RINGO3  is a fast-readout optical imaging polarimeter. It uses a polaroid that rotates once per second to measure the polarisation of light entering the instrument. A pair of dichroic mirrors split the light into three beams for simultaneous polarised imaging in three wavebands using three separate Andor cameras.

Exposures are synchronised with the phase of the polaroid's rotation and each camera receives eight exposures per rotation. All images for each octal phase are stacked to obtain the final signal at each phase in the polaroid's rotation. Observing for longer and thus stacking more images in each octal phase increases the signal-to-noise ratio obtained and therefore improves the polarisation accuracy of the observation.

RINGO3 was commissioned in early 2013. However, observations later in the year demonstrated a problem with polarisation induced by the two dichroic mirrors. This instrumental polarisation unfortunately changes with the mount angle (the angle of the cassegrain stage) used for the observations. The problem is described in more detail here. All users of RINGO3 in 2013 are encouraged to read this document and/or contact us (the LT support astronomer) to discuss their observations PRIOR to publishing.

To combat this problem, in December 2013 we inserted a depolarising Lyot prism into the beam, after the rotating polaroid but before the beam reaches the first dichroic mirror. Note that downstream of the rotating polaroid all one needs to measure is the intensity of the light, not its polarisation state. Prospective users of RINGO3 are encouraged to read this document for further details. We are now confident in the calibration of RINGO3, and there is no need to do observations at multiple mount angles as we were recommending in 2014.

From installation until June 2015 the rotor turned at roughly 1Hz. From 29th June 2015 onwards the rotor is running at ∼0.4Hz. Procedures for data handling, processing and analysis are unaffected by this change. Total dwell time on a target is unchanged so no change is required to phase2 configurations, but it means the integration time per frame is now about 0.32sec instead of 0.125sec. Targets that were close to saturating at 1Hz may now be saturated.