Large Format Camera (LFC) Cookbook
View the original cookbook.
The Large Format Camera (LFC) is a mosaic of 6 ccds located at the prime focus of the Palomar 200-inch Telescope. The ccds are SITe SI-002 chips with 15 micron diameter pixels. Each chip is 2048 pixels across by 4096 pixels tall.
At the prime focus (f/3.33), the pixels correspond to ~ 0.18″/pixel (0.175″/pixel at the field center; 0.185″/pixel at the field edge). The total field of view is 24′ x 24′. The field is unvignetted from the center of the mosaic out to a 22′ diameter circle. Beyond that a circular vignetting pattern increases uniformally (due to the Wynne Corrector) to a maximum of 10% at the outer edge (25.3′ diameter circle). There are 15 arcsecond gaps between each of the 6 chips.
If you choose to write all 6 CCDs to disk for each exposure, then you
will typically end up with about 6-8 Gb
of data per night
in the summer and perhaps as much as 10-12 Gb in the winter, maybe even
more if you're doing something truly exotic.
You can transfer data off the mountain using secure FTP if you prefer,
but please note that internet connectivity to our remote mountain site
is not always as stable as we would like it to be.
Note that when your last night is over, we like to have the data room and instrument ready for the next observer by noon the following day. Please try to get your data backed up before this time. If this is not possible, please let somebody know and we will do our best to accomodate you. Also please note that we do not archive data at Palomar. At best, we delete "old" data first when making room for our new observers. We strongly urge you to backup your data promptly, and let us know if you have any troubles or any special needs.
The Palomar daycrew will set up a workstation in the data room to run LFC. Currently that workstation is planetx. If you want or need to set up this workstation yourself, follow these directions:
Log into the Palomar workstation planetx as
You are ready to start the LFC program. Make sure you are in one the of 4 proper data disks:
Now enter the following commands in this order to start the LFC program (the order is crucial):
The Binning Value and Raster Value parameters can be changed should the above options not meet your needs (the chip numbers readout cannot be changed). Seek assistance if you want to do this. For more information, see the mosaic command.
Logsheets can be printed from planetx using:
The ccds collect light at slightly different rates (gain). We've plotted the linearity curves together to create a tool to help you guage this rate difference. This can be useful if you want to keep the raw, unbiased chip4 below 21000 DN while displaying other chips.
There are five LFC mosaic configurations (modes) that you can use. The
The binned modes (2×2; ~ 0.36″/pixel) have been tested on sky, and they appear to work very well. In fact, the wavy patterns in the bias frames disappear in binned mode. The readout is faster (56 seconds instead of 115 seconds) and the image file sizes are smaller (4.3 Mb instead of 17 Mb). The seeing at Palomar is typically ~ 1.0″ to 1.5″ FWHM, sometimes as good as 0.8″ to 0.9″. In the binned mode, this would correspond to ~ 2 pixels per seeing disk in the very best conditions and ~ 3 – 4 pixels per seeing disk in the average Palomar conditions.
Note: When the LFC is installed in the East-West orientation, the Position Angle of the guider must be manually addressed:
The LFC dewar is large enough to hold LN2 for over 24 hours. The daycrew typically fills the dewar once during the day.
Dry nitrogen is continuously blown across the LFC dewar window at 5 PSI. If the weather is very humid you may want to ask the daycrew for more (~10 PSI). If condensation starts to form on the entrance window, you will see dark blobs grow near the center of the LFC field (upper left of chip 0, upper right of chip2...).
The shutter is comprised of two large rectangular blades. While one moves away from the field to start an exposure, the other moves over the field in the same direction to end an exposure, thus providing uniform illumination. The shutter will NOT go faster than 0.6 seconds. You will receive all sorts of trouble if you try to go faster than 0.6 seconds. The shutter is very stable at and above 0.6 seconds. The linearity curves show that the chips are very linear all the way down to 0.6 seconds.
There is no light leak around the shutter, you can take bias frames during the day if you want.
When you are ready to take your first exposure, take a bias frame to clear all of
the charge off of the chips. Charge
will accumulate on the mosaic during the start-up procedure.
This first exposure will likely show residual charge on the chips, you can throw them away if you want. The next exposures will be fine. To take 5 bias frames in a row, type:
Your bias frame should look like the one on the right.
As with most modern CCDs, there is no appreciable dark current in these chips. You can prove this for yourself with:
You'll find some hot pixels, some hot columns, and nothing to be alarmed about.
Once the dome is dark (~ 4:00 pm), have someone open the mirror cover and make sure all of the dome lights are off. All domeflats are done with the telescope pointed up at the white part of the dome slits (no need to move the telescope). If you want to take u' domeflats, you will need to use the highlamp.
See the Filters section for exposure times.
To take 4 domeflats in a row, type:
Your domeflat should look like the on on the right.
Since the LFC has a light tight shutter (unlike COSMIC), you can take good skyflats.
Note the LFC command
The left sides of the ccds (column 1 to about column 75) are ~ 30 to 100 counts "hotter" than the rest of the chip. This is in the bias, and subtracts out nicely.
The field of view is unvignetted from the center of the mosaic to a circle about 22′ out from the center. From this circle and continuing out to the field edge (22′ to ~ 25′) the field is up to 10% vignetted by the Wynne corrector lens. You will see the vignetting pattern at the outer edge of each chip, especially chips 4 and 5.
Look out for condensation "blobs" that might appear near the center of the mosaic field when the humidity gets very high.
If this happens to you, have somebody check that the dry nitrogen is flowing over the dewar window. If it is, the dry nitrogen flow should be turned up.
Each chip has a few bad columns and a few small dead spots. Chip 4 has two bad columns down the center of the chip. Chip3 has small, in-focus "marks" (dare I say, "spooges") all over the ccd (need light to see them and you have to look close). These are defects on the surface of the ccd, and they will flatten out.
The 6 ccds are wired together in pairs to one of three controller boards. As a result, if something saturates (star, background) on one chip you will see a "ghost" of the saturation on the chip that shares the board. The ghost is usually tens of counts, which you can remove later through filter processing. In the image on the right, the core of a bright star saturated on chip 0 during a long exposure, which led to "bloom" trails. The pixels that actually saturated show up in the same location on chip 1, seen here.
You can use Ximtool as a stand-alone to display your images. If you want to use iraf you need to append a .fits to each image filename to keep iraf happy. And note that the images that LFC writes are unsigned integers; iraf wants to display signed integers by default.
You can use the IRAF toggle to scale the output image data to iraf friendly signed integer values. Here's the original ccdcom explanation (this was verified to work on LFC images in February 2005):
IRAF (among others) is finicky about dealing with 16-bit FITS images, but the dynamic range and A/D converters for our CCDs are such that we want to save all 16 bits. The "iraf" option causes ccdcom to scale all FITS images to pack 16-bits of unsigned data (so-called "ushort") into standard FITS signed 16-bit integers ("short" or BITPIX=16 FITS format). Specifically, this the FITS header parameter BZERO is set to 32768, and the image is written as a FITS-standard 16-bit integer data file. IRAF et al. use the value of BZERO in the FITS header to restore the pixel values to those provided by the camera. By contrast, the old "not IRAF" format images stored pixel data as unsigned 16-bit integers, resulting in a non-standard FITS format image (a "ushort" image in IRAF parlance) that most FITS readers cannot read without special handling.
To start an iraf session from planetx, click on the icon called
You should check the pointing at the start of your first night, especially if a previous observer used a different instrument. Have the telescope operator point to a 5th or 6th magnitude star and take a 1 second exposure:
The star should be somewhere near the center of the field, so display chip 0 first and look at the top left of the chip. A 5th magnitude star is bright enough that you should see the reflection somewhere on chip 0, if it isn't totally there already. If the pointing is good, the star might be behind the 15 arcsecond gaps at the center. Change the low-high display values to see the core of the star, and then move the telescope as appropriate to set the pointing wherever you want. You may need to display another chip in order to see the star.
Take another exposure to verify that you put the star in the right spot. When you are happy, have the night assistant "X" the pointing, which sets the telescope's pointing model to its current position. The pointing should now be good for the rest of your run.
During the winter season, you will likely find focus values from 31.00 to 34.00 millimeters. In summer, you may find focus values from 28.00 to 31.00 millimeters. The Sloan r', i', and z' filters are parfocal. The Sloan g' and u' filters are 1.5 to 1.7 millimeters higher.
Note that due to the curvature of the field at the 6 ccds in the focal plane, the best place to inspect focus stars is about 1/3 out from the center of the mosaic. This will provide the best focus distribution over the entire field. Note the following LFC array focus maps:
You can measure the seeing and do some simple statistics with the
The primary mirror is astigmatic. Use this to your advantage.
With the cursor over a star, hit the
If the seeing is bad, say greater than 2.0″, all bets are off; everything is round.
Here are some examples of what you might do when taking normal exposures:
To see the full list of mcdcom commands, see the LFC Commands page.
Most of the guider information can be learned from the guider's HELP window invoked by the Help menu option on the GUI (view this help file).
To start the guider, click the icon labeled
The LFC guider is a SITe 512 x 512 ccd with 0.54″ pixels, corresponding to a field size of 138″ x 138″. This ccd is located in the southeast corner of the mosaic next to LFC chips 3 and 5 at the outer edge of the field. Since it is in the cooled focal plane (which means that it is behind your filter AND the shutter), the guider is very sensitive to faint stars. Two important points:
A typical exposure and guiding sequence:
The guider crashes and dumps core every once in a while. This may be a feature to keep you awake. Also note that the guider isn't happy when you move the guide box too near to the edge of the field (either manually or with a guided offset). Crazy things will start to happen in this case, so quit the guider and simply restart it.
The strip charts are nice tools to help you gauge seeing and sky transparency.
The display scales automatically by default. If you don't like this, uncheck the Auto Scale button. The background level is determined from the sky box, so if a bright star is in the field, you may not like the autoscale.
The guider provides communication between the LFC and the telescope. The guider must be running for the following telescope commands to work:
For a full list of telescope commands, see the Telescope Commands section.
There are several scripts available in the oasis
To run a script, type
After starting a normal exposure, say you select a guide star that is 20″ east and 30″ south from the center of the guider field. The guider field of view is about 90 arcseconds across. When this exposure is complete:
The most common problem with the LFC is a "hung" utility board.
The utility board controls the shutter and the filter wheel. When it
hangs, you won't be able to move the filter wheel or take an exposure.
Usually, a scary
The utility board hangs about once or twice every 2 – 3 nights.
Sometimes, the filter wheel will fail to converge to the proper
position after a move request, and a warning
"filter wheel didn't make it into fine lock" will result.
If this happens, just issue the
For a full list of LFC troubleshooting errors, go to Rob Simcoe's LFC Troubleshooter Web Page.
The tape drives are located next to oasis in the computer room on the mezzanine floor (next to the E-lab). Your taping options are:
We can typically get 7 Gb on a 112M Exabyte tape (5Gb native : 10Gb compressed) using compressed mode, so that's about a 1:1.4 compression of LFC images. A 160M tape (7Gb native : 14Gb compressed) should hold about 10 Gb.
A 125M tape will hold 12Gb native and 24Gb compressed, but we haven't tested this.
Nothing special needs to be done at the end of the night. The dewar should still have plenty of Ln2, and the daycrew will fill it in the afternoon.
When your last night is over, please try to have your taping done before the next observer arrives in the afternoon. With the limited disk space on oasis, your data can't be guaranteed to remain on disk after you leave. Contact the mountain staff and the next observer if you need to make special arrangements.
Please share your feedback on this page or any other Palomar topic at the
COO Feedback portal