|
|||
| Joint Astronomy Centre |
|
|
|
| Show document only |
Surface changes, June - November 2004The aim of this study is to find:
On 16 June 2004, eight surface maps were made with RxH3 between 17h36 and 01h24 the next morning. Roof and doors were opened at about 17h30 (there was some local fog and clouds). The first seven maps were made at 80 GHz, the last map at 160 GHz. During the night the CSO tau decreased from 0.16 to 0.10 but tau was quite noisy until 19h as can be seen in this plot, This can cause non-existing structures in RxH3 maps (e.g. stripes). The humidity at the JCMT increased somewhat, but not dramatically. The weather station at UKIRT, and the 350 micron tau meter did not work On 11 November 2004 nine maps were observed between 17h19 and 01h25 in the same way as on 16 June: eight 80 GHz maps and one 160 GHz map. The map parameters are listed in Table 2. Because I concluded that phase and amplitude data of calibration measurements do not add more information, I did not make these figures for this date (may do it later though). During the maps the CSO-tau at the JCMT decreased from about 0.10 and then remained fairly (but not very) stable around 0.05. The humidity decreased strongly during the first map from 80% to 40% and then varied only little during the next maps. In addition, 80 GHz maps were made on 5 other nights in June around shiftchange: 5, 6, 9, 12, and 15 June (here the date is the HST date of the beginning of the night), and on 8 August, 9 September, and 17 November. On those nights the tau was better than on 16 June and more stable (see table below). Also, 160 GHz map were made on 11 July, 8 August, 9 September, and 17 November. During most of these maps the phase monitor was not yet or no longer working. All maps can be seen in the reduction summary (click below 'surface'), as well as by clicking on name in Column 1 of the following tables. The scale is here from -50 to +50 micron and one can clearly see the structure (scalloping) of the individual panels. Using the new, preliminary, masking of 80 GHz maps mean rms values were derived for the maps (see tables below). Given are the average rms of the whole map (excluding the masked parts) and the adjuster rms. The time is for the middle of each map (with 80 GHz maps taking 50 minutes and 160 GHz maps 1 hour 40 minutes). Because of still unknown reasons the adjuster rms is larger at 80 GHz than at 160 GHz. Table 1. Maps of 16 June:
Table 2. Maps of 11 November:
Table 3. Other maps:
One aim of the project is to see what are the changes in the surface between opening roof and doors in the afternoon and midnight when the dish has stabilized. On 16 June the rms was quite variable (Table 1), but the maps do not contain clear stripes which are the signs of an unstable atmosphere). These data are combined in the figure below with the results of 11 November (Table 2) and 16 April (discussed separately because an adjustment was made on 24 May). This figure shows the rms of the surface as function of time for the three days when the surface was monitored, as well as for other maps in the period April - November 2004. The circles indicate 160 GHz maps. From days when consecutive 80 and 160 GHz maps were made, it appears that the rms in 80 GHz maps is on average 1.3 micron larger than in 160 GHz maps. The red curve shows a running average (over three datapoints) of the rms, using both 80 and 160 GHz maps, correcting the latter with the above amount. The 80 GHz rms decreases from about 34 micron after opening to about 26 micron around midnight, with the largest change occurring before 20h HST. This is consistent with the results of a preliminary analysis of SCUBA FCF's (from 2003) at 850 and 450 micron. On a number of days maps have been made in the afternoon just after opening roof and doors, which maps can be compared with a reference map made after midnight on 20040909. This comparison is made using these difference maps. At the bottom of that page there is also an analyses of temperature measurements. In general the changes in the surface are very similar on all these days. Maps made before the first adjustment based on RxH3 data are very noisy and only one example is shown (for 20020827). The conclusion is that the changes of the surface at the beginning of the shift are mainly due to the well-isolated centerbeams connecting the counterweight and the telescope through the cabin. Three series of plots show the differences between the maps (with right the difference of the derived moves (make window wide enough). They are on a separate page because the time to load this page increased too much. First I subtracted each map from the first map. This was done both for the observations of 16 June and 11 November. The times indicated above the plots are again for the middle of the maps. All colour plots are scaled from -50 to +50 micron. The map roughly show the same structure as similar maps obtained on 24 April: positive differences near the centre of the dish and near the edges, but details differ: the location of the sectors with the positive differences in the outer dish are not the same. In addition the last difference maps (map 1 - map 6,7) show arc-shaped structures. The second series of plots shows (at the above link) the difference between subsequent maps. Many of these plots show a striking striping pattern, as was also seen on previous occasions. These difference maps show decreasing values with maps 4 and 5 almost equal. But the next map shows again increasing differences. This reminds of the effects seen on 15 June 2002 (these maps were combined with tests of panel movements) and 24 April 2004, where also the difference between the last maps again increased. The third series of plots show differences between different dates at 80 GHz and 160 GHz. Many of the difference maps show a more or less symmetric pattern. While some of them also show the well-known arched striping pattern, others show a large difference in the upper left or right part of the dish. Finally, there are difference plots between 160 GHz maps from 2003 are shown here. The aim was to find which structures are visible in these maps. made on 11 February, 9 April, 19 May, 30 June, and 7 September. In between these maps, adjustments have been made on 5 March, 14 May, and 30 July. To study the reliability of these maps, I show separately an analysis of the difference between seven pairs of 80 and 160 GHz maps. These maps were all made subsequently on the same day around midnight, so the telescope surface is expected to be stable. The conclusion is that differences between those maps match at 80 and 160 GHz, except for some days, such as 16 June 2004. Table 4. The rms values of the difference maps.
An analysis of temperature measurements of the telescope for the same period is given here. I divided the quality of difference maps into three catergories - good, aver, and bad, depending on the structures seen in the map (see below), in a somewhat subjective way. The rms in difference maps which are 'good' is in the range 8 - 16 micron. Structures seen in (some) difference maps:
Possible causes of structures in maps: -Temperature changes of dish -Weather: We know already since the first RxH3 maps that unstable tau values are correlated with horizontal stripes in maps. -Scanning (pointing errors) -Other effects, e.g. bleeding of the signal Typical relative temperature changes of different parts of the dish:
The influence of the atmosphere: calibration measurements: During all maps, 8 (80 GHz) or 14 (160 GHz) 2 second calibration measurements are made at a fixed offset position. These measurements were analysed to see whether they show effects due to the weather (causing the stripes and other defects in the maps). The conclusion seems to be that bad weather causes variations in most cases on time scales longer than 2 seconds. The RxH3 log gives one map which was mad in particularly bad weather on 20030613. This figure shows a differencemap compared to 20040808: there are strong stripes, and possibly one arc in the upper part of the figure. The cause of the arclike structures is not clear. They appear to be much more prominent at 80 GHz than at 160 GHz. Although in a dish which is not flat, one might expect non-linear deviations to occur, im my opinion these structures vary too fast to be real. This is particularly well visible in the difference maps at 80 GHz and 160 GHz for 16 June (wrt 8 August) where the strong 80 GHz arcs are invisible at 160 GHz. Conclusions: -The rms of the nighttime surface varies from night to night by 8 - 15 micron (as seen in difference maps), or less (due to defects in these map such as stripes and arcs). -The (80 GHz) rms of the dish varies from opening to midnight from about 34 to 26 micron (the probably more realistic 160 GHz rms is 1.3 micron smaller); during part of the period the 160 GHz rms was better than 23 micron at night). -The best way to improve the surface in early evening is to remove the cause of the changes, i.e. the large changes in temperature difference between centerbeams (in the cabin) and the other parts of the dish. -Alternatively, with some more measurements, average corrections (moves tables) could be derived, but this would mean frequent adjustments (e.g. every 30 minutes), which process then has to be automated. -The best way to improve the nighttime surface is to repair four adjusters which have large errors. -Further improvements can be made when the cause of arclike structures in RxH3 difference maps can be found, which now prevent more accurate studies in many cases. |
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
