Each science chip is read out with 2 amplifiers; the readout time is about 35 secs (in unbinned mode). The readout noise for these is 4.5 e-, and the gain is about 3 e-/ADU, though it is slightly different for each amplifier used (see below). Digitization is done with a 16 bit ADC, so data values up to 65,536 are recorded. With the current gain, the ADC is saturated before the pixel wells are filled. Dark counts are normally insignificant, but are high after the controller has been powered up, and also after the chips have been exposed to very bright light (such as room light). QE curves for the chips are available at this site.
The guide CCD is read out with separate electronics and computer. The guide CCD can see light when the shutter is closed through special openings, as well as see light when the shutter is open. There is no pickoff mirror, and no TV guider or acquisition for Minicam. It is recommended that short, perhaps highly binned exposures with the science or guide CCDs be used for cases where a field verification is needed. For guiding, before the shutter to the science CCDs is opened, a single 2 second exposure is taken, on which a star is identified in the two parts of the image. Subsequent 2 second exposures are taken automatically for guiding during the science CCDs exposure.
The CCDs are in a LN2 dewar, which holds for
about 12 hours. This dewar is mounted to the Megacam
topbox. Inside the topbox are two filter wheels and a shutter.
Each filter wheel has slots for 5 filters. Special holders accomodate
the future Megacam filters, and also inserts for the smaller filters that
Minicam uses. Currently, there are UBVRI filters in residence, called
the SAO set , and a set of polarizing filters. See
below for transmission curves and instructions for using other filters.
The topbox shutter is a two plate design, similar to that used in SLR cameras. This allows accurate exposures over the large focal plane for times as short as 0.100 second. Short exposures are accomplished by using a narrow opening between the plates, and scanning this slit across the CCDs. However, in that case the guide CCD can not be used, and so exposures shorter than 10 seconds must be unguided as a matter of practicality.
The MMT being an Alt-Az telescope, the focal plane rotates, and so of course a derotator takes that rotation out for imaging purposes. Normal Minicam operation is always to align one side of the CCD array with North, rather than at a random angle.
The Sun station called Packrat is used to acquire the science
images., while Cfaguider acquires the guider images.
Note that different keyboards are required for the different computers.
Under normal conditions there is no need to use Cfaguider. Images are taken
by using a graphical user interface called the Mahgui. These images
are stored on a 72 Gby disk on Packrat, in FITS Extension format. There
are four extensions - chip 1 has extensions amp1 and amp2, and chip 2 has
extensions amp3 and amp4. The directories are called /data/mini/MINICAM/year.monthday
(e.g., /data/mini/MINICAM/2000.1009). The files themselves will have names
like a0001.fits, where "a" is a prefix that the observer can select.
Left: Status Monitor
Middle: Control Monitor
Right: Image Display Monitor
The middle monitor, which is packrat:0.0, is the main window you will type in and runs the catalog and queue manager. Two windows are provided here for general use. If this is the first night of your run, you should reset the minicam account to its default state with the command 'reset_obs', entered into one of the windows. This command will reset all the observing parameters back to their default states and clean up any junk left behind by the previous observer. The data directories are not deleted.
The shutter/filter monitor window (also on the left monitor) shows a cartoon of the current positions of both filter wheels and the shutter. Shutter parts are shown in red when they are moving, black when stationary. The green boxes represent the science and guider CCDs. The names of the filters should appear next to the filters. If the names are incorrect, or you have just loaded new filters, you should click the minicam status button "scan". This will rotate both filter wheels past a bar code reader, and display the results afterward in the cartoon.
The Mahgui consists of two major parts: the catalog of objects
& commands, and the queue of those items in the catalog which
have been selected to be executed in sequence. The catalog
is typically prepared before the start of the night, and contains the details
of how you would like the data taken. The queue is loaded up with
items from the catalog during the night, which are then executed according
to their order. The queue itself can be stopped, or individual items
can be removed at any time. The general idea is to keep enough items in
the queue so that it doesn't empty, thus wasting telescope time.
In the Mahgui, the catalog is at the top of the window, and the queue is at the bottom. The relative sizes of the partitions changes as needed.
The required catalog format is a starbase table, which is a tab delimited ASCII table, with optional header keywords. It is possible to import a standard MMT target catalog using the catalog tool. The general format is one of keywords and values. Many data characteristics will be the same for all or most of the data, such as the binning, the observer name, whether to guide, etc. These can be placed in a header at the beginning of the file and will apply to all subsequent specific entries.
After this list of global values, would come the specific entries, where one puts things like object name, coordinates, exposure time and filters. A command in the catalog manager can expand each of the specific entries to show the global values as well. Here is a short example; detailed examples and samples to edit are found below.
# Dithering (none)
title ra dec
pa filts exptime equinox
----- -- --- -- ----- ------- -------
M81BK5N 10:04:41.1 68:15:22 -180 B R 300 120 J2000.0
M81F6D1 9:45:10.0 68:45:54 -180 B R 300 120 J2000.0
So in this catalog, all exposures will be of type object, have binning 2x2, have a filename prefix "a", etc. If queued, two fields would be acquired in sequence. For each of them, a 300s B and a 120s R exposure would be taken, with the rotator at -180 setting. (As a reminder, there are tabs separating keywords and values in the global value section, and between different keywords or corresponding groups of values in the specific entries section. For instance, a tab separates M81BK5N from 10:04:41.1, but there is only a space between B and R and also between 300 and 120.)
For details on making catalogs, refer to this document.
There are two important details in the catalog format: First is the placement of the object on the chip: obscen. This is described in the catalog.html page, but be sure you ask for what you want. Second, is the position angle of the rotator (this is different from instrot). For objects with declinations less than 31:41:18 (the latitude of the observatory), use 0. For objects north of that, use -180. These values will insure that the rotator has enough room to compensate for field rotation for all hour angles.
There are several standard catalogs in addition to your observing one
that you will want to load into the Mahgui at the beginning of the night:
the bias catalog, the skyflat catalog and perhaps the test catalog.
To load, pull down the File menu at the top left of the Mahgui window and
open the catalog you want. You may open more than one at a time.
Selection among catalogs is then done by a button next to "Names".
Once a catalog has been loaded, an item from the catalog can be queued
up for exposure. Move the mouse to the title of the catalog entry
and highlight that entry. Now click on "Queue", which will transfer the
item to the queue, in the lower part of the window. If there are
no other items in the queue, you'll be queried if you want to activate
the queue (start taking images). If the queue is already active,
the new item will simply go to the bottom of the list and wait its turn.
Pointing: It is likely that when you first start up or after a big change in elevation, the object will not appear exactly where you want it. To correct the pointing, take an image, and then use the "Fix Pointing" tool in the Observe menu of mc9. It will place a red and a green marker on the screen. With the left button drag the red marker to the location of the catalog target. Drag the green marker to the desired location. When you click OK, Az/El offsets will be sent to correct the pointing, and revised "Mc-Offsets" will be sent if you changed the location of the green marker. If you want to dither the object around the CCDs, the way to do this is to change the McOffsets. Either the operator can type them in, or you can use the "Fix Pointing" tool. If you want to change the location of the object on the CCDs, say because you want to dither, or the object falls on a bad part of the CCD, the best way of doing this is to change the McOffsets slightly.
Imaging: If your catalog is ready, load it up and queue the first object you want. After hitting the queue button, you'll be asked if you want to activate the queue (meaning should the program actually begin). Go ahead and activate the queue. The telescope computer will receive the coords, and the operator will confirm the move. Away you go. If you have selected guiding to be automatic, guide stars have been selected. Upon arrival, the guide ccd will take an exposure and the operator will be asked to confirm the guide stars. Because of the many bad columns in the guide CCD, sometimes there are problems with the guiding. Thus this process may take some time. After a while, a timeout message will appear on the Mahgui screen, asking you for further instructions, if the guiding hasn't yet started. If guiding has started, then an exposure will ensue automatically.
The exposure will be readout and displayed automatically, and the next exposure in the queue will be executed without further intervention. If you prefer to look at your images before moving to the next field, make sure you have paused the queue before the exposure ends.
Focusing, Collimating, and Guiding are typically done by the operator. Instructions are contained in the MMT Operators manual. A few salient points are mentioned here. The guider CCD is of the same type as the science CCD and is mounted on the same focal plane inside the Minicam Dewar. The guider CCD views the sky through it's own dedicated filters. For the R and I science filters, the guider uses clear glass. To control scattered light, for the U, B, and V filters we have installed a B-like filter. This dramatically cuts down on the amount of scattered light on the science camera and modestly cuts down on the amount of star light reaching the guide camera. Because the current guider CCD has a large number of bad columns, we have implemented a bias subtraction scheme. Whenever the box location is moved or the exposure time is changed, the guider camera will pause for approx 15 seconds while new bias frames are collected.
The functions of mc9 exist in two forms: (1) pull down menus at the top of the window, and (2), panel buttons. To use the panel buttons, which are the top row of the two rows of buttons, click on one of these. The screen will be redrawn with the options available for that particular function (eg., file, frame, scale, etc). The pull down menus, which are a super set of the panel buttons offer many more options. The manual for the public version of ds9 can be found at this web site, but here are some basic features.
mc9 features a main display window, and two small windows which show the whole image (the "pan" window) and zoomed part of the whole image (the "zoom" window). Note that the pan window shows the image orientation on the sky.
Nice postscript data logs can be made automatically. Generally, all you need to do is type initlog in an iraf window, and answer the obvious questions. To display the logs, type viewlog. To print the logs, type printlog, or print the postscript file in the data directory. For more info, see this website.
|Format||Length||Native Capacity||Compressed Capacity|
|(assuming 2:1 compression)|
The DLT is a DLT8000 and is a unit 1, /dev/rmt/1. In iraf, this is mtdlt1. There are also the iraf options mtdlt1h, mtdlt1l, mtdlt1hic, and mtdlt1loc, for all the hi/lo density and compression options
To write a DLT tape at a particular density, use one of these 4 tape devices:
/dev/rmt/1l "low" density
/dev/rmt/1m "medium" density
/dev/rmt/1h "high" density
(add an "n" to the end of the name for norewind. e.g. /dev/rmt/1cn)
The meaning of "low", "medium", etc. is controlled by density codes
boot time in a system configuration file, and by the type of tape used.
DLTtape type capacity (GB)
l = low
DLT7000 mode, uncompressed (35 GB/DLT IV tape)
m = medium DLT7000 mode, compressed (35-70 GB/DLT IV tape)
h = high DLT8000 mode, uncompressed (40 GB/DLT IV tape)
c = compression DLT8000 mode, compressed (40-80 GB/DLT IV tape)
NOTE: The default is set to DLT7000 compressed ("medium").
Note that all but the "h" and "c" devices for the DLTtape IV will write
tapes compatible with DLT7000 drives, so you use these for backward
In normal use it may be easier to set the density and compression using
the "Select" button on the front panel. If this is engaged, density change
commands via the device name are ignored.
It is possible to bring your own filters for use with Minicam with some advance preparation. We have manufactured one extra filter holder for guest filters which can accomodate filters ranging in size from 4 to 5.75 inches on a side. If you are bringing a large number of filters, you may need to construct extra filter holders.
The guide camera is split into two halves, with each half slightly defocussed from the science focus, but on opposite sides. Once we switch to f/5 this will enable the autoguider to track the telescope focus automatically.
In addition to the science filter, you must provide a piece of glass that goes in the filter wheel over the guider CCD to make it cofocal with the science CCD. For science filters with high background, this guide filter can be clear glass. For narrow-band and blue-band imaging it is recommended that the guide filter be a B-band-like filter rather than clear glass to reduce the potential for scattered light in the camera. The trade-off is the sensitivity of the guide camera is reduced by a factor of 2, but that the dark sky background coming through the guide filter is reduced by a factor of 10.
The required dimensions of the glass are
width 107mm (+0.00 -0.25 mm)
length 72mm (+0.00 -0.25 mm)
(clear aperture 100mm x 65mm).
For clear glass, use Borofloat with a thickness of T = 1.47 sum (ti/ni) - 1.46mm, where ti and ni are the thicknesses and refractive indexes of the individual layers of the science filter.
For a blue guide filter, we recommend 1mm BG12 + Xmm BG39, where X = 1.57 sum(ti/ni) - 2.59mm. We will provide a guest piece of BG12, so you need supply only the BG39 unless you have several filters.
A tolerance of 0.15mm is acceptable on the thickness.
I have purchased such pieces of glass from Paragon Optical http://www.paragon-optical.com 1-800-508-9722.
Make sure to notify the MMT staff that you will need assistance loading filters before you arrive.
There is currently no method of taking dome flats. Rather, you should be sure to take twilight flats or collect enough dark sky exposures to make a dark sky flat.
Zeros should be taken for calibration, but darks are unnecessary.
2 2.7 4.1
3 3.2 4.6
4 3.2 4.8
For a 20th mag zero-color star at 1 airmass, one gets (gain = 3 e-/adu):
U 76 ADU/sec = 220 e-/sec
B 360 ADU/sec = 1100e-/sec
V 280 ADU/sec = 810 e-/sec
R 300 ADU/sec = 890 e-/sec
I 210 ADU/sec = 620 e-/sec
For rough calibration then, the standard magnitudes can be found from
the instrumental magnitudes measured at
[e.g., b = -2.5 log(counts/exptime)] by:
U = u + 24.7
B = b + 26.4
V = v + 26.1
R = r + 26.2
I = i + 25.8
Here are dark sky background counts, with proper baffling installed.
The numbers refer to a 2x2 binned pixel. The scattered light previously
measured has been reduced by a factor of 10 by installing colored glass
in U, B, and V guider slots. The rate of the scattered light (now
negligible) is also listed below, under Scat. This light will appear
at the quoted rate in U, B, and V exposures. It will be highrer
at R and I, but is negligible. The sky background at B-band has not
been measured accurately. An upper limited is given.
Many problems can be diagnosed using the Minicam Status display. The top section shows the status of all the servers in the Minicam client/server system. Green indicates that the server is up. Red indicates that the server is not running. Yellow indicates that the server is up but not responding. Sometimes servers will go yellow briefly if they are very busy. The servers are as follows:
The remaining items indicate the status of the topbox.
If you cannot access the cfaguider window with the packrat mouse , first make sure that there is a strip of the root window visible along the left edge of the middle screen. Otherwise try restarting X2X using the pull down menu from the root window.
Restarting various windows. To restart one of the status windows use the pull-down menu from the root window on the left screen. To restart mc9, in the ICE window type, e.g. "postproc MINICAM/2002.0313/a0001".
To shutdown either packrat or cfaguider, type:
Note that if the console is locked up, it still may be possible to enter those computers via another computer (e.g., hacksaw or hoseclamp), using rlogin or telnet, after which you can try killing the Xsun process. If this fails, then execute the userreboot or userhalt command. If for some reason you need to do a "STOP-A" to reboot, please type 'sync' twice after doing the "STOP-A". This will prevent the disks from being corrupted, and save you a lot of time and pain during the rebooting process. All problems (weather not withstanding), comments and suggestions should be sent to Brian McLeod. Remember, if you don't report it, it can't be fixed!
X. Data Reduction for Minicam
A few preliminary notes about data reduction of minicam images.
Iraf 2.12 has a bug in mscred in the current version.