Kermit Software

Overview

• kermitui is a graphical user interface written in Tcl/Tk
• leachd handles the low-level communications with the Leach box electronics and thence the detector
• soundserver plays sounds in response to various system events.
• traffic is a back-end communications server used to link the various components of the system together.
• ds9, from SAO R&D, is used for quick-look image display.

The GUI provides automatic electronic logging of all exposures, motor moves, and other important events, including support for free-form user comments. This is meant to completely replace the need for taking a paper logfile during the night.

The electronics rack supports hardware coadding, in which multiple images are read from the chip and added together in hardware before being transferred to the computer. The detector can be read out into the Leach electronics at up to 32 Mpix/s (i.e. 1 MHz per channel times 32 channels). However, data can only be transferred from the Leach electronics over the fiber into the computer at 1.5 Mpix/s. A full-frame CDS image is 8 megapixels, and thus takes about 5 seconds to read into the computer. For this reason it is highly recommended to use the hardware coadding feature rather than reading many exposures to disk, particularly for shorter exposure times.

Starting the Kermit Software

In a different xterm, run kermitui. This will start the graphical user interface and initialize the camera hardware.

Also start ds9. New images will be automatically displayed in ds9 as they are taken.
Note: Sometimes the ds9 process hangs. If this happens, run fixds9 to kill the hung one and start a new ds9.

GUI: Main Window

The controls in the main window are grouped loosely as follows. At the bottom of the window are text entry boxes for the FITS file headers and log files. At the top left of the window are the most frequently used settings and the actual Expose button, while to the top right are less frequently changed settings relating to array readout.

Each entry field in the GUI is associated with one of the [Set] buttons, usually adjacent to it. When you make changes to any of the values in the GUI, the appropriate set button will change color to indiate that changes have been made but not yet taken effect. Press the [Set] buttons to apply changes. The two large [Set] buttons are equivalent; both apply all array readout configuration settings - there are two just for convenience. A button or field will turn red to get your attention if it is critical that a parameter be set, or if there is a problem with the selected options.

Exposures are controlled by four buttons, as follows:

• [Expose]: Initiate an exposure (or a series of exposures if NExpose > 1). A countdown on screen will indicate exposure progress.
• [Abort Exp]: Stop the current exposure, even in the middle of things. Warning! There is some chance this may crash things; it's not fully debugged yet.
• [Stop Series]: If NExpose is > 1 (see below), then this button will allow the current exposure to finish but will stop the system from continuing to take more exposures. Unlike Abort, this one's guaranteed safe.
• [Reset DSPs]: This will completely reset the DSPs in the readout electronics. If things get hosed, this is your best bet for fixing them.
• In addition there is a checkbox marked "display". If this is checked, and a copy of ds9 is open, then each new image file will be automatically loaded into that ds9.

The exposure settings you can control are as follows:

• Exposure Time: This sets the integration time that the detector will pause for. WARNING: Currently the actual integration time for CDS reads is the Exposure Time plus the time it takes to read out the array. I'll fix this once we've stabilized what the pixel readout time is. The software records in the FITS header both the Exposure Time and what it thinks the true integration time was. The Exposure Time can be input in either milliseconds or seconds. To toggle between these options, just click on the Exposure Time label and the label will change to reflect the exposure time units and convert the current input time to the new units.
• NReads: Number of CDS reads
• NCoadds: Number of hardware coadds to take per image
• NExpose: Number of separate exposures to take and write to disk.
• Exposure Number: Kermit files are named ker0000.fits through ker9999.fits. This parameter is the number for the next file to be written to disk. This sets itself automatically on startup and should rarely need to be changed by the user.

• OBJECT: Name of target. Written to FITS header.
• COMMENT: Arbitrary comment for FITS header.
• PROGRAM: Title of observing program for FITS header.
• Log: Any message entered here will be recorded into the electronic logfile.

• nresets: Number of times to reset the chip between frames.
• nframes: Number of separate frames to read out at once. Can only be set to values greater than 1 for subarrays <=512 in size.
• subarray:size: Size of sub-array to read out, in pixels. Must be a power of two between 64 and 2048. For configuring subarrays, the blue region on the chip map (just to the right of the "Set" button") will indicate the exact portion of the detector that will be used with the current settings.
• subarray:coadder: Which coadder to read the subarray from. The 32 readout channels are read out by 8 coadder boards (4 adjacent channels each). This architecture places certain restrictions on the subarray position such that you cannot position a subarray completely arbitrarily.
• subarray:offset: The number of rows into the selected coadder that the subarray region should start.
• Pixel readout rate: 4 microseconds vs. 2 microseconds per pixel. Right now only 4 microsecond mode is known to work properly. Right now the actual pixel time is something else, probably around 8 us or so. Changes and tweaks to this await the detector being cold for testing.

  There is also a "4 us RESET" mode. This mode acts like the 4 us mode, except the RESET line is asserted throughout. While completely useless for actual observations, this mode allows detector read noise tests while the detector is at room temperature.

• Readout Mode: The algorithm used for reading the chip out, as follows:
  • Single Read: Reset, integrate, read Nreads times.
  • CDS (Correlated Double Sampling): Reset, read Nreads times, integrate, read a second Nreads times. Subtract the total of the first reads from the second.
  • CDS, no subtraction: Same as CDS, except the subtraction is not performed and both sets of reads are written to disk. The file written is saved as a FITS file of dimension [N,N,2].

GUI: Motor Control

The Motor Control window controls the filter wheel, pupil and M2 mirror slides, and detector focus stage. For each motor, the current position and a desired new position are displayed. You move the the motors to the Desired positions by pressing the Move button; one motor at a time will move into position. Stop will instantly stop all motion, but will leave the motor controller in a confused state and will require re-homing the motors before moving again. Update refreshes the display of status for each motor. This will usually happen automatically but feel free to press Update if for some reason you suspect the display is outdated.

The Filter, Pupil, M2 and Focus labels are grey when the Current and Desired filter wheel positions are the same. The labels are pink when the Current and Desired filter wheel positions are different.

The Current positions (or last recorded positions, if Status unknown) is listed for each wheel.

The Desired or next position is specified using the drop-down lists or the first three motors, and by entering a number between -8700 and 0 for the focus motor. You can home the first three motors by selecting Home from the list. To home the focus stage, choose "HomeFocus" from the Config menu (see below).

The Config menu has a few additional motor-related commands and settings.

• Config | Dark moves the motors to the proper filter positions for taking dark exposures.
• Config | HomeFocus homes the focus stage.
• Config | Enable motorname Jog enables the Jog functionality for motor motorname. The jog command lets you move each motor an arbitrary number of counts in either direction. It is disabled by default for all motors to prevent accidental jogging.
• Config | Advanced | Stop Motors allows you to stop the motors if you experience a problem with motor motion.

• Config | Advanced | Enable Move allows to to make the Move button active if it has been disabled (usually it gets disabled if there was a problem moving motors or homing motors). It is best to make sure any problems are fixed before clicking Enable Move.

GUI: Exposure Palette

The Exposure Palette lists a number of standard exposure times in milliseconds and number of reads. When you click on an exposure time it updates the information in the main ircalui window. The numbers to the right of each button are the number of images (including number of coadds) taken with that exposure time and number of reads. When you input a custom exposure mode, it will be added to the exposure palette automatically after its first use.

The buttons are labelled using a unique identifying string for each exposure mode, known as the "mode string". The format of a Kermit mode string is as follows:

• T_ms for an T millisecond exposure using the full detector and default read and reset settings.
• subX-Y-Z:T_ms for a T millisecond exposure using a subarray X by X pixels in size starting on coadder Y, row Z.
• subX-Y-Z:T_ms:N:R for a T millisecond exposure, with N CDS reads and R resets, using a subarray X by X pixels in size starting on coadder Y, row Z.

• File | Take Darks brings up the Dark Widget for taking dark exposures.
• Config | Rebuild adds exposure times and configurations that have been used to the Exposure Palette.
• Config | Edit brings up a separate Edit window listing all the current entries in the Exposure Palette. Checked (red) checkboxes indicate that the exposure will be included in the Exposure Palette, unchecked exposure times will be deleted. When finished editing, click OK button. If you remove an exposure time from the Exposure Palette, there are a few ways to re|insert it into the Palette. Easiest is to take an exposure of that type and Rebuild the Exposure Palette. If the desired exposure is in one of the default lists (see commands below), you can reload the list. If you have already taken an expousre of the desired exposure time, just Rebuild the Palette and it should reappear.
• The following three commands import a list of exposure times into the Exposure Palette. Exposure times that you have used that aren't on the default list are retained in the Exposure Palette.
• Config | Add default Short Exps loads a default set of short exposure times into the Exposure Palette.
• Config | Add default Long Exps loads a default set of long exposure times into the Exposure Palette.
• Config | Add default All B loads the list of all exposure times for which Reference Biases are available.
• Config | Zero sets the number of exposures taken for each exposure time to zero. This is advisable at the beginning of every night (or observing run) so that when running the Dark Widget at the end of the night (or run) will only try to take darks of the exposures that were actually used during the night.

GUI: Dark Widget

The Dark Widget looks at what exposures have been taken on the Exposure Palette and calculates how long it will take to do 10 darks of 10 coadds of each exposure type. You can modify the number of darks and number of coadds until you reach an acceptible total time for the darks (most observers try to pare down the list of dark exposures so they take less than three hours of darks). Make sure the filter wheels are in the Dark position (motor - Config - Dark). Then press Go and it will start a script for taking the specified dark images.

GUI: Analog Offsets

Once you've set the offsets as desired, press the Set button to apply the changes to the detector. You can load one of three default sets of offsets by pressing the "Fast", "Slow", or "Warm" buttons. These load offsets appropriate for readouts at 4 us, 2 us, and 4us with the detector at room temperature.

The offsets for each of the 32 readout channels can be edited individually. You can also set all the offsets to one value using the Change All button. Enter a value in the input field next to the button and then press it.

The best way to determine reasonable offsets is to put the detector in single read exposure mode. Take an exposure and examine it in ds9. If the exposure is uniformly 0, the offsets need to be raised. If the exposure is uniformly 65535, the offsets need to be lowered.