Jeff Morgan's Collimating and Focusing DIS III Instructions

DIS Red Camera Focus Procedure Notes

Jeff Morgan, April 26, 2002

For both focus and the alignment of the dewar rotation it is useful to use the following spectrograph settings:

High Resolution Grating set at a centeral wavelength of 6700 A

0.9" slit

Ne lamp

60 sec exp.

1x1 binning

Before attempting to focus the DIS cameras you should first make sure that the camera rotation is well adjusted. On the dewar flange there are 4 slots which hold #8 bolts. These bolts hold the dewar and camera optics onto the spectrograph. Loosen all but one of these bolts to adjust the dewar rotation. The current position of the dewar rotation should be marked on the dewar by a piece of tape which has been cut at the interface between the dewar and the fixed spectrograph tube. If it is not, then put on such a tape before attempting to rotate the dewar. For the red camera, a clockwise rotation of the dewar will result in a clockwise rotation of the image on the CCD. A relative motion of about 1/32" between the dewar and the spectrograph tube (at the location of the marking tape) will result in a tilt of several pixels. With a modest effort you should be able to line up the spectral lines with the CCD columns to within about a pixel from top to bottom of the spectrum.

Note that the original camera optics produce spectral lines with a very small "bow" to them. Between the center of a spectral line and one of its ends there will be about 2 columns difference. You may use the IRAF routine "implot" to analyze the spectral line widths when you attempt to focus the spectrograph. Note that the implot command ":l 300 350" will average rows 300 through 350. This significantly increases the signal to noise in the plots. Owing to the aforementioned bow, if you average too many rows (say 300), then you will artificially increase the measured widths of the spectral lines. In the focus example give below this was done. The fwhm numbers given in the example below are about 10% larger than the actual fwhm of the spectral lines because of this. Other implot command that you will find useful here are the commands "e", "r", and "p". The "e" command allows you to look at a blow up of the lineplot by defining a region with two "e" keystrokes while the cursor is positioned at the start and end column positions of the region you wish to enlarge. The "r" command restores the plot to its original state before the "e" command was done. The "p" command allows you to fit a gaussian to the spectral line enclosed by two "p" keystrokes. For details on these commands see the IRAF help file on implot.

There are 3 actuators, labled A, B, and C, at the back of each dewar which can be used to piston and tilt the CCD. Each actuator will rotate approximately 10 full turns. Each turn of an actuator moves the detector approximately 0.002". Turning an actuator clockwise will move the CCD toward the camera optics. The only way to know how many turns each actuator is from the end of its travel is to gently turn it until it hits a stop. When moving the actuators to the end of their travel it is much safer to travel to the full "out" position (i.e. move the actuators counter clockwise until they hit a stop.) The inner stop of each actuator is a slightly dangerous position to move to. If you jam the actuator into the inner stop with sufficient speed you can make it impossible to back it back off. There is a clutch which will start to slip at the inner stop if you do this. Normally you will only make the clutch slip a little before you come back off of the stop, but if you are careless and jam the actuator too hard against the inner stop you will have to take the dewar apart to recover the actuator's functionality! This problem can also happen with the outer stop, but you have to really be a gorilla to cause this to happen at the outer stop. The bottom line here is be gentle!

Each actuator has a thumbscew which must be loosened before any attempt is made to move the detector, and a degree scale marked along the body of the actuator. The zero degree marks are near the thumbscrew positions. There are fixed marks on the body of each actuator which you can use as a fiducial when turning the actuators, but usually these fiducials happen to be inconveniently located. For convenience I simply put a small piece of tape at an easy viewing angle and use that for a fiducial instead of the mark provided on the actuator itself. In the focus example below I mark the position of an actuator by referencing it to some arbitrary starting position ("0") and an angle. For example, the initial entry below shows a starting position of "0+135" for actuator A. This means "0" turns + 135 degrees. Positive degrees are in a clockwise motion and therefore represent motions of the CCD toward the camera optics. Motions of 90 degrees are decearnable in the spectral line widths. When you are near best focus 90 degree turns will change the fwhm by about 0.1 pixels.

Actuators B and C are aligned with the wavelength axis of the spectrograph. Moving these actuators in opposite directions will tilt the CCD with respect to the optical axis of the camera. When viewing a spectrum from the Red Camera you can pretend that actuator B is on the red side of the spectrum (i.e. to the left when using IRAF to view the spectrum) and actuator C is on the blue side of the spectrum (i.e. to the right in the image).

When using the 0.9" slit, the best focus you can achieve is about 2.0 pixels fwhm for each spectral line. You will find that there are times that you will see smaller fwhm values than this, but you cannot achieve this uniformily across the detector owing to the current optics. In the example below you will see that we choose to leave the detector at a place that gave us very uniform results across the entire detector. This was done at the expense of a slightly (about 10%) larger fwhm in the blue end of the spectrum than was achieved in some earlier exposures. This compromise was thought to be the best overall considering both signal to noise and ease of data reduction.

I will finish this discussion by simply showing the log of one of our most recent focus measurements. In the table below we measured the fwhm of 8 lines spaced across the chip. In the last section of the table you will see the column positions of the line measured as well as the fwhm of that line. I have discussed the interpretation of the actuator positions above. The positions in the table below are relative to an initial starting position. You can put these positions in terms of the total range of motion in the actuators only by knowing the starting position relative to the actuator limits. If 0 turns represents the full "in" position, then 10 turns is approximately the full "out" position. In these terms the starting position for the table below is approximately at A=2.1 turns, B=3.1 turns, and C=2.6 turns. The red camera focus is therefore fairly close to the inner actuator limits.

../../instrumentinfo/dis/graphics/focus.tutorial1.jpeg