NRAO Home  >  Green Bank  |  Wiki Topic:    GB > PTCS > PointAndFocusModeling

Changes | Index | Search | Statistics | Go

Models and Performance

• Pointing and Focus Model Recommissioning Results
• See The GBT Pointing Model (URSI 2007) for an end to end analysis of calibrate-offset-track performance. Errata: The error budget for offset pointing erroneously includes only a single contribution for encoder fine-cycle and large-angle errors. They should both be doubled (radial error rather than single-axis errors). There's no substantial change to the result- Total offset pointing errors predicted increase by ~0".1 since these errors are small. Note that the model presented includes an azimuth correction term associated with local focus correction (this might not have the same definition as others use...), but PFM5 will not include such a term. Exclusion of the term increases p68 for cross-elevation error to ~3".0 vice 2".3 when it is used.
• Fitting focus models shows HUGE residual (offset) associated with TPTCSPNT_070319. When it is excluded, the p68 focus performance in benign conditions is ~ 1.3 mm. What was different with this dataset? (RMP - from the commissioning logs, it is very likely there was ice in the dish) NB: Fitting was done with respect to the ant mgr indicated SR_YP. Is it possible that this is not really the position, but some offset position? The receiver for all datasets was X (and note that different receivers will potentially have different focus offsets- how to handle this consistently?).
• Focus model thermal terms suggest that primary reflector distortions (temperature gradients) are the largest contributors to thermal focus shifts.
• Instructions for using the beta version of PFM5 Model
• Inverting the model: The pointing model generated determines the refraction-corrected topocentric cross-elevation and elevation errors as a function of mount azimuth and elevation. In order to calculate a command stream (to main drive servos) the inverse of this is needed, i.e., mount az and el as a function of refraction corrected az (or cross-elevation) and elevation. The pointing model is a contractive map, so the inverse is calculated via an interation... More detail on all of this soon.

Adjunct Information for Observations

• Temperature stability monitor using time derivatives of temperature sensors (or model temperature features) to guide point-ups.
• Angular separation to guide point-ups.
• Results of bootstraps on peak (and focus) fits to flag bad point-ups, e.g., the 1- of peak position.

Architecture of Pointing and Focus Correction System

1. Pointing and focus model database Ver. 5 notes.

Astronomical Data and Data Processing

A note on terminology

If the measurement has conventional or derived power units (i.e., Janskys, W/sr, W/m^2, etc.), and there is no eye response weighting function, it's definitely radiometry. Take a look at e.g. Boyd, "Radiometry and the Detection of Optical Radiation".

1. Accuracy and Anomolies in Pointing and Focus Scans

• Disagreement between Monte-Carlo, bootstrap, and LCP/RCP cal

2. Methods

• "Jack-Scan" method may be suboptimal for high-frequency calibrate-before-offset. Altenatives should be considered.
• FocusFitting
• Current flavors of peak fitting for az and el scans fit a gaussian to the peak. Why do this? Why isn't it sufficient to fit a quadratic in the neighborhood of the peak? Lots of compute time savings to be had, making more extensive bootstrapping possible and quicker turn-around of fits to largish datasets.
• Now are tagging a peak fit with the actual mount az/el at the instant of the peak (as determined by the fit). Some difference between what PrePoint uses and this technique, although not clear what causes the difference.

3. Recommissioning the pointing and focus model

Subreflector

1. Calibration of subreflector link length encoders
2. Inclusion of SR XYZ in SID of TPTCSPNT_070410, scan 29, reduces simulation error from 0".95 to 0".76. This suggests that either the SR motions are contributing to antenna vibrations, or SR motions are directly causing smallish pointing errors. Follow-up by distinguishing using filtering.
3. Several things point to SR motions or calibrations being responsible for dominant pointing and focus errors: a) Improvment of azimuth pointing when LFC is included. This suggests that the LFC motion has a coupled xel component. Could this be due to SR tilt and/or programmed xel motions? b) Temperature feature time derivatives do not correlate with focus residuals, strongly suggesting that the focus errors are not due to temperature gradients. c) Improvement of SID (above, Item 2) when LFC is included

Pointing Model Generation

1. Error Budget
2. Half power track and elevation bearing instruments
3. Effect of azimuth truck wheel out-of-round
4. Hysteresis modeled as encoder running friction
5. Encoder accuracy
6. 0.27 and 0.54 Hz tracking error
7. No apparent correlation between time derivs of struct temps and model residuals except possibly in the case of 060817. See figure attached.
8. Weighted Least Squares
9. No apparent encoder coupling misalignment- Inclusion of terms in 060817 to 070410 all-sky data model fit does not reduce model residual.

Instrumentation Specific to Pointing and Focus Corrections

• InformationForProjectTeam
• Instrument tests from 3/18/07
• WheelPositionSensors
• Differences between Antenna Manager mount angles and Az El 50Hz mount angles

Issues

Presentations and Notes

• Astronomical Data Analysis, DTD 4/16/2007
• Inclinometer results from 070420
• PFM5Alternatives.pdf
• Instrument Tests
• Hysteresis modeled as encoder running friction

Relevant PTCS Project and System Notes

• Thermally-Neutral Traditional Pointing Models and Thermal Corrections to Pointing and Focus
• The GBT Pointing Model
• Pointing Performance and Analysis
• Structure modal frequencies and elevation angle dependence

Other interesting (essential) stuff

• Influence Coefficients
• Bootstrapping and Jackknifes
• Virtual Work

-- KimConstantikes - 24 May 2007

Attachment:	Action:	Size:	Date:	Who:	Comment:
AstroDataAnalysis.pdf	action	1492867	31 May 2007 - 14:34	Main.Main.KimConstantikes
IncTests.pdf	action	130434	01 Jun 2007 - 15:23	Main.Main.KimConstantikes
PFM5Alternatives.pdf	action	187318	09 Jun 2007 - 14:15	Main.Main.KimConstantikes
HysteresisStudy4_18_07.pdf	action	342636	10 Jul 2007 - 19:57	Main.Main.KimConstantikes
results_3_18_07.pdf	action	1174519	12 Jul 2007 - 13:55	Main.Main.KimConstantikes
tempDerivs.JPG	action	91773	14 Jul 2007 - 19:23	Main.Main.KimConstantikes
FocusDatasetProblem070319.JPG	action	53853	04 Aug 2007 - 16:34	Main.Main.KimConstantikes