barton-satchwill / fsoc

Free-Space Optical Communication

FSOC

Free-Space Optical Communication

During the fall of 2013, NASA's LADEE robotic spacecraft demonstrated the use of lasers for high-speed, 2-way communication between the Earth and the Moon.

Light has been used to transmit information for centuries. Today, laser light transmits data through glass fibres that form the vast fibre-optic networks that underpin our modern telecommunication world.

In a sense, all that NASA did was to deploy a fibre optic network between the Earth and the Moon, but without the fibre. They simply sent the laser light through free space.

If NASA can do it, why can't I?

The Hardware

I used two Arduino Uno's for this project, with very simple circuits to drive a 36¢ laser as the transmitter and the phototransistor as a receiver. For most of the development and testing, a bright LED replaced the laser, and allowed the two development boards to be conveniently placed within a few inches of each other.

The Software

The software implements --in a very rudimentary and incomplete way-- layers 1, 2 and 4 of the OSI network communications protocol stack. Let's face it: there's no need for layer 3, there are only two nodes in our setup. The code is small and fast, not much more that 4 kilobytes. Both the receiving and transmitting rates are governed by the 16-bit Arduino timer1, with a single bit being transmitted each timer cycle. As it didn't make any sense to drive the timer faster than the rest of the system could send and receive a signal, I instrumented the code to measure how long it took to send or receive a single bit.

limiting factor	Time per Bit	Bits per Second
Arduino Uno 16 mHz processor	0.0625 μsec	16,000,000
transmission code	8 μsec	125,000
receiver code	13.5 μsec	74,000
laser response time	not a clue	not a clue
phototransistor response time	~ 5-50 μsec	200,000 - 20,000

The phototransistor and laser introduce a lot of uncertainty. The table reports typical response time I gleaned from a bit of reading. I tried to nail down exactly what the response time of my phototransistor was, but was never able to achieve convincing results. Phototransistors are sensitive to saturation, and a pulse of light that is too bright can actually lengthen the response time. Test results varied enormously depending on the position and intensity of the laser.

Stability

Early iterations of the code were not stable, and suffered from bit slip as the clocks on the transmitter and receiver drifted apart. This was addressed by periodically re-synchronising the receiver, and by taking multiple samples from the phototransistor each cycle. Thus, the actual time to read a single bit is better expressed as 13.5+(4.5n) μsec for n samples.

Timer frequency, number of samples per bit, and re-synchronisation period became tuning parameters for the system.

Results

Ultimately, I was able to achieve a remarkably stable system transmitting data at a whopping 15 kilobits per second, beating the 14,400 modems of the mid 90's .

Next Steps

The next obvious evolution for this project is to use orbiting mirrors to send a signal from the Earth to my secret base on the far side of the Moon.