Gray Miller Spectrograph

The Gray/Miller Cassegrain Spectrograph

Introduction

The Dark Sky Observatory Cassegrain Spectrograph, otherwise known as the Gray/Miller (G/M) spectrograph was designed and built in the early 1990's, and saw first service on the DSO 18-inch telescope in 1993. In 1995, the spectrograph was modified by replacing the old, war-surplus AeroEktar camera lens with a modern multi-coated Pentax large-format 200mm lens, and by making it possible to tilt the focal plane. Since August 1995, it has been used on a regular basis on the straight-through Cassegrain port on the Dark Sky Observatory 0.8 meter telescope. To date (November 2003), nearly 22000 spectra have been obtained with this spectrograph.

The G/M spectrograph is a traditionally designed classification spectrograph, consisting of an entrance slit (100 microns width), an optimized doublet (achromat) (350mm f.l.) for the collimator, a reflection grating for the dispersing element, and, as mentioned above, a large-format 200mm Pentax lens for the camera. The detector is the DSO 1024X1024 thinned, back-illuminated, Peltier/glycol-cooled Tektronics CCD.

Five reflection gratings are available for use on the G/M spectrograph, although only two are in common use. The spectral ranges and resolutions, the approximate grating tilts, focal-plane tilts, etc. for these two gratings are summarized in table 1. Since the other three gratings are not in regular use, standard settings have not yet been determined for them.

Grating	Spectral Range	Res	Grating tilt	Camera Focus	F.P. tilt
1200g/mm	3800 - 4600 A	1.8 A	3^o 18'	7.7 (fixed)	12.0 (fixed)
1000g/mm (1st order)	H-alpha region	2.0 A	358^o 48'	7.7 (fixed)	12.0 (fixed)
1000g/mm (2nd order)	3800 - 4250 A	1.0 A	353^o 30'	7.7 (fixed)	12.0 (fixed)
600g/mm	3800 - 5600 A	3.6 A	10^o 18'	7.7 (fixed)	12.0 (fixed)
600g/mm (red blaze)	H-alpha region	3.6 A	7^o 18'	7.7 (fixed)	12.0 (fixed)

Setting Up Procedure

Mounting the Spectrograph

The G/M spectrograph is mounted at the main Cassegrain focus of the 32" telescope. Usually you will find the filter wheel mounted at the Cassegrain focus. To remove the filter wheel, first remove the CCD and temporarily mount it on the west port of the GAM. Screw the dust cover that was on the west port onto the filter wheel. Detach the cable (labeled "3") from the filter wheel by gently unscrewing the plug that attaches it to the filter wheel. This cable, once detached, may be allowed to swing free. Remove the northern bolt (the one that does not have a slot on the mounting flange) attaching the filter wheel to the GAM, and carefully loosen the remaining 3 bolts by about 3 - 4 turns. Grasp the filter wheel firmly in both hands and turn it to free it from the bolts and remove the filter wheel from the GAM. Store the filter wheel in the grey cabinet, with the CCD dust cover & filter wheel mechanism facing up.

To mount the spectrograph, first install one bolt in the northernmost screwhole (using the same circle of bolt holes as the filter wheel). The bolt should be screwed in to the same amount that you found convenient for removing the filter wheel. Now install two other bolts at 90^o to the west and 120^o to the east along the same circle of bolt holes (see figure below). Once you have verified that these bolts are screwed in so that the mounting flange has enough leeway to turn easily, go back to the grey cabinet. Remove the cardboard cover from the top of the spectrograph flange, grasp the spectrograph so that the right hand is behind the spectrograph immediately below the flange and the left hand is supporting the bottom of the spectrograph ( please do not touch the grating adjustment dial or pick the spectrograph up by the camera barrel!!!). Return to the telescope (I find I must do all of the mounting/unmounting procedure from the first or second step of the ladder), and raise the spectrograph flange over the bolts, so that the camera part of the spectrograph is pointing due north, and twist the flange so that the bolts now hold the spectrograph. Use the hexagonal wrench to tighten the bolts in rotation, making certain that the spectrograph is fully rotated clockwise. Before the bolts are entirely tightened, insert the fourth bolt in the southernmost hole and tighten it as well. Tighten all the bolts snugly.

The telescope will now need to be rebalanced for the spectrograph. The weights should be set at 11,500.

Bolt pattern and plug

Mounting the Comparison Lamp

The comparison lamp usually resides on the telescope pier on top of the comparison lamp power supply. Before this lamp is mounted beside the spectrograph, a small penny-sized brass plug must be removed using a screw driver from the base of the GAM (see figure above).

Once this is done the comparison lamp is mounted immediately to the east of the spectrograph using the outer set of bolt holes (see figure below). One cable from the lamp is permanently attached to the power supply - make certain that it is can swing freely. The other cable ends in a two pronged plug - it must be inserted into its mate which will be above your head, hanging from the telescope.

Mounting the CCD

The dust cover on the camera mount on the spectrograph may now be removed. Remove the CCD from the west port of the GAM and backout all the thumbscrews so that they do not stick out from the lower surface of the CCD flange. Position the CCD on the spectrograph camera flange so that the label (CH250) is facing north (see figure below). The CCD shorting plug should be aligned EW. Align and tighten the four thumbscrews which are in the diagonal positions (these are not the set used for the filter wheel). Install the dust cover on the west port.

Now remove the blue shorting plug from the CCD and place it in a safe place. Install the CCD cable to the CCD, and loop it over one of the white plastic loops on the telescope so that it will not flop around. Once this is done, turn the power strip on the west side of the pier on. Check that the temperature on the CCD control box is going down.

CCD mounting

Data Acquisition Computer Setup

The data acqusition computer in the control room is normally left on. Turn on the monitor and click on the icon for PMIS. This will open up a window with a black background. Under "File", click on "New Image", and a square window (where the data will appear) will open up with a white background. Back in the first window, type at the prompt: ROI RECT 426 0 183 1024 . This will draw a rectangular outline in the image window. Go to the image window, and click on the ROI menu. Click on "ROI to Camera". Nothing immediate will seem to happen, but once the first set of data is acquired, the window will shrink to the ROI (region of interest).

Power up the Comparison Lamp

The power strip that powers the CCD also powers the power supply for the comparison lamp. The power supply resides on the telescope pier below the polar axis. Push the red button and turn the knob until the current is at 15 mA. Do not under any circumstance exceed 25mA! You should see a faint violet glow coming through the ventilation holes on the comparison lamp.

Telescope Setup

To set the telescope up for the nights work, follow the standard procedure in the DSO User's Manual. Under the GAM setup, set the slide (motor #1) to position 4. Check to make certain that the LED on the GAM is indeed on position 4. Once the telescope is fully setup, carry out the following steps:

In the spectrograph drawer in the control room (labeled "Gray") you will find an old dark-brown 1 1/8" f.l. eyepiece, plus a brass extension tube. Insert the eyepiece in the brass extension tube, and then slide this tube into the aluminium 1 1/4" to 2" adapter that has an off-axis hole. Tighten the two nylon screws, and then insert the entire assembly into the north port of the GAM so that the two nylon screws are at 45 degree angles and are pointing "down", and tighten the retaining screw on the GAM. The north port (which has a two-inch circular hole which will accept the aluminum adapter) is on a drawer which can be pulled out. Pull the drawer out so that the scale on the western side of the drawer reads 2.0 cm. This drawer is released and locked with a lever under the RHS of the drawer. Make certain to lock this drawer once you have it properly positioned.

TV Guide Camera and Guide Star Setup

Now you are ready to find a set star. Select a set star from the table on the bulletin board in the warm room; the best set star is one that is a little to the east of the meridian. Use the "select library object" command under the Movement menu in TCS to move to this star by typing in the library number for the star, hitting return, answering ``No" and then entering 7 for ``slew". Make certain that ladders, etc. are out of the way before you move the telescope! First center the set star in the finder scope. The finder scope has illuminated crosshairs. Move the star to the single crosshair, and then center it on the western-most of the double crosshairs. Uncover the spectrograph slit by pulling the plunger on the south side of the spectrograph out so that about 13 indentations are showing. Then, look through the North Port GAM eyepiece (the one you inserted a few minutes ago). Center the out-of-focus image of the star (it will be out of focus if the telescope was not used for spectroscopy the previous night; if it was used for spectroscopy, then the star will be in focus. It is best to defocus it at this point) on the slit, and then push the spectrograph cover (decker) back, using the plunger so that the decker covers up the ends of the slit (if the star was originally out of focus, the exposed length of the slit should be roughly the same length as the diameter of the out--of--focus image of the star). When the slit decker is properly positioned, there should be about 8 indentations showing on the plunger.

Once you are satisfied the star is centered properly and the decker is properly adjusted, remove the eyepiece and aluminum adapter from the GAM and bring them into the control room. Remove the eyepiece and brass tube from the adapter. Fetch the intensified TV camera from the cabinet in the dome (it will be in a black suitcase), and bring it into the control room. The video cable is wound up and is hung on a nail on the outside wall of the control room, near to the telescope pier. Plug this video cable into the appropriate plug on the back of the intensified TV camera (it may be necessary to rotate the plug while applying gentle pressure to get it to engage). Then, remove the cover from the intensified TV camera, and insert the front tube of the intensified TV camera into the aluminum adapter, so that the box--like battery holder on the TV camera is centered between the two nylon screws. Tighten the nylon screws, and insert the whole assembly into the north port of the GAM, making certain that you tighten the retaining screw as tightly as possible. The two nylon screws should be aligned as before with the eyepiece. Make certain that the CCD part of the intensified TV camera is aligned parallel with the base of the mirror cell (see figure below). Using the thumb rotary switch on the TVcamera, turn the TV camera on, and rotate the rotary switch (the "gain") about half or one--third of a turn. I usually loop the video cable over the same plastic clip that the CCD cable is on.

Go back into the control room, and turn on the TV monitor and the intensified TV power supply (the power supply is a small metal project box with a push--button on the front. It is found on the top of the TV monitor. Push the button; it used to light up but doesn't any more - don't push it twice! After a few seconds, if everything is okay, the out--of--focus image of the star on the spectrograph slit should appear on the screen (see figure). If necessary, adjust the angle of the intensified TV camera on the telescope so that the slit on the TV screen is perfectly horizontal. At this point, it is also a good idea to use a marking pen to make marks at the ends, center, and one quarter and three quarter points across the spectrograph slit on the TV screen. If there are marks from a previous night on the screen, and they are appropriately place, that is good - use them. If they are not properly placed, remove them with rubbing alcohol (there should be a bottle in the control room) and remark them (see fig below).

Since the set star is a bright star, we usually don't completely focus it, as the in-focus image of the set star on the intensified TV can damage the TV phosphor. Focus this image so that it is about the size of a quarter on the screen. Now center the out-of-focus image as best as possible, initialize the coordinates (using coordinates out of the Bright Star Catalog or the Astronomical Almanac) for the set star (using the Initialization menu in TCS -- "2) Set Telescope Position''), and then find a fainter star (V about 6.0) near to the set star in the Bright Star Catalog, and slew to it (make certain ladders, etc. are out of the way!) You will see the dim, out-of-focus image of this star on the screen. Use the focus buttons (which can be sped up by pressing the middle "set'' button) on the hand paddle until the star is in as good focus as possible. You may need to refocus from time-to-time during the night, as the mirror cools and comes to equilibrium. At this point, you can increase the gain on the TV camera to maximum if you plan to observe faint stars during the night. If you are only going to observe bright stars (< 6 mag), leave the gain where it is.

Observing

Preliminaries

We now use a "macro" to assist in semi-automating exposures during the night. To set this up, carry out the following steps. Using Windows Explorer, make a new folder under drive E: (label Gray) on the data acquisition computer. This new folder should have a name that reflects the current date in the format mmmDDYY, for instance, jan2303, sep1502, etc. The convention is to use the local date, not the UT date. Thus, if the night is Oct 15/16 2003, the name should be oct1503. We use this convention even if we start observing after local midnight. In the main PMIS window (black background) click on the "Macro" menu and select "Edit". Under the macro directory (on the C: drive), enter the "Gray" folder, and then select "Specgrab.cmd". About halfway down this macro you should edit the directory name to that of the folder that you just created. Save the macro and exit from the editor.

On the other Windows computer, click on the icon Fitsheader. This runs a Visual Basic program that grabs info from the CCD camera and TCS and uses that information to stuff the FITS headers in the datafiles.

You are now ready to turn to the first star of the evening. However, it is best at this point to begin exposing a Dark. In the data image window (white background), click on "Acquire" and then on "Dark". Enter 300 seconds and expose.

While the Dark is exposing, enter the basic data for the night into the header of the observing page for the night (blue notebook), which will be in the observing briefcase. Note the sky conditions, the grating used, the date (use the double-date convention, i.e. Feb 10/11 2003), and your name and the name of any co-observers.

Now you can turn to the first star on your list. Slew to the star, and spend a few seconds trying to get the star in the best possible focus using the buttons on the hand paddle. At this point, you can also go into the Rates menu in TCS (selection 3) and set the handpaddle rates. I use 100 for the set rate, and 3 for the guide rate.

By this time the dark will have finished, and the dark will appear in a new window on the computer screen. Click on "File" on this new window and then "Export" and save the dark as a FITS file in the directory for the night. The name of the file should have the format dkXXXX.fit where XXXX is a running number. Check the previous observing night to find out what XXXX should be. Once you have exported the Dark, take two Bias frames by selecting Bias from the Acquire menu. Export these to the same directory as FITS files, using biXXXX.fit as the format for the name. Again, XXXX stands for a running bias number.

On the main data window (white background) click on the Display menu and then click on "show cursor".

Acquisition of Stellar Spectra

Now center the star on the slit, using one of the marks that you placed on the TV screen, and acquire a spectrum by using running the macro that you just edited. This macro will prompt you for the running "PM" number and the name of the object. The standard exposure time built into this macro is 300 seconds. If the star is faint, simply guide the telescope so that the star remains at the same point on the slit. If the star is bright, you will need to trail the star back and forth along the slit. To do this, you can use the Trail option under the Rates menu in TCS. A good trail rate is either 2 seconds or 5 seconds per second, and a good trail length between 10" and 60", depending upon how bright the star is, and where you are in the sky (the trail length is multiplied by cos(delta), so you will need to use long trails near the pole to compensate). It will just take practice to know the right trail length. The trail direction should be 90 degrees.

Once the spectrum has been acquired, the main data window will shrink to the size of the ROI. You may wish to right click on the macro to temporarily get out of the macro to resize the window. I usually go into the Display menu, click on Zoom, and then click once on the spectrum and then resize the spectrum window to nearly the full length of the screen. Resize the window horizontally so that the whole ROI width is visible. I usually also change the scaling from "Optimal Scaling" to "Fixed Min/Max", setting about 1000 for the Min and then, maybe, 4000 for the Max. You may need to change these settings for different stars. Alternately, you can use "Dynamic Min/Max".

Now, check to make certain that the star was not overexposed. Run the cursor up and down the spectrum. If no counts are above about 60000, the exposure was okay. For faint stars, you will want to use the "Fixed Min/Max'' option, and enter appropriate values to scale the image so that the spectrum can be seen. You will need to experiment with the max value; for very faint stars, there may be only 100 or 200 counts in the spectrum, hence a max value of 1200 may be appropriate. Use your cursor to find the approximate maximum number of counts in the spectrum, and then set Max to about 20% above that value. Unfortunately, these menu manipulations cannot be carried out while the macro is running. However, counts can be read out once the exposure is finished, even though the macro is running.

The macro automatically saves the exposure under a name with the format PMXXXXX.fit where XXXXX is a running number. The macro will automatically increment XXXXX if multiple exposures are taken.

It is good practice to take a number of exposures of the same star; these exposures can be combined during the reduction process to increase the signal to noise ratio. I almost always take at least three spectra of any star, even bright stars. For fainter stars, four is the minimum number, ranging up to twelve, all of the same 300 seconds. With stars that are not trailed, I usually move the star to a different mark on the TV screen between exposures so that the spectrum is not exposed on the same part of the CCD every time. This helps to increase S/N as well.

To end the macro once you are finished with a star, right click on the macro after "starting" the exposure (the macro gives you two seconds to do this!).

You should aim for a minimum of about 10000 total counts in the spectrum at the violet end - usually in the vicinity of the Ca II K-line (this will lead to S/N = 100, or 1% statistics). Measure the counts near to Ca II K in your first exposure of the star. Subtract about 1000 (the average number of counts in a dark) from these counts and multiply by three (with good seeing an untrailed spectrum will be about 3 pixels wide -- if broader, use that number of pixels in the multiplication). Divide this number into 10000. This is the minimum number of exposures you need for that star. Adhere to the minimum numbers mentioned above (i.e. 3 for a bright, trailed star, 4 for a fainter untrailed star). If the foregoing calculation suggests the need for 10 or more spectra, you will have to be content with S/N < 100, as other sources of noise (read noise, thermal noise in the dark, etc.) will dominate.

Comparison Spectra

Once you have completed the exposures for your star, it is necessary to take a comparison spectrum before you move the telescope to the next star. To accomplish this, move the GAM slide to position 2 and then turn the comparison switch (located on the comparison control box on top of the TV monitor) from "observe" to "comparison" (make certain the other switch is set to "on" - it can remain there the entire night - see figure above). Using the spectrum window, acquire a comparison spectrum with an exposure of 1 sec. Export to the working directory for the night using a name of the format acXXXX .fit. Move the GAM slide back to position 4 and set the switch to "observe". You are then ready to move to the next star.

Darks

You should try to take Dark exposures at least every hour, more often at the beginning of the night while the CCD is still coming to equilibrium. Try to get a minimum of 4 darks per night, a total of 6 to 7 is preferable. After every dark, take two bias exposures and export to the night's working directory. All darks should have an exposure of 300sec.

Flats

It is not necessary to take Flat fields (flats) exposures every night, as the flats are quite stable over time scales of weeks. You should take a set of flats every month. You can take a set of flats by using the following procedure:

If there is some other grating besides the blue 600 g/mm grating in the spectrograph, it will be necessary to remove that grating and to install blue 600 g/mm grating. The grating tilt should be 8^o6' (note that this is not the standard tilt for observing with the 600g/mm grating). For more information on how to remove and install gratings, see the next section.
Move the GAM slide to position 1, and push the GAM drawer all the way in.
Uncover the slit so that the decker plunger shows about 13 indentations.
In the spectrograph drawer, you will find a metal tube with a diagonal mirror at the end of it. In addition, you will find a white PVC tube with a toggle switch on the end. Push the PVC tube over lens end of the brown 1 1/8'' eyepiece, and then put the combined PVC tube/eyepiece assembly into the metal tube with diagonal, and tighten using the set screw. Insert this entire assembly into the North Port of the GAM, with the diagonal mirror facing down, i.e. toward the spectrograph slit. You can gauge if the mirror is properly aligned by centering the set screw between two screw heads on the North Port of the GAM.
In the spectrograph drawer, there is a little black Radio Shack power supply. Plug this into the power box on the telescope tube, and the other
end into the PVC tube. Turn the switch on. You should see the PVC tube glowing dimly.
Make certain that all lights in the dome are off. Expose for 300ms (milliseconds, not seconds!). To view the resulting flat, set the scaling
on "Dynamic Min/Max'' and "Squeeze'' the frame until you can see the entire frame. The flat will not be uniform, but it should be symmetric from side to side (although not from top to bottom). The maximum number of counts should be on the order of 20000, the minimum (at the top edges) should be about 7000. If you are not getting these counts, adjust the exposure accordingly. If the flat is asymmetric from side to side, this can be corrected by rotating the PVC/eyepiece/lamp assemblage in the North Port of the GAM.
Take about 10 -- 15 flats, and export with the format flXXXX.fit. You may need to turn back a number of pages in the observing notebook to find the running number to use.
At the end, do not forget to take a dark exposure with the same exposure as the flat. Record this with the same format as the normal dark exposures: but with the prefix df: dfXXXX.fit.
I usually take flats the last thing at night before I close down and go home. If you leave the 600 g/mm grating in the spectrograph with the
8^o6' tilt, make a note of it so the next observer will know that he/she will have to adjust or replace the grating before observing.

Changing Gratings

There are currently two gratings (see Table 1) which are in common use. These are the blue 1200 g/mm grating and the blue 600g/mm grating. Both gratings use the same camera focus and tilt, so those parameters will not need to be changed. To change a grating and set the proper tilt, carry out the following steps with the telescope pointed at the zenith:

The gratings are stored in the plexiglass case in the spectrograph drawer. Except for the 1200 g/mm grating (which has a blackened handle) all gratings are marked on their ends with the grooves/mm, and the blaze wavelength. The blue 600 g/mm grating is marked with "600'' and "4000''.
A grating may be removed from the spectrograph as follows: Grasp the black ``handle'' (not the graduated grating tilt knob!!) at the base on the south side of the spectrograph and gently but firmly turn it. The grating ``door'' should come off the spectrograph with very little persuasion. The grating is held in a small ``drawer'' and is secured with a brass thumb screw. Loosen this screw, grasp the grating handle (brass as well) and pull the grating gently out of the drawer. (see figure below)
Remove the grating from the plexiglass box that you wish to install in the spectrograph and replace it with the grating that you just removed from the spectrograph. Holding the grating by the brass handle, gently insert it in the grating drawer. It may take a moment or two to get the angle right, but once the grating is in the proper tracks, it will slip right in; it will take no force to push it into the drawer. Do Not use force in any part of this procedure! Tighten the brass thumb screw and replace the grating door. Secure it by turning the black handle.
The grating tilt must now be adjusted. This requires a bit of black magic. Unlock the grating tilt mechanism (on the east side of the spectrograph, there is a small brass lever -- see second figure below. Push this lever up as far as it will go). Use the graduated knob on the west side of the spectrograph to set the grating tilt recorded in Table 1. Use the vernier scale to set the tilt accurately. Now it is necessary to carry out the fine adjustment of the tilt. This is where the black magic comes in.
Set up the comparison source as described in the main text, and take an exposure of 1 second.
At the front of the observing notebook and at the bottom of this webpage, you will see two graphs of comparison spectra, one for the 1200 g/mm grating, the other for the 600 g/mm grating. Using the appropriate graph, locate the indicated line on the exposure on the screen. For instance, the indicated line on the graph for the 1200g/mm grating should be located at pixel 911. If the indicated line is not at the designated pixel (you can determine this on the screen if you select ``show cursor'' under the display menu') you must change the tilt.
Unfortunately, the graduated knob does not have a vernier motion, and so, with the grating lever unlocked, lightly hold the knob, and with just slight finger pressure rub your fingers over the knob in the direction you need the knob to turn. If you need to increase the pixel number, rub your fingers in the counter--clockwise direction, to decrease, in the clockwise direction. The knob should not visibly turn during this process. After doing this a few times, lock the grating, and take another exposure. If you are doing things correctly, the position of the line should have changed, in the right direction, by 5 to 10 pixels. With a little practice, it is possible to get the line to the exact pixel desired, although a leeway of +/- 5 pixels is acceptable. If the line is off by more than about 50 pixels, it is not practical to change the tilt this way; try to use the vernier scale to get the knob closer to the recommended setting.
Once you have the tilt within acceptable limits, lock the grating, and you are ready to go!

1000g/mm comparison (blue-violet, 2nd order)