RxH3 - The JCMT Surface

Ever since the JCMT started operation, the main method of measuring the shape of the surface has been "millimetre-wave holography". This involves making very detailed measurements of the beam of the telescope and exploiting the Fourier transform relationship between the beam pattern and the fields in the aperture to obtain a map of the surface errors. A new system for making this measurement is now in operation. These pictures shows the new receiver during its construction and as it is now, installed in the cabin (just to the left of the entrance doorway).

The new receiver is called RxH3. Its predecessor, RxH2, was retired on 20 February 2002 after almost 15 years of service. The old system operated at 94 GHz and consisted of a source at UKIRT and a single-channel receiver, which simply measured the amplitude of the signal. This meant that maps of the beam had to be made at two different focus settings and then a complicated analysis performed, involving least squares fitting, to recover the phase in the aperture. (Surface errors show up as deviations in the aperture phase.) This system had several disadvantages: it took at least two hours to obtain the data; the phase-retrieval technique did not work well if the errors were very large, so several iterations of measurement and adjustment were needed if panels were a long way out of position; and the accuracy was limited because of the presence of spurious signals due to reflections, in particular from the membrane.

RxH3 is designed to overcome all these limitations. It has the following features:
- It measures both phase and amplitude. This is done by bringing a reference beam through a hole in the primary surface and into the back of the receiver, where it is correlated with the main beam arriving via the primary, secondary and tertiary.
- It operates at either 80.35 and 160.7 GHz. Both these frequencies are emitted simultaneously (on orthogonal polarizations) by the new source unit, which is located inside the UKIRT dome. There are two sets of mixers in the receiver to collect these signals.
- The frequency can be stepped rapidly over a few tens of MHz. This makes it possible to remove the effects of spurious signals.
- The backend obtains very high dynamic range by having both high- and low-gain ranges.
- The real-time VME-based data-acquisition system can take samples at a rate of up to 1 kHz and these are time-stamped and then associated with accurate readings of the telescope position. (The maps are taken "on-the-fly" with the telescope performing a raster scan at speeds of up to 400 arcseconds per second.)

The new instrument was designed and built as a collaborative effort involving MRAO, JAC and RAL. Most of the design and development was carried out at MRAO, who also provided the analysis software. Much of the hardware was built at RAL, while the JAC designed and built the data-acquisition system and integrated the instrument into the JCMT environment. A great deal of time and effort has been taken up in commissioning and de-bugging the system, but this is now essentially complete. There is still a good deal to be done in the way of understanding all the properties of the new instrument, finding the best way of using it and making automatic such steps as the generation of a set of moves which can be sent to the surface adjusters. RxH3 itself and its software is, nevertheless, fully operational and producing good data, as can be seen from the next picture.

This shows a high-resolution data set obtained at 160 GHz. The two upper images are the measured phase and amplitude of the beam pattern (after re-gridding and the application of various calibration steps) and the lower ones are the phase and amplitude in the aperture. These were obtained by taking a Fourier transform and applying corrections for near-field effects and the illumination pattern of the receiver.

Prominent features of these images are: the shadows of the legs that support the secondary mirror; and the rings towards the outside of the aperture, which are due to diffraction at the edge of the secondary mirror. These are, however, of mostly academic interest and are normally removed in further processing to reveal the errors in the surface. The resolution of these images is about 80 mm so we can easily see individual panels which are out of position (e.g. at the bottom of the dish) and in fact it is also clear that there are significant errors within some of the panels - essentially "curled edges".

With the new holography system it should be possible to make measurements with an accuracy of about 5 microns and to do so rapidly enough to understand the thermal deformations of the telescope in the course of a night. At the time of writing the surface error is higher than we would like and this is mostly due to adjusters that are not responding when commanded to move. This is presumably due to problems with either the adjusters themselves or the connections, limit switches, etc. It is hoped that it will be possible to repair these over the coming months. When the adjuster system is working well enough to allow frequent updates to the surface, it should be possible to remove most of the thermal deflections. At that stage users should start to enjoy the best beam patterns and highest efficiencies that the telescope has ever produced.

It should be remembered that at 450 microns the aperture efficient of the JCMT is still quite low and there is a significant error lobe (a "skirt" around the main beam) which can cause misleading results. We are on a steep part of the curve where even quite small reductions in the surface error will bring considerable benefits. It is expected that, whereas in the past the errors in measurements were a substantial contribution to the overall surface error, this will no longer be so with the new measurement system.

Current information about the JCMT surface can be found on the Surface homepage.

---

back to:> March 2002 Newsletter Index

Click here for printable version.