MINUTES of the 17 January 2003 GBT Commissioning Meeting AGENDA 1. Spectral Baseline work -- Rick 2. PTCS project work -- Richard 3. Az Track work -- Bob A. 4. K-band and other commissioning -- Ron 5. Project and observing program scheduling -- John, Carl 6. AIPS++/Dish status -- Joe M. 7. Spectrometer status -- Rich 8. Software status -- Nicole 9. Any other business REPORTS & DISCUSSION 1. Spectral Baseline work Spectral baseline report 1/17/03 Reduction of the 12/27/02 L-band test of continuum source calibration transfer tests is now complete. These initial results indicate that calibrations are stable over periods of about 30 minutes and deteriorate slowly with time. These tests are to the sensitivity of a 5-minute-on/5-minute-off, 50 MHz bandwidth spectral line measurement of a 3 Jy source. Ratios of continuum source spectra near beam half-power points to beam-center spectra showed the expected spectral slope due to changing beamwidth with frequency. The elevation offset spectra showed about 0.2% rms ripple at the 2.3-MHz period corresponding to reflections between the subreflector and main reflector, presumably due to the specular reflections between the subreflector and main reflector moving on and off thge surface edges. Azimuth-offsets spectrum ratios showed no such ripple component but showed a bit of waviness on the 10- to 20-MHz scale. The 800-MHz bandwidth, X-band focus offset measurements initially reported last week have provided roughly 20 cm resolution of the time delays associated with the multipath interference of spillover noise. Total-delay features are seen at 5.3, 5.9, 7.5, 10.5, 15.8, 20.6 (broad), 27.7, 29.2, 30.0, 30.3, and 31.8 meters. Seven of the eleven delay features have been tentatively identified with various scattering and reflection paths involving the subreflector, particularly the top and bottom edges which are tangent to constant-phase locii. The rms ripple amplitude associated with each of the multipath interferences is on the order of 0.1% of the system temperature, or about 30 mK. Normally this ripple cancels out in an on-off measurement of a source with little or no continuum flux, but this effect is part of our investigation of the structure in an (on - off) / off measurement of a strongish continuum source. Presumably these multipath effects are stronger at lower frequencies, but it is hard to get enough RFI-free bandwidth to resolve the individual components. Tsys/Tcal measurements at X-band show 5% peak-to-peak ripple with a 280-MHz period (108 cm multipath delay distance) and about 1% p-p with a 23-MHz ripple period (13 meter delay distance). The long-period ripple might be associate with the length of the X-band horn, which is about 56 cm, but this is only preliminary speculation at this point. The origin of the 23-MHz ripple period is presently a mystery. The 8400-9200 MHz (on~-~off)~/~off spectrum of 3C48 shows a lot of fine-scale structure that would seems hard to explain with reflections within the feed system. We need to confirm that this is upstream of the mixer. Tsrc/Tcal spectra taken at different overlapping frequency ranges have significant offsets from one another that need to be investigated. A number of stability tests were run on 1/14/03 using the DCR. Routine total power stability checks on the IF channels found nothing remarkable, except that one fiber channel (ch 3) showed a step in transmission near 5 degrees elevation. The X-band receiver tracking cold sky showed fluctuations of a few percent over ten minutes correlated across both polarizations. These fluctuations could easily have been atmospheric noise variations. Similar variations were seen in Tsys/Tcal. Similar measurements at Ku-band showed that one channel (R2) exhibits brief dips in total power at irregular times, and needs follow-up. Fluctuations in total power of a few percent over ten minutes, correlated across the other three channels, were also likely due to atmospheric noise variations. On 1/10/03 the IF noise source was used to investigate the stability of the 2.4 MHz ripple in the optical driver modules (ODMs). The original external modulator (ODM 2), the new external modulator (ODM 6), and the direct modulator (ODM 8) were tested using the spectrometer in four-bank mode for a continuous time duration of ~3.5 hours with 30 second integrations. The data are currently being analyzed. We are also in the process of reducing data on C-band focus-offset and continuum spectra data and Ku-band focus-offset data. The C-band receiver was removed from the telescope this morning (1/16/03). New cryo amps will be installed during the next few weeks. Discussions with Rich Bradley on a noise model of the GBT receivers continue with useful preliminary results. His calculations deal mainly with how waveguide-loss noise couples into the feed/OMT/amplifier system in a frequency-dependent way. Similar coupling effects need to be modeled for external noise sources. -- JRF Discussion: Rick added in the discussion about the fiber modulators that although he new ODM has about 20dB lower ripple than the older units, there is still room for improvement. They are investigating whether this could be achieved through temperature stabilization, etc. Roger noted that the Ku-band receiver was being pulled off for repair of the bad channel. The receiver is needed again at the end of February, but should be back well before then. Phil asked if the ripples with origin in the optics can be eliminated by focus modulation. Rick stated that since some of the ripples involve two reflections, it will affect the amount of the modulation. In general, the month of February will be spent trying out mitigation strategies for the baseline effects found so far. 2. PTCS project work The meeting to discuss the PTCS system design has been scheduled for Wednesday 22nd. This meeting is "by invitation only" simply to try and keep it well focused. After the next iteration, the proposal will be circulated more widely for comment. Kim has circulated a proposed spec for the GBT Data Acquisition and Control package (a generic small-footprint common interface for telescope instrumentation). The accelerometers have been re-installed on the antenna, and plans are being made to restart the structural vibration testing. Satish should be in CV for discussions with Fred and Lee on Friday (weather permitting). Amy, Melinda, Joe B., Don and myself met to discuss the Active Surface Charter. For the next development cycle, the SDD will continue to work on the 100ms control loop, and also work with Don to resolve the "projection to normal" issue, and update the FEM interface to Don's latest code. We will then be in a position to compare the different models back-to-back. Don will contact Fred Schwab to discuss the projection question. Jason and J.D. continue to service actuators; currently there are 16 inoperative, all of which required physical inspection. This will be scheduled when the weather improves. RMP is making slow progress on the commissioning data back-log, but should have some preliminary results to show at the meeting. -- RMP Discussion: Richard showed a number of focus and pointing curves. In general, these curves looked quite good, although there was still a sizeable pointing vs. elevation curve. There are still more data to analyze. The concensus was that the present performance was adequate for K-band observations this winter/spring, but that focus and pointing would need to be checked frequently. This will be refined by next fall. 3. Az Track work Phil reported on a new problem that has arisen with the azimuth track. Yesterday, as the telescope maintenance crew was about to lift a wear plate to shim under it, they noticed a 3-inch crack on the top surface of the plate, propagating from the end of the wear plate, more or less down the length of the plate. The plate was lifted, and the crack was found to be about 11-inches on the bottom. When the shimming was done and the plate replaced and torqued down, the crack propagated farther. The plate was lifted again, inspected, and re-torqued. The crack is now 15 3/4 inches on the top side. The wheel has been run over the crack numerous times now, and the crack has not propagated any further. In addition, small cracks were observed on the wear plate on the other side of the joint. [During the course of the day, following the meeting, the wheels were run over the crack numerous times. The crack did not propagate any further. The engineering team believes that it is safe to operate the telescope for the time being, provided that the crack is monitored every few hours, and that we avoid observing programs that cause the wheels to pass over the affected joint an excessive number of times. We are trying to procure a replacement plate as rapidly as possible, and will also initiate a program to strengthen the joints in an effort to avoid cracks at other joints.] Phil also noted that a project team for the az track work had been formalized earlier in the week. The core team consists of Lee King, Bob Anderson, and Dennis Egan. Lee will serve as Project Engineer and Bob as Project Manager. 4. K-band and other commissioning There have been no commissioning activities this week. Frank is on vacation as well and no commissioning activities were scheduled on the telescope. When one considers the limitations imposed by the M&C and Aips++ software in multi-feed, multi-IF, IF-balancing, and multi-bank software, we've done about 3/4 of the observing we need to do at this time. Two more stints at commissioning, hopefully this weekend if the weather is good, will finish K-band commissioning until we get these other capabilities. Next, we have to find the time to process all of our recent data. Next week we return to observing checkout and training. -- RM 5. Project and observing program scheduling This week I have been reconciling our current plans with the plans we outlined in the GBT review meeting. Fortunately, they have not diverged too much. Monday, January 13th we discussed at length the ongoing high-priority projects: PTCS, Az track, and Baselines, as well as our K band commissioning project. All are making good progress, some better than others. There was concern voiced by many in the group that our reach is exceeding our grasp, that is, too many ideas are clamoring for attention when we have too few resources to handle them. We will be working this problem in the near future, and our clear priorities will serve to be the deciding factors. Monday, January 20th, we will discuss milestones for February for each of our highest priority projects: Az. track, Baselines, and PTCS. These milestones will be published in the GBT news page. I have not published the January milestones, due to a bit of indecision and incompetence on my part. This week's activities got a bit scrambled due to a couple of factors: 1) The M&C group needed more test time for "sponsors" to test the requested modifications and bug fixes. In the future, all sponsors of sdd MR's should plan to test whatever it is that they have requested, and schedule time accordingly. 2) Mother Nature fouled up plans to paint the track baseplate with a zinc paint, for the pupose of testing the anti-fretting properties. This required rearranging observing and maintenance, for a possible try at it next week or the following week. It may prove infeasible to paint outdoors in January in GB! In spite of this, there seems to be enough telescope maintenance and commissioning time being made available to the projects that need it. Everyone's patience with these last-minute changes are appreciated, and I'd like to remind everyone that you should be thinking a week in advance where possible, in order to avoid disappointment with being denied time to work on you project. -- JF Last Week ======== Completed about 15 hours of Pulsar Observations Successfully completed the second of three radar experiemnts on Mercury January ===== Asronomy observations remaining ~ 25 hours (mostly pulsar) February ======= Astronomy scheduled - 168 hours (includes K band) no backup projects yet scheduled Maintenance - 106 hours Test+Comm - 36 hours Un scheduled commissioning time - 247 hours (all required) March ===== Not yet started -- RCB Discussion: Phil noted that Jean-Luc Margot was very pleased with the two bistatic radar runs, Mercury observations with Goldstone, and near-Earth asteroid observations with Arecibo. Jean-Luc has preliminary asteroid observations posted on his web page, which indicate that the asteroid is probably a contact binary. The polarization team, Troland, Robishaw, and Crutcher (and Heiles, working from Berkeley), seem pleased with their results. At L-band, the beam squint is measured at ~0.1%, which is probably within the measurement error. A typical squint value for other telescopes is ~1%. The beam efficiency is also ~100%, owing to the low sidelobes. It appears that those on the GBT project team who calculated the optics got it exactly right! 6. AIPS++/Dish status o Stable link has been changed to point to build 18.391. o Multi-x discussions. There was a coordination meeting on Tuesday to help converge on the outstanding issues. There will be a note on this posted to (Note 254 if all goes well): http://aips2.nrao.edu/daily/docs/notes/notes/notes.html This describes the details and decisions made. One key point that became clear was that we could better exploit some knowledge of the GB observing system for the multibank modes. A new target has been added for next cycle to enable filling of multibank data to a single MS. o January targets - bob gbtmsfiller: support bs data 30-jan-03 bob gbtmsfiller: support multi-beam data 30-jan-03 The FEED table and associated columns in the MAIN table are basically in place for the multi-beam, beam-switched cases. Available for testing early next week and projected to be on time for the snapshot at the end of the month. bob gbtmsfiller: multi-bank data in single MS 28-feb-03 Added target. joe beam switched calib refinements 30-jan-03 joe beam-switched imaging 30-jan-03 Bob has added filler changes for dealing with several of the existing procedures properly, though still across banks. Initial work on updating the calibration is underway. We are also working on a document similar to the above to better describe the calibration. Ongoing. Some initial work on jim bs calibration refinements 30-jan-03 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). o User Support jim demo and training with Crutcher, Troland and Robishaw o Defects resolved: Use http://aips2.nrao.edu/ddts/ddts_main to examine enhancements to GOpoint displays (including Tsys listing) (defect list provide at meeting). -- JMcM Discussion: Steve added that the first draft of the AIPS++ GBT Requirements Document is out. It can be found at http://www.aoc.nrao.edu/~smyers/aips++/gbt/naug_gbt_eval.pdf 7. Spectrometer status The sources of the 100 MHz comb in the 800 MHz bandwidth data has most probably been found. If anyone sees this again please let Holly or me know. The problems were traced to two chips in two Low Speed Samplers. The chips should have been "high speed version" but instead were "low speed version" (slower by 1.3 ns on average). Because of tight timing tolerances between the sampler and the correlator, the low speed chips could work in some sampler slots and not in others. During the modification process, Low Speed Samplers (LSS) were not put back in their original locations so the problem showed up after the modified LSS were re-installed. What do the LSS have to do with the 800 MHz data from the High Speed Samplers? The data from the HSS goes through the LSS to get to the correlator. The 1600 Mega-sample/second data is transmitted as 16 paralled 100 Mega-sample/second data streams. The slow chips caused one of these data streams to slightly corrupted, resulting in a feature that had a nominal period of 16 samples or 100 MHz. Two melted wires was also found in one of the samplers. It's not clear if these were involved in generating the comb. An attempt was made to duplicate the large steps that have been seen in the auto-correlation functions (acf)of some integrations. The data that has been examimed so far looks clean. Updated Correlator Card Firmware was installed in quadrant 0. It was successfully tested. Offsets in a two of the Low Speed Samplers were traced to mis-adjusted potentiometer. These were re-adjusted and re-tested successfully in the system. A day was spent on pulsar spigot system test. A pulse generator was used a a signal source. It was a long frustrating day during which very little worked. A complete lack of data was traced a a faulty spigot card. With the spare installed, and some firware manipulations, good data was seen at the output monitor point of the spigot card. However, the data received by the spigot computer was corrupted. The software scripts which were supposed to make test observations easier did not work because of conflicts within the software. Plans for next week include the following. We will attempt to repair the failed pulsar spigot card. The time allocated for pulsar spigot tests next Wednesday may be spent troubleshooting the system if the causes for this weeks failed tests are not found. The firmware in a second correlator control card needs to be updated for these tests. The task of modifying the High Speed Sampler Phase-Locked Loop will be organized so that it can be handed to technicians. The cause of glitches in the auto-correlation function of multi-integration scans will be sought. We will again try to reproduce the "large acf step" problem and to isolate it's cause if we can reproduce it. -- RL Discussion: Rich noted that the scripts that were tried to facilitate spigot card setup ended up causing conflicts with the serial port, etc. This may mean that the setup will have to be done manually. This will be discussed more off line. 8. Software status Software Development Division #20 - Friday, January 17, 2003 We are finishing Week 5 of a 5-week development cycle. This cycle culminated in the release of M&C v3.10 on January 16, 2003. Out of the original 5 features planned for release, 3 were successfully completed, 1 was deferred to the next cycle in favor of operational support for the spectrometer, and 1 was handed over to Ron (continuous real-time total power displays to support baselines work). 6 extra items were coded, tested and included in this release. Release notes are available at http://tryllium.gb.nrao.edu/release/MC_v3.11 We received many requests, and through a selection and release planning process based on Green Bank priorities, the SDD work plan for the next cycle has been identified. This can be viewed at http://tryllium.gb.nrao.edu/images/POR_Feb03.jpg. Requests which were not selected for this month's release will be considered once again at next month's release planning meetings. Many thanks to those who submitted requests or gave us information about your upcoming requests (including Frank, Glen, Ron, Tim, Gary). The first suite of regression tests (jointly developed between the SDD and aips++ groups) was successfully conducted this past Sunday and Monday. Beginning next month, SDD will be able to ensure that the new software it releases is capable of handling the same key astronomical scenarios as previous versions. New regression tests will be added as the M&C and aips++ groups complete the joint release of major new capabilities. (Complete status is available online at http://tryllium.gb.nrao.edu/status/20030117.html) --NM Discussion: Toney added a few words about the astronomical regression tests. Some problems were had with the Spectrometer, but these were probably long-standing ones (i.e., not caused by the new release). Rich asked if these were on the list to be addressed. They are, but have been bumped by other project work to this point. 9. Any other business Joe asked about the dates of the single dish summer school. They are roughly 10-17 August. The SOC has its first meeting next week. Hands-on experiments with the GBT will be featured, and this will involve aips++. Dana asked if polarization projects can now be scheduled with the Spectral Processor. This should follow after Heiles et al. finish analysis of their test data. They should be able to proceed with their own projects, using their own reduction software. Use of the Spectrometer in cross-pol mode is on the list, but with no firm schedule (perhaps by next fall). Same for aips++ data reduction software. PRJ 17 Jan 2003