Minutes of GBT Commissioning Meeting 6 December 2002 N.B. Those giving oral reports at the Commissioning meeting are now providing me with written versions for inclusion in the minutes. My additional comments will be inside []. -- PRJ 1. Spectral Baseline work From Rick: Tests of IF gain stability showed the analog filter rack SF modules to be stable to about 1 part in 10^5 over 10 minutes. The converter rack modules are about 5 to 10 times worse than this with more variation from module to module. Two converter rack modules were removed from service for repair because of excessive gain fluctuations. None of these gain changes are suspected of causing spectral baseline problems, but a better understanding is prudent. Anomalies in continuum flux density measurements has led us to make system linearity measurements, which will require several maintenance days plus time for correction. One fiber modem is in the lab for linearity checks. We have seen spectral baseline ripple periods in the 50 to 80 MHz range on IF channels 1, 2, and 3 in the course of baseline observing tests. Further isolation of the cause of these ripples within the IF system remains to be done. The "improved" optical modulator is back from fiber splicing vendor. Initial lab tests do not detect the 2.4MHz gain ripple, but we need more sensitive spectrometer measurements. Mechanical and electrical modifications of the X-band receiver are underway to incorporate improved (more stable) InP amplifiers. Interim InP amplifiers have been received from the CDL. Lower noise units are expected by the end of the year. S-band receiver channel 2 is less gain stable than channel 1 (2% vs 0.5% total power drift in 20 min), but this doesn't seem to affect spectral baseline quality on cold sky. Continuum source (on-off)/off spectra show structure on scales of 120 MHz (4% p-p) and 10 MHz (0.5% p-p) plus one or more weak (0.5%) dips ~2 MHz wide within the observed 200 MHz. All of these features were found to be ahead of the mixer by changing the first LO frequency, and we are looking into feed/OMT/amplifier match issues without any definite suspects at this point. At least some of the 10-MHz period ripples change phase when the subreflector is moved, and the 10-MHz period is consistent with the 15.5-meter distance of the subreflector from the feed. The 120-MHz period baseline structure did not change with subreflector position. There is a 20-30% discrepancy in Ta/Tsys for continuum sources measured with the two receiver channels which needs to be resolved. The K-band R1/L1 feed has had one baseline test run using 800 MHz BW. Aside from IF mentioned in the first paragraph, the 5-minute on/off cold sky baselines look fairly good, but these need to be remeasured with a cleaner IF. Continuum source spectra show about 10% slope across 800 MHz in both channels and a couple of narrow dips of this magnitude. Channel 1 shows a ripple of about 100 MHz period with p-p ~1%. Channel 2's most obvious ripple is about 30 MHz, 2% p-p. A 2% jump in total power was seen between on and off scans, which resulted in a strong 35-MHz periodicity. Further tests to isolate effects remain to be done. The C-band has had one baseline test run using 800 MHz BW. The 5-minute on/off cold sky baselines show 50- and 80-MHz baseline ripple periods which also appear in the IF noise source spectra. Also, some slope develops in the spectra after about 10 minutes. Continuum source spectra show about 15% slope across 800 MHz and a very prominent ~68-MHz period, somewhat irregular ripple of about 3% p-p in the continuum flux in both channels. There is a roughly 25% discrepancy in Ta/Tsys for continuum sources measured with the two receiver channels. Three different continuum sources show very similar baseline and flux density measurement anomalies. Initial work next week will concentrate of comparing the old and "improved" optical modulator performance, including a more systematic measurement of fiber bending effects. 2. PTCS project work From Richard: Kim Constantikes, Jim Condon and myself continue to refine our ideas on the material which will be required for the Conceptual Design Review, and producing first drafts of this documentation, which will then be discussed by the larger project team. These include refined scientific objectives and the resulting system requirements; critical experiments; prototypes; the PTCS system design itself, and so on. One of the key prototypes is the Engineering Measurement System, which will demonstrate our ability to routinely convert LRF measurements to physically useful results in near real time. Kim has already proposed a high-level architecture and system design for the EMS, and Ray Creager will commence work on this next week. A (human) communication failure, complicated by the unexpected power outage, meant that the active surface system was not available for the start of K-band commissioning on 24/25th November. However thanks to the hard work of Jason Ray, J.D. Nelson, John Ford and Rich Lacasse, the system was put back into operation before Thanksgiving. We have now formed a dedicated project team of Jason, J.D. and Amy Shelton, who will perform a thorough review of the complete system, and address the remaining outstanding issues. As noted by Jim Braatz and Ron Maddalena, there are clearly some issues with the temperature dependence of focus, which will need to be resolved in the near future. In other areas, work continues to resolve the LRF calibration problem, to complete the phase closure experiment, and to update the FE model to the latest version. As a result of the review of priorities, starting this month the PTCS will have available at various levels of effort Joe Brandt, Ray Creager, Randy McCullough and Amy Shelton. This will allow us to make significant progress on the EMS, the Active Surface, and the new instrumentation package (whose first application will be to the accelerometers and the structural health monitoring program). 3. Az Track Work From Bob Anderson: * 3 splices were shimmed with teflon/bronze material this week (19, 20, 21) to give a good test section for teflon. Zinc shim material also ordered to use in a test in a similar section. * Research for a good zinc rich paint to test for anti fretting continues. * Inspection of a wheel bearing showed the beginnings of a wear pattern in the outer race. No shadows on the rollers, or brinnelling. * Next week's focus to be on plan and scoping for analyses. 4. K-band and other commissioning From Ron: Since our last meeting, astronomical commissioning has concentrated on K-band and polarization tasks. I'll concentrate on the K-band exercises, which occupied about three partial nights in the last two weeks. [Anish had several sessions doing polarimetry tests at L-band in preparation for Tom Troland and Tim Robishaw's (Carl Heiles's student) visit in early January.] Jim Braatz will probably comment separately on the 16 or so hours he spent observing last weekend. o The active surface has been put back into service. After a bit of fumbling, we've measured efficiencies, both with the surface on and off, that are identical to those derived last spring. o We've measured feed offsets for the lower section of the K-band system and verified that the new multi-beam antenna control software and GO procedures work. Beam-switched observing is now relatively easy to execute. o There's some indication that beam offsets may be a function of elevation, not unexpected since the effective focal length changes with elevation (which changes the telescope's plate scale). The measured axial focus was 20 mm from what we expected, probably due to the expected temperature dependence of the focal length. Parameterizing these affects have always been on our list of things to do. Eventually, the M&C system will need to incorporate the models the commissioners and the PTCS group derive for these effects. Next week, we will hold training sessions for Harvey Liszt and Barry Turner. Al Wootten plans to be in Green Bank toward the end of the week of Dec 16 for similar training. As I stated at the last meeting, the idea is for Harvey, Al, and Barry to learn enough about the GBT to check out and support their own proposals, and eventually help check out and support other K-band proposals similar to their own. Since those doing the training next week are also the commissioners, I'm not scheduling any significant K-band commissioning tests next week. The commissioning for the next month is sufficiently prioritized and planned out that, even with the limited resources due to the upcoming holidays, vacations, AAS meetings, etc., we can expect steady progress. [N.B. The web URL for GBT commissioning status items is www.gb.nrao.edu/gbt/gbtstatus] 5. Project and observing program scheduling From John: An infrastructure has been created to help schedule our development and commissioning projects. Microsoft Project skeleton projects have been created for most of the active projects, and a summary project to tie them all together is under development. The skeletons for the Spectrometer commissioning and development have been fleshed out, along with the PTCS, RFI mitigation. The rest of these will be fleshed out during the next week. Project managers are expected to keep these projects up to date with the progress of their projects. To better manage the scarce resources of telescope time and other equipment time, a consolidated calendar has been developed to allow us to more easily assign time on the telescope and other equiment. [The consolidated calendar tool is located at http://192.33.116.100/webcal/webcal.dbw This is only available through the NRAO Intranet.] There is a project scheduling team that meets Monday morning to revisit the week's scheduled activities, and to plan for the following week's activities. From Carl: Fri Dec 6 Successful completion (today) of GBT02C-024 ( HI Survey of Local Compact Galaxies). Even though they lost some observing time due to winds they completed all of their goals and appeared very happy with the data. December Schedule ============== Astronomy - 200 hours (completed ~ 64 hours) Maintenance - 143 hours Commissioning & tests - 364 hours January Schedule ============ Astronomy - 35 hours Maintenance - 142 hours Commissioning & tests - 567 hours (227 hours scheduled) 6. Software status Nicole submitted the weekly Software Development Division status report. Highlights are: - Next M&C release (v3.9) is on Wednesday, 11 December - Deadline for requests for the February release is 13 January 2003 (the deadline for the January release has passed) Active projects include: - Active Surface Improvements Project charter under development by Amy, J.D., and Jason - Spectrometer root cause analysis and improvements - Spectrometer Multibank Capability New code completed on schedule. Test runs proceeded without error for up to 12 hours before they were stopped when the time ran out. - Regression test suite development Will not be released this cycle. Work will resume in January. - Software Project Office Online Bugzilla replacement to be released on 16 December. - Antenna Control Unit Improvements January release is planned. - Q-Band Receiver control Release of the Manager and related IF changes deferred to January to allow testing. Code is complete, however. 7. Spectrometer status From Rich Lacasse: The Phase Locked Loop of a high speed sampler was optimized on the bench. Plots of the spectral purity of the 1.6 GHz oscillator will be shown at the meeting. Considerable improvements were made to the spectral purity of the Sampler Clock generated by the Phase Locked Loop. This is important because the during the sampling process, the spectrum of the sampler clock is convolved with the input signal. The sampler was tested in the spectrometer and found to work satisfactorily. The nine additional High Speed Samplers will be modified as technician time allows. All the 800 MHz and 200 MHz modes on John's "top 16" list have been tested, except for those which involve multibank. We tried one multi-bank mode. The system did not crash, however Rick's analysis program could read data from only one of the four data files generated. We saw another one-lag acf glitch. This is an intermittent problem that somehow needs to be fixed. One sampler has a large glitch in the frequency domain. The cause of this will be investigated after the meeting. Plans for next week include - updating documentation - re-testing spectrometer modes to verify M&C version 3.9 has not had unintended consequences. - testing multi-bank modes. 8. AIPS++/Dish status From Joe McMullin: Targets for upcoming stable release ----------------------------------- ips: support abcissa ranges braatz, jim 80 15/12/02 enhance aver task for improved weight garwood, bob 40 15/12/02 dish assay functions garwood, bob 80 15/12/02 multi-if implementation design mcmullin, joe 40 15/12/02 multi-beam implementation design mcmullin, joe 40 15/12/02 v2.0 newsletters mcmullin, joe 20 15/12/02 sd benchmarks mcmullin, joe 80 15/12/02 We are currently on track for delivering the above targets. There have been extended discussions on the improved weighting schemes, broadening to include revisions to the calibration (involving Maddalena and Liszt). Completed targets will be reported in all their glory in subsequent minutes. Working with Nicole and Richard, we are making efforts to better coordinate the GB/AIPS++ interaction. The current development plan has been approved for the short term; monthly meetings have been setup to maintain alignment. User Support ------------ Braatz has done the initial K-band observing this season (reports filed). He has also worked in support of the Pisano/Garland run. Defects/Enhancements -------------------- AOCso04156 spectrometer multi-bank problem (not a bug) AOCso04064 white space in source names causes problems AOCso04116 indexing of integration is incorrect (not a bug) AOCso03817 aips++ hangs up regularly while the gbt is operating AOCso04166 can't handle flagged data AOCso04044 crashes aips++ AOCso04162 d.fileout() reports a SEVERE error under normal conditions duplicate of AOCso04161, AOCso04034 AOCso04163 d.help() should be more helpful AOCso04067 header disappears AOCso04167 range values don't default to full range (not a bug) 9. Any other business To avoid a conflict with the weekly astronomy colloquium in Socorro, the Commissioning Meeting will henceforth be at 10 AM ET, 8 AM MT each Friday morning. There is another video-con meeting at 11 AM Eastern on Fridays, so we must end promptly at each meeting. PRJ 8 Dec 2002