So perhaps I'd better explain what goes on and what this experiment is in very general terms (and in layman's terms because I'm a Software Engineer not an astronomer). This project is an experiment to test a new adaptive optics system. Now adaptive optics are a way of compensating for distortions in the earth's atmosphere so you can get clearer images from ground based telescopes. It basically works by observing a light source in the part of the sky where you want to observe your target object and by watching how that distorts you can work out how the atmosphere is changing. Once you know this you can use a deformable mirror, basically a mirror that changes shape, to compensate for these distortions and you get a clearer image. Obviously you have to make these measurements may times a second to make this work and at regular intervals, otherwise it doesn't work. Also the time taken to capture your image, calculate the shape of the mirror and the time for the mirror to change must be very short otherwise the atmosphere will have changed to something else and you have to start again. The first experiments in this used stars as the light source to work out how to change the mirror but these days we use lasers and make our own light sources by reflecting laser light off of layers of dust, etc in the atmosphere.
OK, so that's the science bit. In terms of Software Engineering what we have here is a hard real-time system. The data has to flow through the system and meet stringent timing criteria to work. The numbers we're talking about are four cameras, each taking an exposure once every 4ms (250 times a second) which then need to be measured and a shape for the mirror worked out and sent through before the next image has been taken. To do this there are a number of computers, each linked by a fibre-optic bus which allows the data to flow from one computer to another with little delay.
Now the part I'm responsible for (apart from fixing anything that seems to go wrong if nobody else is about) is telemetry. At various points in the system I take the data and then send it via a normal ethernet network to a number of sub-systems which can then use the data to make secondary calculations with slightly less stringent timing constraints. So to recap, I have the job of taking the data flowing though this high capacity fibre-optic network and transferring it to a normal style network (like the kind you'd have around your house say). This is the equivalent of sending a suitcase on a high speed train and at the same time sending an identical suitcase via horse and cart to the same destination and expecting them to arrive at the same time. Some days I have no idea how but I manage to make it work!
So in less that two weeks time we're packing all our equipment up, putting it into a container truck and sending it off to La Palma in the Canary Islands. So this is the last of the time we have left to test anything before we arrive on the Island at the end of August. To say things are hectic is an understatement.
