The GBT antenna pointing model (as of PFM5) implements a hysteresis correction, which is based upon velocity. It has been found that the method used to generate this velocity input unintentionally placed a feedback path into the pointing model. In some rare circumstances, this produces a tiny oscillation in velocity, which is of course undesirable. This MR corrects the issue.
As noted above the GBT antenna pointing model requires a velocity input to compute one component of the pointing model corrections. When the pointing model was integrated with the antenna, the required velocity input was tapped from a point downstream in the computation pipeline. This added a small amount of feedback to the command stream through the pointing model. This problem was reported during the analysis of daisy and billiard-ball type scans. Details and plot of the reported problem are available on this page. Of particular note the zoomed view of scan 24.
It should be noted that the problem only existed in the command stream, the indicated position stream does not have this problem (as it uses the correct procedure to generate velocity).
- 3.1 Change the method used to generate the velocity input to the pointing model for the command stream.
- 3.2 Update the FITSVER keyword in the antenna FITS files to denote these changes.
- 3.3 Update antenna FITS documentation with a description of the latest changes.
The changes are illustrated in the diagrams below. The top diagram shows how the velocity input was obtained from a point downstream of the command pipeline, creating a feedback path. The lower diagram shows the corrected method which derives velocity from a point upstream, thereby avoiding any feedback through the pointing model. This change affects both the azimuth and elevation axes.
 |
| Velocity Input for Hysteresis Correction (Incorrect) |
 |
| Velocity Input for Hysteresis Correction (Corrected) |
- 5.1 Will the changes described here change the pointing?
- The changes proposed here should not change the pointing. The reasoning is that the difference between the velocity derived from the topocentric position stream, and that of the encoder position stream is extremely small.
- 5.2 What about the indicated position feedback path, is it not affected also?
- The indicated position pipeline velocity was correctly implemented, and does not have the same problem.
What has to get done to integrate this completely into the system. This checklist must be completed before Cycle Integration Testing begins.
- Communication with Computing group needed? No.
- What documentation needs to be updated? The antenna primary header keyword FITSVER will be updated to a value of 2.13, and the antenna FITS documentation will be updated to reflect this change.
- Training Needed? Is this being released to staff astronomers or everyone right now? General release
- Notification Needed? via release notes
- Verify antenna units tests pass.
- On the simulator, set H00 values larger than normal (>0.001) and H01 values smaller than normal (<25,000). Verify oscillations are not present in sampler commanded velocity fields. (Note: H00 and H01 are fields of the pointing model, which control the magnitude and other characteristics of the hysteresis correction. These fields can be set manually using devexplorer.)
- Verify antenna FITS files are written with a FITSVER of 2.13.
- Run one or more billiard-ball or daisy-pedal scans, and inspect the velocity fields of the Antenna-AntennaManager-azElCommands sampler. Verify there are no fast (about 1-3 Hz) oscillations present.
- Verify that the primary header of the antenna FITS files have a FITSVER value of 2.13.
- Verify that pointing scans execute normally.
- Data stored under Project Id TINT_042808
- Scans 1-4: (xband pointing)
-------------------------------------------------
pazCE1: -0.451 dazCE1: 0.034 tazCE1: -0.417
pazCE2: -0.435 dazCE2: 0.051 tazCE2: -0.384
pel1: 5.009 del1: -0.132 tel1: 4.877
pel2: 5.044 del2: -0.134 tel2: 4.910
-------------------------------------------------
OldAz2: 0.000 OldEl: 0.000
dAz2: 0.043 dEl: -0.133
NewAz2: 0.043 NewEl: -0.133
- Scans 5-8: (xband pointing)
-------------------------------------------------
pazCE1: -0.407 dazCE1: 0.018 tazCE1: -0.389
pazCE2: -0.392 dazCE2: 0.004 tazCE2: -0.388
pel1: 4.879 del1: -0.034 tel1: 4.844
pel2: 4.913 del2: -0.023 tel2: 4.889
-------------------------------------------------
OldAz2: 0.043 OldEl: -0.133
dAz2: 0.011 dEl: -0.029
NewAz2: 0.054 NewEl: -0.162
- Scans 9,10 and 11 are 120, 240 and 240 second Daisy pedal scans.
- sampler2log Antenna AntennaManager azElCommands start halfway through scan 10 and run through scan 11.
- No oscillations noted.
- log filename is 2008_04_29_01:48:10.fits
- Scans 20-23 : (xband pointing)
-------------------------------------------------
pazCE1: -0.412 dazCE1: 0.003 tazCE1: -0.409
pazCE2: -0.398 dazCE2: -0.002 tazCE2: -0.400
pel1: 4.835 del1: 0.090 tel1: 4.925
pel2: 4.869 del2: 0.118 tel2: 4.986
-------------------------------------------------
OldAz2: 0.054 OldEl: -0.162
dAz2: 0.000 dEl: 0.104
NewAz2: 0.054 NewEl: -0.058
- Files checked and confirmed FITSVER is 2.13
- Verify that the primary header of the antenna FITS files have a FITSVER value of 2.13.
- Verify that pointing scans execute normally.
APPROVED: To the best of my knowledge, the request in this MR is complete. I have thought through this request, and believe it to be an important feature to implement or bug to fix.
ACCEPTED: I acknowledge that I have validated the completed code according to the acceptance tests.
Symbols:
- Use
%X% if MR is not complete (will display 
)
- Use
%Y% if MR is complete (will display 
)
CCC Discussion Area
-- JoeBrandt - 05 Mar 2008
Revision r1.8 - 30 Apr 2008 - 18:17 GMT - ToddHunter
Parents: PlanOfRecordC22008
|
Content copyright © 1999-2007 by the contributing authors.
All material on this collaboration platform is the property of the contributing authors.
|
| |
velgenbefore.png 
