Add on-board optical delays for different interferometers

We currently neglect any on-board optical delays of the laser beams, effectively assuming all S/C to be 'point masses'. Such delays might negatively affect laser frequency noise suppression in TDI. To study this, we should add them to the simulation.

To get a first crude estimate of the order of magnitude of these delays, I did some 'pixel-counting' on the OB diagrams, starting from the first beam-splitter (common paths before shouldn't affect laser noise residuals): and converted this from pixels to meters in a mathematica notebook, ob-distances.nb .

Most of the delays are sub-m and probably negligable for our overall noise residuals. However, the backlink fibre might be several meters long and will presumably give the most significant contribution.