Minutes of the GBT Commissioning Meeting 24 January 2003 Also posted on the web at http://www.gb.nrao.edu/GBT/gbtstatus/minutes.html AGENDA 1. Az Track Status -- Bob A., Phil 2. PTCS project report -- Richard 3. Spectral Baseline work -- Rick 4. K-band and other commissioning -- Ron 5. Spectrometer status -- Rich 6. Software status -- Nicole 7. AIPS++/Dish status -- Joe M. 8. Project Scheduling -- John 9. Observing Program Scheduling -- Carl 10. Any other business Minutes 1. Az Track Status For the Azimuth Track Update: o A total of 7 Wear Plates have been inspected. 3 have cracks. o The bottom edges are being ground to minimize possibility of crack initiation. (Bevelling tool on site). o Cold weather has slowed progress. o Magnetic Particle machine purchased to help identify cracks. o Target for completion is mid February, but more realistically late February. o Receiving quotations for replacement (in kind) plates. o Requests out to 3 companies for baseplate machining and welding approaches. Other activities: o 2 Meetings held on bolt strain gauge test. Test devices being instrumented by Metrology. Metrology to approach UVa about helping with calibration. Test scheduled for February 3. o Telescope weighing discussed at same meeting. Weighing targeted completion February 28. o Wheel surface profiling also discussed. Metrology would like to do this during weighing. o Wear plate strain gauge test discussed at same meeting. Targeted completion March 28. -- RA Discussion: Barry asked about the status of the plan to address the cracks and other track projects. Phil replied that a plan for the overall track program had been developed over the past several weeks and that the specific plan to address the cracks had been worked out over the past several days. The plans were to be discussed with Fred Lo on Friday afternoon. 2. PTCS project report We are making excellent progress on the System Design, Active Surface, EMS and Antenna Instrumentation work packages. Kim presented the proposed high-level system architecture in a meeting held on Wednesday. The design was extremely well received, and we will continue to develop it, including elaboration of the scenarios, interfaces and simulation. As the design proceeds, we are starting to identify "person-sized" components which can be allocated to other team members for further investigation. The Active Surface and EMS charters are complete, and the SDD are working on both of these. Don Wells will deliver the complete updated (97f) FEM package by 31st January, for release in February. We will be meeting next week to finalise the spec for the Antenna Instrumentation small footprint Data Acquisition System. Ron and Frank have continued to characterize the Y-focus dependence on elevation and temperature, as described in Ron's report. We will incorporate these results into improved focus-tracking coefficients. RMP performed a few additional calibration observations, looking at the elevation dependence of X-focus; these data have not yet been analyzed. Performance and analysis of commissioning data remains a problem area, we plan to get Jim Condon more heavily involved to address this. -- RMP Discussion: Phil commented that the PTCS discussion had gone very well in his view, and that both the presentation and participation by the audience were effective. 3. Spectral Baseline Work Spectral baseline report 1/24/03 Simultaneous operation of the DCR and spectrometer in GBT observing is working to first order. A problem with data drop-outs in the cal phase at the end of scans needs to be resolved. Concurrent use of the DCR and spectrometer will help diagnose temporal stability problems with spectral baselines. Linearity checks/retrofits of the optical driver modules (fiber modem transmitters) is nearly complete. Checks of the linearity of the Converter and Analog Filter Rack modules so far indicate that these modules are operating within their linear range, but these tests are still in progress. Higher resolution temperature monitors in the Equipment Room are now in place and are being regularly logged. Little improvement has been achieved on the stability of temperature control so far; this may require the installation of proportional hot water valves, but Bob Simon continues to experiment with the controller. We know of Equipment Room temperature effects on spectral baseline stability, and we are investigating the best combination of improving the room temperature control and reduction of the temperature sensitivity of the electronics modules. Reduction of the 12/30/02 C-band test of cold sky focus offset and continuum source calibration transfer tests is now complete. The results are consistent with similar tests conducted at L-band. The amplitudes of the reflections from the circumferential panel gaps are down by factors of 4 to 10 at 5 GHz as compared to 1.4 GHz. Two or three relatively weak ripple periods in the 7 to 30 MHz range remain to be fully understood. A series of on-off total power measurements of a 5-Jy continuum source over 3-1/2 hours showed a remarkably stable spectrum, and the transfer of this spectrum to the calibration of the continuum spectrum of another source about 45 degrees away in elevation was reasonably good. Further study of this calibration strategy is required, however. A second look at the continuum source spectra from the X-band (8-10 GHz) receiver shows the fine scale structure mentioned last week has a cause that is upstream of the first mixer. Full analysis of these data is still in progress. Cold sky measurements of the Ku-band (12-15.4 GHz) receiver showed a lot of 30-MHz period ripple in the total power spectra of two receiver channels and an unusual amount of spectral baseline instability in all four channels. We don't have enough diagnostic information to determine where in the full signal path to the spectrometer these problems may have arisen. This receiver is now in the lab for the repair of a previously known stability problem in receiver R2, and we have asked for test time in the equipment room to run cold load spectral baseline tests before the receiver goes back on the telescope. From Roger: Total power data was recorded for thirty minutes from both polarizations of the L-band receiver, as the GBT tracked what appeared to be cold sky. A half percent step in detected power was noted in one channel, but appears to be localized to the detector in ODM1. Fluctuations of a few tenths of a percent which were uncorrelated across the polarizations were seen. The fluctuations before the IF Rack appeared to dominate any fluctuations in the fiber IF or Equipment Room equipment. Some spikes in total power correlated in the two polarizations were seen, and probably were RFI. Total power data was recorded for thirty minutes from both polarizations of the S-band receiver, as the GBT tracked what appeared to be cold sky. The appearance of what was probably a RFI signal during the period makes it difficult to draw conclusions about the health of the equipment. We should devote more time to characterizing this receiver and the related RFI environment, using both the spectrometer and the DCR. Total power data was recorded for thirty minutes from both polarizations of the X-band receiver, as the GBT tracked what appeared to be cold sky. Nothing remarkable was seen. All eight channels of the K-band receiver were tested using the DCR. It was found that one channel, R1, fluctuated by several percent in a few seconds in the early part of the test period, but gradually settled down. There seems to be an intermittant problem somewhere between the feed and the IF Router. A few other less significant phenomenon were noted and need further study. Discussion. Phil asked about the observation that the standing waves from the panel gaps are worse at L-band than C-band. It might seem that these problems should be worse at shorter wavelengths. Rick stated that he did not fully understand this, but that it probably has to do with the offset of the feed from the optical axis. For the Ku- and K-band receivers, the feeds are offset, and this problem is not seen at all. 4. K-band and other commissioning We had a very full week with three observing sessions committed to commissioning activities. These sessions conclude the first wave of K-band commissioning tasks and we will need changes to hardware and software to start the second wave. The findings from this week include: o Offsets between beams as a function of elevation. Richard presented last week observations taken at elevations above 45d. Last weekend I made similar observations from 5d to 25d. A quick analysis of the data indicates that the elevation and azimuth offsets are constant with elevation to within about 1 arcsec. If this holds true, then the results suggests that the Gregorian house has no appreciable rotation with elevation. It also implies that the plate scale and, thereby, the effective focal length of the telescope remains constant with elevation. Since the y focus tracking algorithm implies we are changing the telescope's magnification by about 4% from horizon to zenith, the only way I know to keep the effective focal length constant is for the focal length of the primary to change by about 1m and in exactly the opposite sense as the magnification. o Frank determined how to set receiver gains within the K-band receiver in order to balance the power for the two I.F.'s when doing beam-switched observations. As Frank reported in his e-mail on Monday, there are some frequencies where no settings of receiver gain can balance the output power levels. o Wednesday night's observing consisted of using the Spectrometer to observe across the full extent of the lower and upper K-band receivers. The observations were 2-minute, 200 MHz wide, on-off measurements of the calibrator 3C147. The direct results of these measurements include: 1) the determination of the noise diode intensity in units of Jy (i.e., SCAL). 2) examination of bandpass shapes (a quick look at the data revealed two or three 'suck-outs'. 3) measure of RFI (none seen at first glance) 4) baseline shapes (very nice -- essentially ruler flat with an rms of 30 mK in 2 minutes on, 2 minutes off. One of the disappointments from these observations is that with the current I.F. power leveling, there's little chance for someone to observe with bandwidths that exceed about 400 to 800 MHz. I recommend that we don't plan on scheduling the large number of proposals that want to use multiple spectrometer bandwidths arranged in a series, an observing technique that synthesizes very large bandwidths. Thus, we're currently limited in the types of observing we can technically run at K-band. o Since Wednesday observing required maximizing the efficiency of the telescope, I measured the focus of the telescope every hour for 10 hours as 3C147 rose and then set. When comparing the results with the focus graph that Frank distributed a few weeks ago, I easily see the same trend of focus change with elevation. But, my focus measurements are consistently 10 mm different then his. Frank's observations were taken when the air temperature was 0 C and the sky was overcast (when radiative cooling is not effective and the air and steel temperature probably aren't too different). Mine were under similar sky conditions but at -10 to -15C. This suggests that the focus may be changing by 1 mm / deg C. The data also imply that the focus tracking curve we are using has a slightly wrong elevation dependence. The magnitude of the error is not unexpected and is about that of the standard deviation of the data that were used to determine the focus curve. -- RJM Discussion. Roger stated that he and Ron would get together early next week to discuss what needs to be done about the IF power balancing problem. 5. Spectrometer status Work on the spectrometer this week concentrated on two main areas. First, we looked into a problem identified by Balser where data processed in the third quadrant consistently had a small step at the midpoint of the acf (auto-correlation function) and had additional noise in the first half of the acf. Second we supported pulsar spigot tests. We were unable to reproduce the problem identified by Balser, but the spectrometer self test identified a problem in the third quadrant which may be related. It found errors in one correlator card. After a fair amount of trial and error, we found that these errors would move when we swapped the two LTA cards in that quadrant. We eventually found a XILINX chip of the wrong speed version on the LTA that produced errors. We replaced it with the correct speed version and the errors disappeared. With a little luck, this will take care of some of the errors seen by astronomers when using the third quadrant. The spigot card that failed in the system last week worked fine on the bench for an extended period of time. It was put back into the system and worked well there as well. Spigot system tests this week with Dave Kaplan of CIT successfully produced reasonable data using this card. Four modes were tested: one quadrant 16, 8, 4 and 2 bit modes. Artificially generated pulses were seen in all these modes. Kaplan plans to analyze the data more fully to look for subtle problems. We also confirmed what we suspected last week: that the spare card does not frame the data properly. Work is underway to isolate the cause of this. Plans for the coming week include the following: - fix spare spigot card - work on multi-integration glitches - work on low level structure in acf of some multi-quadrant modes. -- RL Discussion: Phil asked about the census of Spectrometer modes that have been checked out astronomically and are ready for use. Ron replied that all but one of the 16 modes on the priority list have now been checked out in single-bank mode, but that there multi-bank modes have not been checked out. Karen O'Neil, who joined the NRAO staff this week, will take this project on. Ron stated that there were 50-60 multi-bank modes that have been requested. Priority rankings will be established through contact with observers during program checkout. 6. Software status We are completing Week 1 of a 4-week development cycle, which will culminate with the release of M&C v3.11 on February 12, 2003 +/- 1 day. The SDD has made tremendous progress towards its commitments this week. MRs have been written this week for all of the release commitments, and most of the background tasks for this cycle (9 total), and are currently under sponsor review. Updating the FEM and reviewing wavefront to normal projection calculations has been collated into one development activity. Unit test development and coding has begun for frequency and polarization information in the IF FITS file, as well as the Active Surface 100ms issue. The group has also been actively participating in analysis. The Active Surface Charter is complete, and the EMS Charter has been submitted to Kim for review. A preliminary meeting was held with Brian Mason in support of the Caltech Continuum Backend CDR to determine the scope of SDD support this month. A large contingent of the SDD attended the PTCS design review meeting on Wednesday. The following were patched into the M&C product this week to support other GB needs: an update to the OnePps manager that allows users to change the sample rate for the OnePpsStatus sampler through the OnePps configuration file, correction of an X-band routing error, widening of the receiver filters in the IF Manager, and a change pertaining to LTA delays for the spectrometer. One network problem was resolved that was preventing communication between managers. As of today, the SDD is ahead of schedule for February deliverables. February commitments can be reviewed by going to http://projectoffice.gb.nrao.edu and clicking on the February graphics on the left hand side of the screen. (Graphical status is available online at http://tryllium.gb.nrao.edu/status/20030124.html) -- NR 7. AIPS++/Dish Status o Stable link has been changed to point to build 18.391. New stable available (Build 447). Stable Snapshot scheduled for 01/31. For details see: http://aips2.nrao.edu/daily/docs/reference/updates.html Default documentation is stable build with links to release. o January targets - bob gbtmsfiller: support bs data 30-jan-03 bob gbtmsfiller: support multi-beam data 30-jan-03 First version of correctly labelled multi-beam (and beam-switched) data. Currently being tested for the calibration routines (targets below). bob gbtmsfiller: multi-bank data in single MS 28-feb-03 Added target for next month's cycle. joe beam switched calib refinements 30-jan-03 joe beam-switched imaging 30-jan-03 jim bs calibration refinements 30-jan-03 Initial work underway. Worked a bit with Fabian Walter in preparation for his run 1/24-1/26. The existing routines should be adequate for this. After the additional indexing information is being used, a toggle between the use of the analytic and empirical calibrations for the beam switched data will be put in. sanjay multi-beam imaging 30-jan-03 An initial sketch of the non-switched multi-beam imaging has been done. We would like to try this on some data if available (a K-band map with TPWCAL or NOD:NONE:TPWCAL) -- Frank has pointed us at some continuum data available for testing the first draft of this. Update next week. o User Support jim,joe Walter, Carilli et al run. o Other Investigation of run-time fill/calibrate sequence. Investigation of 16-channel DCR fill problems. Enhancements to GOpoint (IARDS) displays. o Defects resolved: Use http://aips2.nrao.edu/ddts/ddts_main to examine AOCso04100 misbehaves when zooming over multi plots AOCso03839 writing optical velocities not possible -- JMcM Discussion. Roger asked if the problem in the IF Manager that was affecting the DCR might affect whether IF channels were getting identified correctly. There was some uncertainty about this, but it should be looked into. 8. Project Scheduling This week I started collecting milestones for the next month's effort. We have chosen to try and keep milestones based on the M&C release schedule in order to keep to a common schedule. I plan to publish these milestones next week. We agreed with the plan for the software over the next month. The track problems caused a large perturbation in the schedule for the next few weeks, with maintenance being allotted a larger amount of time. Looking out a few weeks, we plan an overnight shutdown for the end of February to do a track strain gage test, and to finish off the servo system RFI box. Overtime and double shifts may be required to finish the RFI box in one overnight period. The Digital group is readying itself for the task. The GBT resource calendar was moved to a new machine, causing a few problems for those with the page bookmarked. Reloading pages with the links and rebookmarking should have it working. -- JF 9. Observing Program Scheduling Last week ======= Observations for GBT02B-019, GBT02C34, GBT02A-062, GBT02B-021, GBT02A-049 Proposal GBT02A-049 is completed (radar of Mercury). February ====== Astronomy - 165 hours Maintenance - 159 hours Tests & commissioning - 349 hours Currently working on scheduling low frequency background projects March ===== Just started scheduling -- RCB 10. Any other business Jim asked if a decision had been reached about how to handle dynamic scheduling on the GBT (our manual version). Carl will send out the proposal again, and anyone wishing to make comments should do so the early part of the week of the 27th. We will make a decision at the meeting on the 31st. PRJ 27 Jan 2003