One of the functions of the PFS is to distribute the light for science observations using flat mirrors that can be moved into position to send the light from the telescope to the science instruments. When they are deployed, these mirrors are the sixth mirror (M6) in the telescope’s optical path. The PFS can use one of two mirrors: one for the Nasmyth instruments (at lateral focus, known as M6N) and another for the Coudé instruments (known as M6C). These M6 mirrors can also be fully retracted from the telescope’s optical beam so that the instrument at the centre of the Nasmyth platform (the straight-through instrument) receives the telescope’s light.

The M6N mirror has an optically flat elliptical shape and is extremely flat: to within a few tens of nanometres. Its optical surface has dimensions of 1440 mm x 1042 mm and it has a highly reflective metal coating. The M6C mirror is slightly smaller with an optical aperture of 691 mm x 500 mm along the major and minor axes respectively.

During science observations, the ELT measures images of stars close to the science target in addition to the science object itself. These additional stars are known as guide stars. The guide-star images are analysed by the telescope control system to obtain information about the alignment of the telescope, which in turn is used to actively position and control the mirrors and telescope main structure to keep the telescope well aligned.

There are several effects that might cause the ELT to slowly drift out of optimum alignment during science observations. Slight changes in temperature inside the telescope chamber cause the metal structure and optics to thermally expand, which changes its dimensions and causes the mirrors to move out of alignment. The direction that gravity acts on the mirrors and their supporting structures also varies as the telescope is tilted up and down to follow the motion of stars in the sky. This causes deflections in the telescope structure and in turn errors in its optical alignment. These effects, which typically range in magnitude from hundreds of micrometres to a few millimetres, seem small on the scale of the ELT telescope. However, the optical alignment of the telescope is very sensitive to misalignments of this size, so the mirror positions and shapes are actively controlled to minimise the optical errors that build up during the science exposure: a process called active optics.

The PFS Natural Guide System includes an adapter subsystem, which has the task of observing and measuring up to three guide stars simultaneously with the science observation. It does this with the help of its three sensor arms mounted on circular bearings. The sensor arms are motorised so that they can move on a circular path around the bearing and radially towards the centre of the adapter. In this way, the sensor arms can find guide stars in the telescope optical beam and follow them as the sky appears to rotate during a science observation. 

To observe a guide star, each sensor arm includes a flat pick-off mirror at its tip that reflects a small part of the telescope’s light into the sensor arm. Inside the sensor arm, the optical beam is split into two channels using a wavelength splitter known as a dichroic mirror.  

The longer wavelength channel, covering 800–950 nm, is sent to an imaging camera that makes an image of a guide star. This tip-tilt camera measures the position of the star and generates a signal that can be used to control the pointing direction of the telescope.  

The short wavelength channel, covering the wavelength range 500–800 nm, is sent to a wavefront sensing camera. This camera determines how much the starlight is distorted by both changes in mirror alignment and turbulence in the atmosphere. It does this by measuring aberrations in the telescope’s optical beam, for example the focus error and astigmatism. The wavefront sensing camera also has the capability to measure higher order aberrations, up to around 400 spatial modes at a frame rate of up to 500Hz or more. Aberrations can be corrected using a deformable mirror, like the ELT’s M4, a process known as adaptive optics. 

Each sensor arm is approximately 1200mm long and has a mass of around 300kg. The required positioning accuracy of each sensor arm is a few hundred micrometres under all operating environmental conditions and for motion speeds up to approximately 5–10 mm/second.

The ELT primary mirror, with its 798 individual segments, acts as a single continuous optical surface when the telescope is observing. When the shape of this surface is ideal we say that the mirror is well-phased. The wavelength of light captured by the ELT ranges from a few hundred nanometres to 24 micrometres. To produce near perfect (diffraction-limited) images, the positions of the segments relative to each other must be controlled to tens of nanometres. Motors known as position actuators control the position of the segments to maintain their relative positions using signals from edge sensors attached to each segment. However, over time these sensors will drift slightly, which can be caused by natural variations in temperature, humidity and the inherent drift of the measuring electronics. This drift of the edge sensors will cause M1 to lose its ideal shape over a projected time scale of around two weeks. 

The PDS subsystem of the PFS will periodically measure the shape of M1 using guide star measurements and allow the position errors of the segments to be calculated and corrected by updating the segment control positions. This phasing measurement can be performed by using a wavefront sensor to analyse the characteristic diffraction patterns produced by unwanted steps between segments that are not well-phased. The PDS is installed in the PFS structure on the lower level and receives the telescope light reflected downwards by the M6C mirror.  

The PDS also contains several other optical sensing modules that will help test the telescope when it is first constructed and later enable troubleshooting and maintenance of the ELT over its operational lifetime.