Current APO Collimation

J. Morgan, Oct. 2, 1997

This report discusses the current method by which the APO 3.5-m telescope can be collimated. This method does not depend on making any modifications to the telescope. Although it would greatly benefit from the addition of a few modifications to the test equipment that we currently have at APO, this procedure can be accomplished with the equipment currently on hand. I offer this description as a help in our future collimations of the telescope as well as for a starting point for discussions about what changes we might wish to make to the telescope to make things easier or better in the future. For descriptions of the APO optics and requirements for the collimation of the 3.5-m, please see my earlier report on this subject.

The current procedure

Aligning the secondary and tertiary mirrors

1. ) Center the primary on the Cassegrain hole lifting fixture.

    

2. ) With the tertiary removed, mount the secondary on the telescope.

    

3. ) Mount the Auto-Reflecting Telescope (ART) in the Cassegrain hole. Note that shims are required to get the ART to reposition accurately. One shim should be placed directly opposite the set screws which hold the ART in the fixture and the other placed 90 degrees to the first. Take care to place the shims in so that they extend the whole length of the mounting tube. I also suggest that we adopt a standard orientation to the ART: with the focusing knob pointing toward the NA2 port of the telescope and the snout of the ART even with the top of the fixture. After placement of the ART, the set screws need to be securely tightened. It is also recommended that a safety wire be strung up around the focusing knob to insure that the ART will not fall to the floor if it slips from the fixture.

    

4. ) Use the reticle lamp on the ART to illuminate the bottom of secondary. With this lamp, the dimple on the secondary can clearly be seen as a dark spot in the field of view even with other dome lights on. Likewise, the back-reflection from the ART can also be plainly seen. The dimple on the secondary has been placed at the vertex of this optical element.

    

5. ) Focus on the direct image of the secondary dimple and then use the telescope vane turnbuckles to translate the secondary until it is centered on the ART crosshairs. It is convenient here, but not mandatory, to mount the Panasonic CCD video camera on the ART to view the dimple as the vane turnbuckles are adjusted. It is NOT necessary at this time to worry about the back-reflection from the ART. At this time you are NOT trying to set the tilt of the secondary, you are only trying to set its translation to be coincident with the axis of the Cassegrain hole lifting fixture. Note that the tertiary mount has been pinned to have its rotational axis to be coincident with the axis of the Cassegrain hole lifting fixture.

    

6. ) Replace the tertiary. Note that this must be done by first removing the secondary, using the crane to install the tertiary, and then replacing the secondary. We are depending here on the pinning of the secondary truss structure to accurately reproduce the secondary position.

    

7. ) Move the ART to the center of the Nasmyth port. There is a special mounting plate which was built for this purpose. Eventually, you will discover that this plate does not hold the ART at the exact rotational center of the NA2 port, but it is close enough and we may ignore this problem for the moment.

    

8. ) Place the tertiary actuators in the middle of their range of motion. Note that I include this step only for future reference. It is not really practical to do this at the current time. The tertiary is held to its mount by three 40 tpi lead screws which in turn are attached to harmonic drives which have 60:1 reductions. The harmonic drives are attached to stepper motors which are currently unpowered. As a result, these actuators must be rotated by hand. With this arrangement, one revolution of the lead screws moves the mirror 0.025", but one rotation of the stepper motor shaft moves the mirror 0.00042". The actuators have a range of approximately 1" which means you must rotate the stepper motors 1200 turns to move the lead screws through 0.5"! Even if you have the fortitude to spin all three stepper motors this much by hand, good luck keeping count!

    

9. ) By eye, find the rough position of the tertiary by focusing the ART on the direct image of the secondary dimple. To align the dimple on the ART crosshairs in the horizontal direction rotate the tertiary mount. To align the dimple on the ART crosshairs in the vertical direction rotate the tertiary in its mount. The tertiary actuators are arranged at the vertices of an equilateral triangle on the back of the mirror. One leg of this triangle is parallel to the horizon. The easiest way to rotate the mirror to place the dimple in the correct vertical position on the ART is to move only the top actuator. Alternatively, you can move both of the bottom actuators in equal amounts.

    

10. ) With the Panasonic CCD video camera on the ART, mark on the TV screen the initial location of the secondary dimple. Also mark the location of the ART crosshairs. Rotate the instrument rotator 90, 180, and 270 degrees, marking the location of the dimple at each rotation angle. The pattern of dimple locations will be a circle on the screen. Below we will refer to this circle as the alignment "error circle". If the camera is exactly on the rotator axis, then the error circle will be centered on the ART crosshairs. If the camera is off the rotator axis, then the error circle will be displaced from the ART crosshairs. In either case, the center of the error circle is the location of the rotator's axis of rotation. Measure on the TV screen the diameter of the error circle. If you have measured the dimple location at the four rotator angles above, then you will have two estimates of the error circle diameter. Take the average of the two measurements as the estimate of the error circle diameter.

     

11. ) Move the image of the secondary dimple to the center of the error circle by use of the tertiary rotator and the tertiary actuators as was done in step 9. Remeasure the diameter of the error circle by repeating the measurements of the secondary dimple positions at 0, 90, 180, and 270 degrees. This step might have to be repeated in order to ensure that a minimum error circle has been achieved. Error circles as small as 0.05" on the TV screen are possible without excessive effort. This is approximately the diameter of the image of the secondary dimple itself. If this is not the case, then don't panic yet. This could be the result of incorrect tertiary pistoning which we will deal with in step 17.

     

12. ) Change the focus on the ART to view the back-reflection from the secondary. The back-reflection is an image of the front of the ART which is seen reflected back through the tertiary and secondary mirrors. Its image on the TV screen is much larger than the image of the secondary dimple, approximately 0.5" in diameter. In the back-reflection image there is a pattern of rectangles that form a cross-hair pattern. With the current sensitivity limits of the Panasonic CCD it is fairly difficult to see this pattern on the TV screen. At its best focus it is just possible to make it out. If this pattern can be seen, use it to estimate the center of the back-reflection image. Otherwise, you're on your own to make your best eyeball estimate of the back-reflection image position. The secondary mirror mount is designed to rotate the secondary around its center of mass. This is close enough to the mirror vertex that for small angles of tilt, the direct image of the secondary dimple is not moved appreciably. Unlike the direct image of the secondary, the back-reflection of the ART telescope is sensitive to the both the translation and the tilt of the secondary. This means that the error circle for the back-reflection will be as big or bigger than the error circle for the direct dimple image. The goal for this next two steps is to minimize the back-reflection error circle without increasing the error circle of the direct dimple image.

     

13. ) In a manner similar to that used for the direct dimple error circle, measure the diameter of the back-reflection error circle and determine its center. Note that in principal the back-reflection and direct dimple error circles are concentric. Therefore, knowing the center of one should be sufficient, but it doesn't hurt to have another measure of the instrument rotator's axes position.

     

14. ) Using a host session to the secondary controller, tilt the secondary until the back-reflection is centered on the back-reflection error circle just measured.

     

15. ) After the tilting of the secondary, remeasure the size of the back-reflection error circle. If the mirror was tilted properly, then this error circle should be significantly smaller than before.

     

16. ) Return the ART focus to the direct dimple image and remeasure the diameter of the direct dimple error circle. If the secondary tilts were large, then its likely that this error circle has increased a little. If this is the case, then you should reiterate step 11. When the back-reflection error circle has about the same diameter as the direct dimple error circle, you have done everything you can hope to accomplish with tilts of the secondary and tertiary mirrors. It is rather important here to record the secondary tilts and tertiary actuator positions.

     

17. ) At this point what we have accomplished to to make sure that the optical beam from the tertiary towards the ART is parallel to the instrument rotator axis and in a vertical plane that goes through the rotator axis (see Figure 1). The optical axis can still be above or below the rotator axis by the size of the radius of the residual error circle at the end of step 16. This alignment error is corrected by pistoning the tertiary to the correct position. The tertiary is pistoned in or out by moving all of the actuators by an equal amount. Recall from the discussion in step 8 that it takes a large number of turns to piston the tertiary any significant distance. In fact, since the tertiary pistons at an angle of 45 degrees to the optical axis, this problem is worse than what we have already stated by a factor of cos(45). For instance, one revolution of the tertiary actuators moves the mirror up or down by 0.00042 * cos(45) = 0.0003". Therefore, to move the mirror up 0.05" takes 168 turns of all three stepper motors. Note here that we currently have not measured the TV image scale when the ART is mounted to the Nasmyth port, but from estimates of the size of the dimple it appears to be close to 1:1 imaging. Therefore, if at the end of step 16 you have a residual error circle with a radius of 0.05" on the TV screen, then you will need to piston the tertiary by approximately 168 turns to decrease this radius and properly align the optical axis. Remember that pistoning the mirror out moves the image up and pistoning it in moves the image down. As before, you want to move the mirror so as to move the dimple towards the center of its error circle. As before, you will probably want to reiterate this step after moving and re-measuring the error circle.

Alignment of the primary mirror

Overview

At the end of step 17 the secondary and tertiary mirrors have been aligned with each other. The translation of the secondary is approximately lined up with the primary, assuming that the optical axis of the primary is close to the center of the Cassegrain hole. Because we have now constrained the positions of the other optical elements, we can now use on-sky imaging to define the best tilt and translation of the primary with respect to the secondary.

The primary mirror adjustments will come in two stages. The first stage will align the mirrors well enough to attain good on-axis imaging and should be easily accomplished in a single night of observing or less. This stage requires only that the primary be tilted in its mirror cell. No translations of the primary will be required for this stage to be accomplished. We know this because it has been possible in the past to achieve good on-axis imaging by means of only tilting the secondary once it has been properly translated to the mechanical center of the Cassegrain hole. There is no optical difference between this and tilts of the primary.

We have two options about how to go about the second stage of the alignment. This stage will consist of using off-axis imaging to determine the true optical center of the primary with respect to the current primary position. The imaging required to accomplish this task can be taken within about 1/2 night of observing. The mirror adjustments can then be made at a later time in one of two ways. Either the primary itself can be translated and tilted to new corrected values, or the realignment of the secondary and tertiary can be redone with the the secondary in step 5 translated to the calculated position of the optical center of the primary. Which of these two options will be easier to accomplish will depend on which direction we believe the primary needs to be translated to. If the translational motion is all or primarily in the direction of the transverse axis, then it will be easier to take the first option. If we are required to move the primary in the direction of the Nasmyth axis, then it will probably be easier to take the second option.

Stage 1: Obtaining good on-axis images

The adjustments for the first stage are relatively easy. The primary mirror can be tilted by adjustments of the hard-point positions on the Primary Mirror Support System. The hard-points positions are adjusted by screwing them in or out. The adjustment screw pitches are 20 tpi. The hard points are located 42.45" away from the center of the Cassegrain hole in an equilateral triangle pattern. Each motion of a hard-point tilts the mirror about an axis that goes through the opposing two hard-points. The distance between a hard-point and its axis of revolution on the mirror is 63.675" (see Figure 2). One revolution of the hard-point screws is a 0.05" motion of the mirror, which is equivalent to a tilt of 0.05/63.675 = 0.00079 radians = 162 arcseconds. We ought to be able to adjust the position of these screws to about 1/10 of a turn, which is to within a 16 arcsecond precision. The procedure for these adjustments is given below.

1. ) Use the guider to place a star image in the middle of the Nasmyth port. Choose a star which has an elevation angle of 40-60 degrees. A fairly bright star will be needed because the Panasonic CCD sensitivity is poor.

    

2. ) Mount the ART on the Nasmyth port with the CCD camera attached. Move the telescope out of focus so that the primary doughnut may be seen on the TV screen. Move the TV monitor to the bottom of the telescope

    

3. ) Record the initial primary tilt position by writing down the LVDT readings for the A, B, and C sectors.

    

4. ) Adjust the hard-point screws so that the primary doughnut is symmetric. The secondary obscuration should be centered in the doughnut pattern.

    

5. ) Record the new LVDT readings. At this point the telescope should have good on-axis imaging. For future reference, also make sure that you record the position of the primary with respect to the Cassegrain hole, the tilt values of the secondary, and the positions of the tertiary actuators. Make sure that the bracket which locks down the tertiary rotator mount is lock down tightly. At this point, the secondary has been placed on the "neutral point" arc (see the 3.5-m collimation report, Figure 1), where translation misalignments cancel tilt misalignments. If the optical axis of the primary lies at the center of the Cassegrain hole, then we are done and no further adjustments will be needed. The second stage of the primary alignment procedure is designed to determine where the optical axis of the primary lies relative to the axis of the Cassegrain hole.

Stage 2: Obtaining good off-axis images

We first need to gather the information which will enable us to compute thetrue position and tilt of the primary's optical axis relative to the current secondary position. We will require off-axis images from the guider to determine where the primary's optical axis lies. Software will be developed for the purpose of analyzing the off-axis guider images and deducing from them where lies the optical axis of the primary. The basis of this software is described in the 3.5-m collimation report in the section titled "Zernike coefficient fits to off-axis images". I will not repeat this information in this report. Here I will only be concerned with describing how to go about acquiring the data needed for this analysis. This data can be obtained any time after the completion of stage 1.

1.  ) Acquire a bright star on the guider.
2. ) Record the initial rotator position angle and take a series of guider images as the telescope is moved through focus. It is desirable to have images on both sides of best focus. We need to be able to define the position of best focus as well as to measure the major-to-minor-axis ratio on each of the out of focus images. For this second measurement, a fairly large focus range is desirable. I estimate that five images centered on best focus covering a range of approximately 500 microns of focus travel will be sufficient. It is very important to record the focus position of each image. Record the end rotator position angle after this sequence of images has been taken. Store the series of guider images taken for future analysis.
3. ) Move the guider to a rotator position angle 90, 180, and 270 degrees from the initial rotator position angle in step 2 and take a through-focus sequence at each rotator position angle.
4. ) Let J. Morgan know where all of this data can be found! For the initial analysis and development of the software, I will be responsible for the analysis of the off-axis images. As I acquire facility with this techniqueI hope to train others to do this analysis.

The analysis of the off-axis images should give us the translation and tilt of the primary axis relative to the Cassegrain hole axis. With this information in hand, we will then be in a position to decide how best to align the primary mirror. The procedure we choose from here will depend primarily on where we need to move the primary. If most of the motion is in the transverse direction, then it will be relatively easy simply to move the primary using the transverse hard-point and then to live with a small misalignment in the orthogonal direction (the direction of the Nasmyth axis). If we are required to move the primary a significant amount along the Nasmyth axis (ie. more than 2 mm) then it will probably be easiest to redo the alignment of the secondary and tertiary instead of actually trying to move the primary in this direction. In this case, we will need to repeat steps 2-17 given above with one difference: we will align the secondary to the new offset position of the primary axis rather thanto the middle of the Cassegrain hole as we did in step 5.