To: David Woody Douglas Bock Steve Scott Melvyn Wright CC: John Carpenter Stephen White From: Stuartt Corder Re: Optical Offset Pointing Summary: We repeated the Optical Offset Pointing measurements which are described in Carma Memo 34. The findings were similar to those desribed in that memo, namely, that while the evolution of the radio offset over sunrise and during the day can be large, the difference between the optical and radio offsets remains small. For eight antennas (C4 C5 C10 C11 C12 C13 C14 C15) the RMS is improved by more than a factor two. We also report on our attempts to utilize the existing optical cameras to observe stars during the day. Radio-Optical Tests Revisited: Radio pointing observations were obtained on the bright radio source 0530+135 at 100GHz. The azimuth and elevation offsets derived were applied to the system. Between the radio pointing observations , the very bright optical source, AOri, was tracked and the optical offset of AOri measured with the TV camera, relative to the newly centered pointing model was measured. The data was taken starting at 11.3 hours UT, finishing at 16.2 UT. The sun has elevation of 0 degrees at 13.95 UT. The radio pointing is plotted in Figure 1. Plots of the components of the optical to radio offset vector are also attached (two panel plots, one page per antenna). As expected, sunrise triggers some variation in the pointing model of the 10m antennas. These changes are strongly correlated with air (?) temperature, shown by the dashed line on the right-hand panels. The largest departures are in elevation. The RMS of the radio pointing measurements are summarized in Table 1. Optical offsets were measured for the entire track and are also shown in Table 1. The optical-radio offsets for antennas C1 and C2 were undistinguishable in noise from the radio data alone. Antennas C4 and 5 show noticable changes in elevation offset when the temperature increases while the effect of this change is mitigated, by a factor of 2, in the optical-radio offset. Antenna C6 shows some increase in scatter in radio points over sunrise but the effect is less pronounced and the improvement offerred by the optical to radio offset is smaller. For the 6m antennas, very large drifts, especially in elevation, were seen on a number of antennas. While some of these (C8, C9) seem to be triggered by sunrise, most are not. The radio pointing plots show the dashed verticle line as sunrise. Unfortunately, most of the 6m antennas were not able to follow AOri past sunrise with the exception of C11 (14.6 hours), 12 (16.2) and 13 (15.5). The radio pointing numbers in the table below have been corrected to only include the time range over which optical data is available for a given antenna. C12, which shows a very step like change in optical-radio offset exactly at sunrise, may have some problem with the flap machanism and it may be changing the orientation of the primary lens. In general, while the tracking of the optical to radio offset is not perfect (if optical reference pointing were perfect the optical-radio plots would be perfectly flat), it typically removes a factor of two or more from the peak to peak variation over time for the poorly performing antennas. The ratio of the total rms values below was found to be a good proxy for the factor by which the peak to peak variation of the radio pointing was reduced. At worst the optical to radio offset shows no improvement over the radio methods alone, consistent with the results of Memo 34. This imperfect removal may indicate some difference in the thermal changes to the entire dish and optical cameras but it may also be a result of the time taken to perform the radio pointing coupled with the rather large (10 degree) separation of AOri and 0530+135. In the presence of large errors in the mount model, more nearby optical counterparts should be used. A more detailed discussion of the implications of coordinate velocity, separation and time to measure radio pointing is included in an upcoming CARMA Memo (Corder & Wright). Daytime Optical Imaging: Over the last several days, we have developed a test rountine to coadd images and subtract the sky. This routine is not available for online use due to the methods used to grab and coadd the images. However, it is instructive to explore the brightness limit of our current cameras. A 10m antenna, C6, was used to observe some faint optical sources during the day. A unix utility (getremoveframe) was provided by A. Beard. It was used to dump a user supplied number of frames to FITS files. Then, the telescope was offset by approximately a camera field of view in azimuth and an equivalent number of frames were taken. The images were then read into IDL and averaged. The averaged background image was subtracted from the on source image. These methods are very slow though integration into the RTS would allow ~10 frames to be taken per second. We took 40 frames (20 on source, 20 off) for most of our observations incidating that our limits could be reached in 4-5 seconds plus the time to offset the antenna ~15'. A variety of stars ranging from V~2.25 to V~4 were observed on a partly cloudy day. Care was taken to observe in regions of less cloud cover but it is unclear if some of the failures were due to high clouds. The results are described in Table 2. We can clearly go to a V magnitude of 2.75 and we even detect a star at 3.5 magnitudes. There are about a 100 stars in the optical catalog brighter than 2.75 magnitudes. While the 3.5 magnitude star is not likely to produce a good centroid with our current methods, further work could take us to a 3.5 magnitude limit, providing ~200 stars with the current catalog. The daytime limit of the 6m cameras seems to be V~0mag though a handful of very red stars are also visible. The exceptions are C12 and C13. C13 has a 10m style camera. Conclusions: Daytime optical-radio pointing produced significant benefits over radio pointing alone. For eight antennas (C4 C5 C10 C11 C12 C13 C14 C15) the RMS is improved by more than a factor two. While the current system does not allow most optical sources to be seen during the day, we have demonstrated that, by subtracting the background and coadding the images, sufficiently many optical objects are made available as to surpass the on-sky density of 3mm pointing sources. Future Work: More work is needed to integrate this into the online system. On the software side, we need an additional option in the centroiding code that allows background subtraction and coadding. The current centroiding routine is optimized to center bright objects with excellent outlier rejection. We need a "faint" or "daytime" option. The porting of the routine down to the antenna computers will provide the increased speed needed to make the optical pointing measurement limited by the time taken to move the telescope to an offset position for background subtraction. On the hardware side, allowing the flap to open during the day on C13 would allow us to test the 6m pointing improvement later into the day, possibly all day. 10m style optical cameras mounted on all 6m antennas may then allow us to utilize daytime optical reference pointing. If we could open the C13 camera flap during the day, we would like to repeat the tests we report on here, taking the optical-radio offset measurements more into the heat of the day. Table 1: Antenna number is indicated in the first column. The next three numbers are the rms of the optical offset relative to the radio corrections for azimuth, elevation and combined. The next three numbers are the same but for the radio corrections only. The intepretation is that the rms of the optical offsets is what you could get if you were using the optical reference pointing. C12 was divided into pre-sunrise (a) and post-sunrise (b) times as the optical offset vector is strongly step like right at sunrise; possibly implicating a shift due to the closing telescope flap. Ant rmsOptRadDiff rmsRad Az El Tot Az El Tot '' '' '' '' '' '' 1 1.80 1.92 2.63 2.07 1.74 2.70 2 2.16 2.58 3.36 1.85 2.92 3.45 4 1.98 1.62 2.56 2.69 4.90 5.59 5 1.86 1.86 2.63 1.53 5.31 5.53 6 1.92 1.26 2.30 1.23 3.06 3.30 7 3.42 5.04 6.09 5.22 6.37 8.23 8 1.08 1.44 1.80 1.37 1.43 1.98 9 1.26 1.50 1.96 1.05 2.00 2.26 10 1.92 2.28 2.98 5.01 5.19 7.21 11 3.48 4.32 5.55 5.47 20.5 21.2 12 9.00 31.0 32.2 3.34 8.51 9.14 12a 2.10 2.04 2.93 4.09 5.88 7.16 12b 1.50 1.20 1.92 1.80 5.06 5.37 13 3.54 2.88 4.56 3.97 7.22 8.24 15 2.82 2.64 3.86 2.28 8.01 8.33 Table 2: Sources observed optically with a 10m antenna optical telescope. Some of the failures are likely due to patchy cloud cover. The brightest sources were seen in 20 frames on source. The one detection at 3.5 magnitudes was with 30 frames. Source V_mag Az/El See it? (HP######) degrees 100453 2.2 65/40 yes 102488 2.5 70/33 yes 079593 2.75 174/49 no 080331 2.75 7/65 yes(twice) 093747 3.0 108/44 no 086974 3.5 106/66 yes (marginal) 081833 3.5 76/82 no 087933 3.7 100/64 no 090139 3.8 107/56 no 092043 4.2 104/51 no Enclosure: pntfit_vs_temp_07oct04.pdf (radio only az and el offsets vs time) pntfit_rad_opt_07oct04.pdf (radio only az and el offsets vs az/el) (Both: Verticle line is sunrise, dashed line is temperature. Thanks Stephen!) carma#.pdf (az/el optical-radio offsets vs time)