Understanding and Calibrating the Argus Positioners

This document is meant for the technical staff. It is provided to the users in case they want more information on the general (xy) calibration of Argus.

1. How The Argus Positioners Work

Argus has 24 uniformly spaced positioners. They turn on pivot points nominally 176 mm from the center, 15 degrees apart. The nominal field is 160 mm in diameter. Each positioner has a radial linear actuator which moves in 10 micron steps from the edge of the field to slightly past the center. Thus, moving a positioner radially 100 steps moves it one mm toward or away from its pivot point.

The positioners are rotated about their pivot point by similar linear actuator arms, moving in 10 micron steps. The line of force is nominally 135 mm behind the respective pivot point. Thus, the physical distance a fiber moves in a step is approximately d = 10*R/135, where R is the distance the fiber is from the pivot point at that instant.

In order to position Argus properly, a very accurate coordinate transformation must be made between astronomical space (R.A. and Dec.) and positioner space (radial and angular steps). To simplify this transformation, it is done as two completely independent processes. Astronomical and mechanical positions are transformed to and from X Y space in mm in the focal plane, i.e.

Transformation 1: Astronomical (arcsec on the sky) <--> X,Y (mm)

Transformation 2: Mechanical (radial and angular pulses) <--> X,Y (mm)

All of the information which Argus uses to do this is contained within the file argus.par, which is presented in an Appendix, for reference. When using Argus, it is important to know that argus.par is read only when the Argus program is started, so to change these values permanently, the program must be exited, new values entered in argus.par and the Argus control program re-entered to load the new values.

2. The Mechanical Model

In principle, the transformation from pulses to position is a matter of simple geometry. In fact, the positioners are not perfect. To compensate for their mechanical imperfactions, each positioner has a model for which there are seven independent variables. The meanings of these variables are mechanically intuitive, but the user has no pressing need to know just what they are, so we won't bother with explanations here. See the Appendix.

Whenever a positioner is mechanically adjusted, the mechanical model MUST be checked. Any adjustment of the arm is likely to change the parameters and affect the positioning. This is done with a special reference reticle. The reticle is made with a very precise xy grid of lines etched on it at 5 mm intervals, with a precision of one micron.

Essentally, an Argus positioner is calibrated by installing the reticle and sending the positioner until it is over one of the intersections, noting the number of steps which were required to get there. This process is repeated a number of times. These data are then processed by a computer program and new coefficients for the mechanical model are calculated. The coefficients are then entered manually in the argus.par file.

The calibration process is time consuming if it has to be repeated for all 24 positioners, but it is not difficult.

3. How to Do the Mechanical Calibration of Argus

To calibrate an Argus positioner, one must perform the following steps:

1. Install the reference reticle. This is done with Argus on its stand. The periscope and fibers are all withdrawn to the home position to get them out of the way. The reticle is installed in the Argus focal plane from underneath. The reticle is mounted on an aluminum plate with captive mounting screws and locating pins. Installation is straightforward and obvious.

2. Next, a special viewing device must be installed to permit viewing the lines on the reticle on a TV screen. The viewing system is installed in place of the periscope head. The viewer consists of the WATEC (ice cube) TV camera and a transfer lens which mounts on top of it. Once installed, it can roam around the field using the periscope motion control system and position itself under all the parts of the reticle grid.

3. Once the reticle and the WATEC have been installed, focus the WATEC on the grid of lines in the reticle and move it until it is over the center intersection of the reticle (identifable by a small arrow next to the lines). This is the center of the field. Move the positioner until it is directly over this spot and define it this as the center using the command "define center position" . This is done in the normal way, as should be done when using the WATEC to center the positioners at the start of every night.

4. Now, go to the Argus "Command Mode" by pressing F3.

5. Issue the command "cal n" , where n is the positioner number you wish to calibrate.

6. the program will now move the positioner to put it where it thinks it has to go to be over one of the intersection points on the calibration grid and jump into keypad mode. If the model is perfect it will go exactly to the intersection. If there is any error, center the fiber on the intersection and hit ESC. This will record the correct position.

7. The program will now go sequentially to a number of grid intersections. Repeat the above procedure until the program has moved to all positions. At the end, a file named "newN.tst" will automatically be generated ( N corresponds to the positioner number, ie. new5.tst). This is a list of the number of steps which this positioner actually needed to get to all the intersections and is the information which it will use to make a new, more precise model for that positioner. DO NOT try to stop this program in the middle. It expects to go to a certain number of positions and then exit. It is not forgiving of an attempt to escape the sequence.

8. Repeat this procedure for all the positioners you intend to calibrate. Each cycle will generate a file newN.tst.

9. Copy the newN.tst file(s) to the SUN and run the program "tsm". It will ask you which positioner number you want to calibrate, take the information contained in newN.tst and calculate new coefficients for the model. It will display the new coefficients on the screen.

10. Copy down these coefficients and carefully edit them into the "argus.par" file. An example of what this file looks like is shown in the Appendix.

Important!

Whenever you edit the argus.par file, please be sure and add a comment line or lines at the end specifying who last updated this file, when it was last changed and why. See the example argus.par file.


The tsm program currently resides in "/ut00/ricardo/argus". There is also an archival copy in /www/argus.


4. The Astrometric Model

The astrometric model was made by Guo at Harvard by scanning photographic plates and is believed to be very good with a precision of better than .5 sec of arc. It is expressed in terms of the focal length of the telescope F and the third and firth order distortion coeffients D3 and D5. Optically the field is symmetric and the mapping of sky onto the flat focal plane is given by:

arcsec on the sky -> physical distance from the center


R = F*A*(1 + D3*A**3 + D5*A**5)

where A is the angular distance in radians and r the physical distance in mm.

F  = 11452.9 mm
D3 = 356.5 (mm^-2)
D5 = .7e6 (mm^-4)

5. Correcting for Centering and Rotation Errors


T. Ingerson
G. Schumacher
31 March 95


Appendix - Example of an argus.par file


version-5.0
*      fiber_size   krad   kang   krot    d3    d5   focalen
0.10     0.000      0.000       -0.169  356.5  0.7e6 11452.9
*      per_center_x,y   per_calib_x,y   safe_limit   tv_nominal
         4281  3827      4642  2890         50        1885
*      mirror positions [1-5]  
  679   734   854   896   906
*     xt      df      ap    ad   fk  x_cen y_cen
 1  135.50  176.30   0.00   0.50  0   7852  4452
 2  135.60  176.50   0.00   0.20  0   7789  4502
 3  135.60  176.70   0.10   0.55  0   7797  4440
 4  135.10  176.00   0.00   0.65  0   7807  4517
 5  135.70  176.60   0.05   0.60  0   7778  4462
 6  135.10  175.90  -0.05   0.30  0   7813  4574
 7  135.10  176.30   0.00   0.30  0   7712  4589
 8  135.80  176.90  -0.05   0.20  0   7756  4623
 9  135.40  176.40  -0.05  -0.10  0   7729  4460
10  135.60  176.00  -0.15  -0.30  0   7809  4555
11  135.30  175.90  -0.05  -0.20  0   7807  4550
12  135.40  175.80  -0.10   0.00  0   7788  4524
13  135.50  176.50   0.00   0.00  0   7810  4471
14  135.30  175.70  -0.10  -0.50  0   7803  4508
15  135.50  176.50  -0.05  -0.60  0   7761  4510
16  135.20  176.00  -0.05  -0.10  0   7792  4511
17  135.50  176.50  -0.05  -0.50  0   7717  4538
18  135.40  176.20  -0.10  -0.40  0   7817  4520
19  135.20  175.80  -0.05   0.10  0   7778  4488
20  135.20  176.20  -0.10   0.00  0   7786  4572
21  135.20  176.10  -0.05  -0.30  0   7834  4547
22  135.10  175.90   0.00   0.20  0   7838  4570
23  135.20  175.70  -0.05  -0.20  0   7845  4499
24  135.10  175.70   0.05   0.20  0   7748  4530
*
* Last updated by: T. Ingerson  31 March 1995
* Format changes only
*



tingerson@noao.edu
gschumacher@noao.edu