10 m, while a 10 kb/s link could retain similar angular coverage at approximately 1 km. Adding collimating optics narrows the beam. The modeled coverage was roughly 1 degree with the C240 lens and 0.4 degrees with the ACL25416, illustrating the exchange between broader alignment tolerance and concentrated light.
Integration and practical constraints
The experimental chips were mounted on evaluation boards controlled by FPGAs. In that configuration, the transmitter and receiver packages measured 14 × 18.5 cm and 12.5 × 20.5 cm, respectively—larger than a CubeSat’s 10 × 10 cm form factor. The evaluation boards contain components not needed in a final design, so a smaller transceiver is a development possibility, not a demonstrated implementation.
The full operating setup consumed 5.48 W, with most power attributed to FPGA control. The system had not been optimized for power. The authors estimated that if a spacecraft could reuse its existing FPGA resources, the micro-LED and SPAD arrays themselves might add less than 1 W of consumption. This estimate depends on the final design and should not be treated as a measured power figure for a flight-ready unit.
Several environmental and optical factors also remain important:
- Background light: Additional light at the SPAD receiver raises the signal power needed for reliable communication. Spectral and spatial filtering may help, while direct solar irradiation could saturate the receiver.
- Collection optics: The simulations assumed an ideal receiver collection area. Real optics introduce practical limitations and losses.
- Space environment: The performance of these specific devices in orbit is not established. The study identifies radiation effects as an open consideration, particularly for the less mature GaN devices.
What engineers should take away
LED and SPAD hardware presents a promising research direction for high-sensitivity optical links under small-satellite SMaP constraints. The measured 100 Mb/s result establishes experimental data-rate capability, while simulations suggest that link range and angular coverage can be adjusted through optical design. Broad LED emission may reduce pointing demands, but it comes with a range and received-power trade-off.
For system designers, the next steps are practical: refine the hardware form factor and power use, test longer-range communication through air, and evaluate the link with realistic optical and background-light conditions. The results support further development, but flight integration and in-orbit performance still require validation.