| No | Date | Project Code | Observer | Support Scientist | Description & Scheduling Blocks | Response |
|---|
| 1 | 04 May 2006 | None | N/A | N/A | Roger reported a problem with the motor rack manager. The standard procedure to update motorrack control settings is to turn the manager(s) off then back on. He had set the default state of PS manager #2 to off, but when he turned the coordinator off/on; it also turned PS #2 manager on. JoeBrandt | The motor rack manager needs a 'subsystem-select' parameter, so that subsystems which are not in service can be masked out of normal operations. JoeBrandt
Checked the code, and this is already done. I was unable to reproduce the problem reported. JoeBrandt
|
| 2 | 08 May 2006 | AGBT06A_061/063 | Zwaan | O'Neil | The observer wished to look at two lines - HI (1420 MHz) and OH (~1612 MHz) simultaneously for source with very high redshifts (z > 2). Since we have only one LO, and everything is based off that, the frequency correction was made only to the first line (HI), and the second line (OH) was shifted out of the narrow filter band he had chosen. One of two things should be done about this:
(1) at the least, a warning message should be produced stating that the 2nd frequency will be out of band
(2) at best, the CR LOs should be taken advantage of, the both frequencies should be set to the center of the band. -- KarenONeil | The issue is in the use of the config_tool and the use of catalogs. If no velocities are set in the configuration, the filters etc. are based on the frequencies without the velocity adjustments. When a scan procedure is run using a source in a catalog, turtle changes the velocity values directly in the LO manager. This can result in this problem. Several possible solutions were discussed with KarenONeil. This issue should be presented to the observing issue committee. Main.Melindamello |
| 3 | 14 May 2006 | AGBT06B_019_02 | Minter | Minter | I tried changing the attenuation levels on Converter Modules 4 and 8 but this did not
change the readings of the Converter Module power levels. Either the samplers for the power levels is bad or the attenuators are not changing values. | This has been determined to be a hardware issue. -- RamonCreager |
| 4 | 22 May 2006 | AGBT06A_067_04 | Lovell | Ghigo | Antenna FITS files missing for scans 107-115. GO FITS files seem ok.
Spectra look ok. What happened to the
Antenna files? | Joe has started looking into this but has not found anything yet. PaulMarganian |
| 5 | 22 May 2006 | AGBT06A_067_04 | Lovell | Ghigo | Scan 132 is unreducible by both DISH
and GBTIDL. This may be a case of a scan starting on a minute boundary. Reference the messages from Bob and Jim:
My guess is that it's certain combinations of :00 seconds with some hours that round strangely. In this example, the start time seems to be 16.4 hours which should result in 16:24:00 but apparently resulted in 16:23:00 in the M&C file names (the GO file name got it right, though). In fractional days, this works out to 0.68333333.
-Bob
James Braatz wrote:
Hi Bob, Frank, Ron,
I'm going to include the SDD here. There was a known bug relating to scans that start on minute boundaries. I tried to reveal the bug during the last regression test by using a track scan with a specified "StartTime" ending in ":00" seconds, but my test showed no problem. Maybe this data set will help find the bug.
Jim
Bob Garwood wrote:
I don't see the GO FITS file for scan 132.
The ScanLog?.fits file has this path for the GO file: ./AGBT06A_067_04/GO/2006_05_22_16:23:00.fits and that file isn't there. However, there is a GO FITS file that does have scan 132 in it:
./AGBT06A_067_04/GO/2006_05_22_16:24:00.fits
There is no other 2006_05_22_16:24:00 timestamp in this data.
So, that's why the aips++ filler complained about it. sdfits's apparent behavior in this case is to skip this scan completely - so that's why gbtidl can't reduce the data. dish has trouble reducing the data, even though it has been filled, is that it is missing the GO information and d.calib uses that to work out why type of calibration it should try (FS, PS, Nod, TP). However, you can force the issue in dish by using FScal:
d.FScal(132)
that works and seems to produce something reasonable.
A better alternative, though, would be to simply rename the apparently mis-named GO FITS files.
Someone in M&C should be told about this because it's clearly a bug. My guess is that this time was a round up from 59.99999 seconds or some such and that either the code that generates the GO FITS file's name got it wrong, or the code that generates the name for the M&C files got it wrong. Given the times in the FITS files, I think the M&C timestamp is wrong and the GO one is correct. Even so, renaming the GO FITS file is a faster fix that should make sdfits/gbtidl and gbdish happy (the data will obviously need to be refilled).
In the Antenna FITS file for this scan, the first value of DMJD in the ANTPOSGR table is 53877.683333 which is 16.4 hours or 16h 24m, not 23m (time's from the SP FITS are similarly after 16:24 not 16:23).
-Bob
Frank Ghigo wrote:
Our observer today (Amy Lovell) has some problem with
a scan, which neither gbdish nor gbtidl will reduce.
The scan is number 132, project AGBT06A_067_04.
This is one of a long series of scans, with no anomalies
immediately before or after.
gbdish complains about the GO FITS file for scan 132,
but I looked at the GO file and it looks normal to me.
Let me know if you can find what the problem is.
-- frank
| The scan did indeed run at 16:24:00 and the function that computes the time string for all of Ygor generated 16:23:00 because of a rounding error. A tentative correction is in place for the next release. MarkClark |
| 6 | 30 May 2006 | AGBT06A_018_10 | Balser | Balser | During peak/focus scans using the DCR the IF rack channel 3 is having problems balancing. We are at Q-band with a frequency of 43.4 GHz and a bandwidth of 320 MHz. The attenuation gets set to zero with an IF power level of 10 Volts (saturation). We were able to get 1 Volt by manually setting the attenuator to 12 db. During spectral line observations the IF rack 3 channel seems to be fine.
Dana | Beats me. I will try to reproduce it in September when the Q-Band goes back up. Two differences between the continuum and spectral line balances were the bandwidths and target levels (1 vs 5 volts). MarkClark |