CERN's surveyors are pushing back the frontiers of precision

Aiming at a target on the other side of the Alps, 730 kilometres from CERN, or controlling the positions of thousands of devices to a precision of one tenth of a millimetre, these are just some of the painstaking tasks undertaken by the surveyors in the Positioning Metrology and Surveying Group. These masters of measurement are pushing precision to its very limit.

Go down into the LEP tunnel, walk about half a mile and then try to imagine how you could possibly take precise aim at something hundreds of kilometres away without any reference to the surface. Absurd, you might think? Not entirely, for that, in a nutshell, is the geodetic challenge of the Gran Sasso project. Indeed it is just one of the challenges faced by the surveyors in CERN's Positioning Metrology and Surveying Group, whose task it will be to aim a neutrino beam at a detector located in an underground cavern 732 kilometres away at INFN's Gran Sasso laboratory in Italy.
The tools for solving such problems are provided by geodetics, the branch of science devoted to studying the shape of the earth and its field of gravity using an array of earth and space-based positioning techniques. The group has installed a geodetic reference network enabling it to determine the relative positions of the CERN accelerators with great precision. But we also need to know very precisely where we are with respect to the rest of the world and for that it was necessary to link up with national geodetic networks. The CERN network therefore had to be integrated into the Swiss network and connected to the closest points on the world geodetic network. Such geodetic measurements are performed by special use of the GPS navigation system (see inset), which allows measurements to precisions greater than 0.5 mm over ten kilometres.
This gives the theoretical bearing (azimuth) and inclination of the beam line. But the exercise is not quite as simple in practice since there is only one connection point to the surface network, via the single access shaft. To keep their own bearings when down in the tunnels our intrepid surveyors are equipped not with simple compasses but with precision gyro-theodolites (2 arc-seconds). But even once the azimuth is properly oriented, the greatest difficulty still lies in setting the correct inclination, which itself depends on the local vertical... and that's where the fun really starts. While satellites use a mathematical surface known as an ellipsoid as their system of reference, altitudes are calculated above a reference equipotential surface known as the geoid - and this is complicated by the fact that the zero surface is not regular - its shape and position vary according to structural or topographical gravity anomalies. Despite the obstacles, the SU group has undertaken to obtain a precision of plus or minus 30 metres at a distance of 730 kilometres.

One of the SU group's major tasks is the positioning of the accelerators. The magnets must be positioned to a precision of one tenth of a millimetre. Above, the positioning of the magnets for the LHC string test.

Not only are the CERN surveyors skilled marksmen, capable of hitting the most distant target, but they are also masters of minute precision. One of their most important tasks is metrology for the accelerators and experiments. For the accelerators, the aim is to position all the magnets perfectly so that the beams are not affected by magnetic disturbances due to alignment faults. The positioning accuracy has continued to increase from one CERN project to the next: from 0.6 millimetres for the PS and 0.15 mm for the ISR to 0.1 mm between two quadrupoles over the 7 kilometres of the SPS circumference and the 27 km of LEP!'We're pushing existing technologies to their limits, largely thanks to the special instruments developed by our group,' Michel Mayoud , Head of the Positioning Metrology and Surveying group, points out. 'But it wasn't long before we were being asked to do better,' he adds, 'and we've now been working at precision levels of hundredths and even thousandths of a millimetre for quite some time.'
Metrology for accelerators is not confined to the installation of the machine. 'A position is only precise at a given moment,' Michel Mayoud points out. For example, over the course of ten years LEP moved by up to 18 millimetres. Such movement can be caused by all sorts of things, such as the thermal expansion of the structures, the effects of water, etc., so geometrical monitoring is vital. 50 km of beam line, which represents 10,000 components fitted with fiducial targets, thus have to be monitored. Naturally, once LEP has been dismantled only half of these will remain, but the group is already preparing for the positioning of the LHC magnets, which means setting up 3000 new pieces of equipment.
The teams of surveyors also perform metrology for the assembly, installation and maintenance of the experiments, determining, for instance, the relative positions of all the sub-systems within a detector. While this does require certain conventional measurements they mainly use the technique of softcopy photogrammetry, with carefully calibrated 6 million pixel CCD cameras. This involves taking a large number of 'shots' from all angles and then mathematically reconstituting the targeted points of the object by spatial intersection of all the perspective line bunches. The data are then processed by powerful number-crunchers and the results obtained are astonishingly accurate, even on large objects such as the 15-m diameter CMS 'ferris wheel'.
But the surveyors have yet more strings to their bows. CERN's future projects will present new challenges. The Gran Sasso project is a prime example. The future accelerator project CLIC (Compact Linear Collider) will also give them food for thought. For such machines that are highly sensitive to alignment errors and with extremely narrow vacuum chambers, precision at the micron level over several kilometres will be required. After various developments in partnership with industry this challenge has already been met for a length of around 10 metres - the CTF2 accelerator components have been maintained within a 3-micron window!