Theodolite survey of dome from ao9 monument before and after shimming

Feb. 2003

links to plots:

04feb03: compare 04feb03 survey (before shimming) with aug01 survey.
04feb03: ao9 survey results.
12feb03: ao9 survey results.
17feb03: ao9 survey results.
17feb03: ao9 survey 1,3az,za fits to theodolite data
17feb03: comparing the 3 surveys 17,12,04feb03
17feb03: calibrating the tilt sensors to 17feb03 ao9 survey.
18feb03: creating a pitch,roll model from the 18feb03 tilt sensor swings.
19feb03: creating a pitch,roll model from the 19feb03 tilt sensor swings.
20feb03: creating a pitch,roll model from the 20feb03 tilt sensor swings.
25feb03: creating a pitch,roll model from the 25feb03 tilt sensor swings.
28feb03: creating a pitch,roll model from the 28feb03 tilt sensor swings.

links to sections:

Intro
04feb03: ao9 survey before shimming.
12feb03: ao9 survey after first shimming.
17feb03: ao9 survey after 'most' shimming.
Comparing survey aug01 with 04feb03 survey before shimming.
Comparing the feb03 ao9 surveys: 17feb03,12feb03,04feb03.
Calibrating the tilt sensors using the 17feb03 ao9 data.
Using the 18feb03 tilt sensor swings to generate a pitch,roll model
Using the 19feb03 tilt sensor swings to generate a pitch,roll model
Using the 20feb03 tilt sensor swings to generate a pitch,roll model
Using the 25feb03 tilt sensor swings to generate a pitch,roll model
Using the 28feb03 tilt sensor swings to generate a pitch,roll model (after dome move)


Intro.

        The elevation rails were shimmed and the dome raised in feb03. The relevant dates are:     The changes to the platform took longer than expected. There was no final ao9 survey with the theodolite to measure the pitch,roll, focus errors after all of the changes (the survey was finally done in jul03). The models made were relative to the 17feb03 theodolite survey. There were large changes after this..


04feb03 ao9 survey before shimming.   (top)

    A survey was done from ao9 on 4feb03 before any shimming was done. 2 zastrips at azimuths of 242.87 and 302.87 degrees were.  The the plots show the results of the survey for the two za strips.


ao9 survey 12 feb03 after first shimming.  (top)

    The first round of shimming was completed and then a survey from ao9 was done on 12feb03. 2 zastrips at azimuths of 242.87 and 302.87 degrees were done followed by an azimuth swing at za of 10 degrees. The the plots show the results of the survey for the two za strips.


ao9 survey 17 feb03 after 'most' shimming.  (top)

    Shimming was completed (or so we thought) on 17feb03.  A survey from ao9 was done that evening. 2 zastrips at azimuths of 242.87 and 302.87 degrees were done followed by an azimuth swing at za of 10 degrees. The the plots show the results of the survey for the two za strips and the az swing.     The second plot shows az,za fits to the theodolite data.
    processing:survey/030217/reduc/pltfit.pro


Comparing survey aug01 with feb03 survey before shimming.  (top)

    The plots show the difference in the pitch, roll , and focus between the aug01 survey and the feb03 survey before the shimming.


Comparing the feb03 ao9 surveys 17feb03,12feb03,04feb03.  (top)

    The plots shows the results of the ao9 surveys for 17feb03, 12feb03, and 4feb03 .
The rails were shimmed above 10 deg za and on the carriage house side. The pitch and roll data for za 6 to 10 degrees have not changed for the 3 surveys. The pitch and roll data below six degrees changed for the 3 surveys. The ch side was shimmed so this should affect out to maybe 3 degrees za (38feet is extent of gr trolleys).  What is not believable is the roll change below 6 degrees za 04feb to 12feb and then the return to the 04feb values on 17feb03. I don't believe any shimming is that accurate (even if they tried it). The difference must been in the computations of the pitch, roll. It is not a movement of the platform since the two za strips 242 and 302 were separated by 2 hours yet they do the same thing.
processing:survey/030217/reduc/comparesurvey.pro


Calibrating the tilt sensors using the 17feb03 ao9 data.  (top)

    The tilt sensor data taken during the 17feb03 AO9 survey was used to calibrate the tilt sensors.
The plot shows the theodolite data and the fits.
processing:survey/030217/reduc/cmptsoffset.pro

A description of the method can be found under tilt sensor calibration using the theodolite.


Using the 18feb03 tilt sensor swings to generate a pitch,roll model  (top)

    The tilt sensor data taken on 18feb03 was corrected to the theodolite reference and then a 2d model was fit to this data.
The plot shows how well the model works.
  1. Fig 1. Shows the pitch  of the tilt sensors, the tilt sensor model, and the theodolite measured points.
  2. Fig 2. has the roll of the tilt sensors, the tilt sensor model, and the theodolite measured points.
  3. Fig 3. plots the residuals of the model - the tilt sensor data.
  4. Fig 4 has the residuals of the Theodolite - model for the theodolite points.
The 2d fit worked pretty well for pitch and roll. It fit the tilt sensor data and was able to recreate the theodolite data. The next day (19feb03) had trouble fitting the roll.
processing:survey/030217/reduc/do2dfit_18feb03.pro

A description of the method can be found under tilt sensor calibration using the theodolite.


Using the 19feb03 tilt sensor swings to generate a pitch,roll model  (top)

    The tilt sensor data taken on 19feb03 was corrected to the theodolite reference and then a 2d model was fit to this data.
The plot shows how well the model works.
  1. Fig 1. Shows the pitch  of the tilt sensors, the tilt sensor model, and the theodolite measured points.
  2. Fig 2. has the roll of the tilt sensors, the tilt sensor model, and the theodolite measured points.
  3. Fig 3. plots the residuals of the model - the tilt sensor data.
  4. Fig 4 has the residuals of the Theodolite - model for the theodolite points.
The pitch is fitting pretty well. The roll does not fit very good. The 2dmodel fits the tiltsensor data, but the tilt sensor data does not agree with the theodolite data. There are two separate  tracks over each za strip  which differ by about 3 hours. They track each other very well. The shape difference is hard to reconcile with the calibration of tilt sensor to theodolite (they only use linear terms in za).
processing:survey/030217/reduc/do2dfit_19feb03.pro

A description of the method can be found under tilt sensor calibration using the theodolite


Using the 20feb03 tilt sensor swings to generate a pitch,roll model  (top)

    The tilt sensor data taken on 20feb03 was corrected to the theodolite reference and then a 2d model was fit to this data.
The plot shows how well the model works.
  1. Fig 1. Shows the pitch  of the tilt sensors, the tilt sensor model, and the theodolite measured points.
  2. Fig 2. has the roll of the tilt sensors, the tilt sensor model, and the theodolite measured points.
  3. Fig 3. plots the residuals of the model - the tilt sensor data.
  4. Fig 4 has the residuals of the Theodolite - model for the theodolite points.
The pitch is fitting pretty well. The roll does not fit very good. The 2dmodel fits the tiltsensor data, but the tilt sensor data does not agree with the theodolite data. This looks the same as the 19feb03 data.
processing:survey/030217/reduc/do2dfit_20feb03.pro

A description of the method can be found under tilt sensor calibration using the theodolite


Using the 25feb03 tilt sensor swings to generate a pitch,roll model  (top)

    The tilt sensor data taken on 25feb03 was corrected to the theodolite reference and then a 2d model was fit to this data.
The plot shows how well the model works.
  1. Fig 1. Shows the pitch  of the tilt sensors, the tilt sensor model, and the theodolite measured points.
  2. Fig 2. has the roll of the tilt sensors, the tilt sensor model, and the theodolite measured points.
  3. Fig 3. plots the residuals of the model - the tilt sensor data.
  4. Fig 4 has the residuals of the Theodolite - model for the theodolite points.
The pitch is fitting pretty well. The roll does not fit very good. The 2dmodel fits the tiltsensor data, but the tilt sensor data does not agree with the theodolite data. This looks the same as the 19feb03 data.
processing:survey/030217/reduc/do2dfit_25feb03.pro

A description of the method can be found under tilt sensor calibration using the theodolite


Using the 28feb03 tilt sensor swings to generate a pitch,roll model  (after dome move)(top)

    The tilt sensor data taken on 28feb03 was corrected to the theodolite reference and then a 2d model was fit to this data. This data was taken after all of the shimming and after the dome move.
The plot shows how well the model works.
  1. Fig 1. Shows the pitch  of the tilt sensors, the tilt sensor model, and the theodolite measured points.
  2. Fig 2. has the roll of the tilt sensors, the tilt sensor model, and the theodolite measured points.
  3. Fig 3. plots the residuals of the model - the tilt sensor data.
  4. Fig 4 has the residuals of the Theodolite - model for the theodolite points. This shows how much we moved the pitch and roll from the 17feb03 survey.
We need to do another theodolite survey to get the focus correct (after moving the dome). At that time we should be sure and do at least 2 az swings (along with the za strips) to nail down the az dependence.
processing:survey/030217/reduc/do2dfit_28feb03.pro

A description of the method can be found under tilt sensor calibration using the theodolite
 
 

home_~phil