The project work was divided into three work packages:
Package 1, develop the low-noise interferometric readout: The group in Birmingham has significant expertise with interferometry, and in particular with devices compact enough for integration in a seismometer. With this advanced start, our first challenge was to consistently perform sensitive measurements at frequencies between 0.01 and 1Hz. We built a number of interferometers in several different sizes to allow us to systematically investigate the way in which external fluctuations, such as sound or vibration or temperature changes, influence the instrument. Our second interferometer design, pictured below, already produced excellent results, reaching the sensitivity required for new seismometers. More importantly, the results were repeatable, and we were able to move forward with confidence.
Package 2, integrate an interferometer with seismometer mechanics: The integration of an interferometer with the bulky, cylindrical mechanics of our chosen conventional seismometer required several design iterations. The first generation was deliberately over-sized and cumbersome, designed to allow us the easiest access to tune the optics. Despite its ungainly form, initial performance measurements were very promising. Without a vibration-isolation system, or multiple devices, we were unable to characterise the performance thoroughly, but at some frequencies the ground motion is small enough that we could see our instrument’s self-noise, which was substantially lower than that of the conventional readout. We used the large prototype to characterise how defects in the optics, particularly in the polarisation of the light, effect the output signal. This kind of test allows us to determine the required quality of the optics, and the maximum speed we can accurately track.
Package 3: construct a ‘field-unit’ capable of deployment at a gravitational-wave detector: Five different design iterations were required, solving a series of practical problems, before we had a solution that was simultaneously small, practical, functional, robust, and simple to align. The final ‘deliverable’ unit was not much larger than the base seismometer, with the interferometer bolted onto the side, and it is ‘powered’ by an optical fibre carrying a small amount of infra-red laser-light. We tested the unit by placing it inside a box to shield it from air-currents and temperature fluctuations (standard practise with seismometers), and as with our first prototype, at frequencies where the ground vibration is low, we demonstrated performance close to the thermal-noise limit of the mechanics.
