THE UNITED KINGDOM INFRARED TELESCOPE
ANNUAL REPORT
2001-2002

4. New Instrumentation

4.1. MICHELLE (10-20µm imager and spectrometer)

MICHELLE passed acceptance testing at the UK ATC in May 2001, although some concerns remained over flexure performance. A strengthening rib was added and improvements to the grating mountings were identified, which resulted in repeatable 1 pixel spectral flexure from slit to array. A software look-up procedure using the translation stage to compensate for this motion was implemented and proven to work well (observations showed that minimal peaking up is needed when placing a point source in the slit). Final waveforms were developed which gave the necessary array performance and efficiencies on telescope, for all observing modes. MICHELLE would have benefitted from additional time in the laboratory post-acceptance, but the project cash limit made this impossible. A request for access to contingency funds was approved in June 2001 and the closing cost of the project was £6446k, within the approved total.

After a successful engineering period in Hilo and transfer to the summit, the cryostat was fitted to UKIRT on 21st August 2001. First light followed on 22nd August, producing excellent images. Whilst the transfer and installation of the MICHELLE hardware and software to UKIRT went well, some issues were raised by the compressed timescale. The data reduction recipes had not been adequately tested, and quite a lot of effort was needed to get them working on site. It was decided to adopt ORAC3 in conjunction with MICHELLE's arrival, which increased the number of software bugs in the Observation Manager and Observing Tool software. An engineering interface was needed for much of the on-telescope commissioning. The process of integrating chop control software was completed close to the start of PATT observing. After a seven week commissioning period, the first PATT MICHELLE run took place on 15th October 2001.

Considering the complexity of this instrument, the MICHELLE fault rate was remarkably low during 01B, %lt;0.3%. However this disguises the fact that MICHELLE had to be warmed up in December and was unavailable for three weeks, during which several MICHELLE runs were scheduled. Backup programmes using other facility instruments were done instead. The warm-up was required due to water damage of the hygroscopic window, probably incurred during a major storm in November, which resulted in large regions of unusably high emissivity. Since the window is part of the cryostat, to replace it required a full warm-up. Unfortunately the same problem occurred in 02A, again after a major storm. To avoid a reoccurence a sacrificial thin window was installed in front of the main window into the cryostat, which can be swapped quickly without a warm-up, and dry-gas purging of the entire calibration unit was implemented. The former required design work and procurement of suitable samples of KBr; the latter required the acquisition of a large compressor and various installation work on the telescope. Both were ready by the time of the second down time in February 2002. Problems were also experienced with the Joule-Thomson (J-T) cooler and with the optics closed cycle cooler; both cold heads were serviced. Temperature fluctuations experienced with the J-T cooler were believed to be due to blockages caused by contaminants in the supply lines, and regular purgings were implemented. In 02A the fault rate increased to 4% as new modes were tested and the instrument was heavily used. Lost time was dominated by a calibration unit fault, fixed by re-engineering a drive shaft nut. Also, a grating microswitch failed, the J-T cooling system failed, and there were occasional problems with the EDICT array controller. Fixes or workarounds were developed. Two gratings (LowQ and MedN2) were incorrectly fitted, affecting two PATT programmes in 01B and 02A.

The sensitivity of MICHELLE in all modes is up to a factor of two lower than predicted. Causes were investigated, though hampered by poor weather which severely restricted engineering and prevented an investigation of chopping rates. Despite these problems, and the poor weather, MICHELLE was successfully used for numerous PATT runs and papers are in preparation or already published.

4.2. UIST (1-5µm imager and spectrometer with 1024x1024 array)

During semester 01A the UIST team at the UK ATC were heavily engaged in making the instrument's performance acceptable for delivery to the telescope, in particular in the area of electronics and array control. The read noise was measured to be 150 electrons, which was within a factor of three of the expected noise but a factor of ten away from the 15 electrons that can be achieved with these arrays. Booting the full EDICT system in an operationally viable manner was also an outstanding issue. Effort for development of the UIST high level software was made available as team members overlapping with MICHELLE returned. Integral-field spectroscopy using the IFU is the most novel mode and the UIST graduate student at the UK ATC worked on the ORAC-DR recipes required for this mode. Software in general was a critical item, in particular the interface to the telescope system.

The noise performance could not be brought within acceptable levels in time for the initially proposed February 2002 delivery. Acceptance tests were delayed to April 2002, and excluded array performance. The instrument was accepted on provision that certain essential items were completed before delivery. JAC and UK ATC staff held detailed discussions of the predicted scientific performance of the instrument with current noise and agreed a programme of continued engineering work. Significant improvement in the array performance was made post-acceptance, using digital averaging techniques combined with multiple non-destructive read-out. By early 2002 the high-level software for UIST was largely complete, and the instrument fit within the ORAC environment including the Observing Tool and Observation Manager. Reduction software for the imaging and spectroscopy modes were adapted from the other instruments by JAC staff. ORACDR reduction recipes for the IFU were effectively completed at the UK ATC. The project closed at £2982k, within the total allocation of £2984k.

Acceptable noise performance was achieved prior to delivery in August 2002. Six engineers and scientists from the UK ATC came to the JAC for commissioning, which began on August 19th. UIST was installed on the telescope in September 2002 and first light was achieved on September 24th 2002. Commissioning took place over 20 nights between October and December. UIST integrated well with the telescope and software systems. On-telescope flexure measurements were acceptable and a UIST-specific pointing model was installed. Cryogenic performance and image quality is good. All but two of the grisms and one of the blocking filters performed well; the problem grisms and filter will be replaced. Thermal readout, the IFU and polarimetry were all successfully implemented. The most problematical issue with the instrument is the throughput, which is lower than expected (25% compared to 40%). This is under investigation.