Jim Harwood
June, 1999
TABLE OF CONTENTS:
This document describes the procedures for performing a sky map, which acquires a data set of telescope pointing error, and for transforming the sky map data set into a file of polynomial coefficients which is transmitted to the TCS over the RS-232 communications link. The transformation is performed by a program that does a least-squares fit of the data set to an analytical surface that is described by a set of orthogonal polynomials. Coefficients of the polynomial terms are computed fron the input data set and are used in determining the pointing error at a particular coordinate.
A discussion of how to plot the pointing error data is also included.
A summary of the sky map procedures is near the end of the document..
The following procedures are performed on the Sun machine, unless otherwise noted.
Currently, the sky map executables are in the directory /benchmark/skymap on herschel (accessible directly from any other Sun computer on the IRTF local net). Switch to that directory using the cd command. The programs are executable from any login.
There are two communications lines to the Forth TCS computer. The workstation max is connected directly to an auxiliary RS-232 port on the LSI-11/23. That channel is used for typed commands from a terminal emulation window (Xterm) on max, the operator's station. If the terminal emulation window is not running, you can run your own terminal emulation from any Unix workstation by remote logging into max and typing
cu -l /dev/cua/a
where "-l" is dash ell, not minus one.
[Important note:] At this writing, this command doesn't work on max. See a computer specialist or telescope operator for a temporary workaround.
The other communications line is from stefan via /dev/cua/b through A/B switches to the principal (boot) RS-232 port on the LSI-11/23, the one shared with the local TCS terminal. This channel is for programmed commands to the TCS from a C-language application running on a workstation or PC, using the "tcsd" daemon running on stefan.
The telescope operator can run the sky map programs from an Xterm window on the same monitor (max) he is using to operate the TCS in a terminal emulation window.
Test the communication link to the LSI-11/23 TCS computer. To test the link, do the following on the Sun: (Assumes TCS is booted and on-line, and relevant A/B switches set properly):
Type tester. The character string ?BEAM is sent to the TCS, which responds by sending back ABEAM. If nothing or garbage comes back, try it again; sometimes, spurious characters are sent with the string the first time.
If the tester program fails, check the continuity of the serial lines from the Sun computer (stefan) to the TCS. Usually, some A/B switch is in the wrong position or a cable is not plugged in.
Check for the communications daemon (tcsd) running on stefan. Follow the instructions for doing this check and for reloading tcsd.
If the comm link still fails, reboot the TCS and bring it on-line and try tester a couple more times. Rebooting both stefan and the TCS has been known to work when everything else fails. Assuming you now get a response from tester, proceed with the pointing run. Communications with the TCS should be OK. Unless you can get tester to work, skymap will not work.
Set the track rate to sidereal; remove any differential rate. Position the secondary mirror in the preferred position, making sure it is not free floating (controller powered off) or chopping. Autoguiding should be disabled.
For best accuracy in slewing to sky map stars, do an index/collimation coefficient procedure (PB6 on two stars).
Initialize the coordinate system on a star near the zenith with a Pushbutton 5.
It is not absolutely necessary to do the Benchmark. However, to get a picture of the pointing accuracy before and after a sky map operation, it is useful to do a Benchmark before and after the sky map. The Benchmark operation is a variant of the skymap program. skymap is discussed in detail below. Go to skymap right now if you aren't going to do a Benchmark.
Make this entry in the TCS window. The skymap program requires this TCS command.
Benchmark differs from the full sky map in that the pointing error it records is the net guiding error at each star, not the raw slew error. Also, Benchmark uses a much smaller sky pattern file. It takes about 20 minutes to do a complete Benchmark.
Refer below for a discussion of the skymap output file naming convention (yymmdd .*).
Call the sky map program on the Sun workstation with the -b option, using the standard benchmark pattern file:
skymap -b bmark.pat yymmddbm1.map
where we named our .map file with the date and "bm1" indicating the first benchmark run for this date. The bmark.pat file contains 16 stars 30 degrees apart on declination scans separated by 1 1/2 hours HA. This file is listed at the end of the hardcopy version of this document and as a link in the WWW version.
The Benchmark immediately starts, with the coordinates of the first star appearing in the operator's TCS window. A double tone sounds, which is the cue for the T.O. to press PB1 to slew.
The message is displayed in the Unix window: "Keep the data for this star? (y/s/n)". Ignore the message for now. You will answer it after guiding to the star.
Press PB1 and hold down the Slew Enable switch. The telescope will slew to the star.
When the slew is complete, guide the star to the crosshairs.
Answer
<CR> in the Unix window to keep the data for this star, "n" to reject the star.A new star is selected, sounding the beep. If you answered <CR> to the previous question, proceed back to "Guide to the star". Otherwise, to "Slew to the star". This loop continues until all the coordinates in the pattern file are exhausted. It will take about 20 minutes.
The message "FINISHED" appears in the TCS window and a double beep sounds, requesting a press of PB1 for a slew to the zenith. [NOTE: at this writing, FINISHED doesn't appear for some reason.]
The following command string will reduce the Benchmark data set without sending any data to the TCS computer (Benchmark data is not used by the TCS):
skyfit -v yymmddbm1.map > yymmddbm1.out
You will be asked if you want to overwrite a file named AandC.data. Answer yes (y). That is the default file that receives the computed coefficients data. We don't care about coefficients for Benchmark.
Inspect the Benchmark output file .out with the more command: more yymmddbm1.out, or use your favorite editor. The first section of the output file shows the guide errors for the 16 stars. At the end of the section is average and RMS errors. This RMS error for each axis is the figure of merit for this Benchmark. Ignore the rest of the file, including bogus RMS numbers at the end. (We aren't fitting anything with the Benchmark operation.) Excellent Benchmark RMS numbers are 2 - 3 arcseconds.
The individual errors in the slews to the Benchmark stars shouldn't be much over 10 arcseconds, if that. If there are large outriders, something is wrong with the slewing, expecially if this Benchmark was done immediately after a pointing run. The first star may show an anomalously large error. If so, delete that star from the data set. I don't know why this happens, but the error on the first star of a sky map is often very high, not representative of the general pointing.
If this is the Benchmark prior to a pointing run, continue with the sky map operation. Otherwise, type -POINTING to restore the MCC pushbuttons operation and continue with normal observing.
This is the full fledged sky map procedure that slews to 130 or so stars and provides a calibration of the pointing error over the major part of the sky.
Initialize the coordinate system on a star near the zenith with a Pushbutton 5.
Choose a pattern file for the desired map pattern. You can see the available pattern files by typing ls *.pat at the Unix prompt. For example, file irtfWE.pat contains the pattern that was used in previous sky maps. (At this time, this is the preferred pattern file.) Inspect a particular file by typing more file.pat where "file.pat" stands for the chosen pattern file, or use your favorite editor. irtfWE.pat is listed at the end of the hardcopy version of this document and is referenced in a link in the WWW version.
Note that the hour angle coordinates in the pattern file are expanded by the cosine(dec) factor, for uniform weighting of the data points in the hour angle space. If you want, you can create your own pattern file by copying and editing one of the current pattern files.
Choose the sky map output data file name (*.map). If it is a currently existing file, you will be asked for permission to overlay it. Let's use the following naming convention, to keep from accidentally overlaying data files:
yymmdd.map
For example, if the sky map were being done on the day I am typing this, the file would be named-
990615.map
This format is useful when listing files in ASCII numerical sequence: the files will be displayed in order by dates.
The following is to be typed in the TCS window:
Define the focus that will be in force for this sky map by typing one of these commands.
+POINTING typed on the LSI-11/23 disables certain pushbutton functions that would be dangerous in a sky map. If you have just done a Benchmark, you are probably still in +POINTING mode.
Call the sky map program on the Sun workstation. Use one of these forms:
Starting from the beginning (the usual):
skymap irtfWE.pat yymmdd.map
Resuming from the middle of a previously interrupted sky map:
skymap -l <line#> -a irtfWE.pat yymmdd.map
where yymmdd is the year-month-day; the -l (ell, not one) permits entry of a pattern file line number, if you are starting from the middle of the pattern file; and the -a says to append the data to the end of this file, rather than starting a new file.
Skymap will automatically (not optional) produce a TPOINT data file as the sky map progresses. This file has the name yymmdd.tpt, which is the same prefix as chosen for the .map file. (TPOINT is a graphical data analysis program for analyzing telescope mount alignment.)
[NOTE:
The TPOINT file is in the wrong format. I misunderstood Pat Wallace's documentation. Andrew Pickles at the 88" wrote a converter program that converts from my format to the correct TPOINT format.]The sky map immediately starts, with the coordinates of the first star appearing in the operator's TCS window. A double tone sounds, which is the cue for the T.O. to press PB1 to slew.
The message is displayed in the Unix window: "Keep the data for this star? (y/s/n)". Ignore the message for now. You will answer it after guiding to the star. Proceed to slewing (below).
Press PB1 and hold down the Slew Enable switch. The telescope will slew to the star. We won't mention anything about taping down the slew enable switch, since the Switch Police will undoubtedly swarm in and arrest you if you do.
When the slew is complete, guide the star to the crosshairs.
The message had previously been displayed in the Unix window:
"Keep the data for this star? (y/s/n)". After guiding to the crosshairs type <CR> to accept the star and start slewing to the next one. For the record, these are your alternative entries to the question. Use a <CR> after each one:A new star is selected, sounding the beep. If you answered <CR> or s to the previous question, proceed back to "Guide to the star". Otherwise, to "Slew to the star". This loop continues until all the coordinates in the pattern file are exhausted. It will take upwards of two hours.
After the last star in the pattern file, the message "FINISHED" appears in the TCS window. [FINISHED doesn't appear now for some reason.] A final slew to the zenith is started. Now, continue to "Skyfit" below to reduce the sky map data set.
a) Do not do a Pushbutton 5. (Ordinarily, if +POINTING is still in effect on the TCS, you wouldn't be allowed to do a PB5.) You don't want to change the coordinate reference between runs.
b) Exit the skymap program by typing the <Ctrl-C> character.
c) Identify the index (ID) number of the last sky map star successfully measured. The list of consecutive sky map stars with their ID numbers should be displayed in the operating window for max. You can also find out the ID number by looking at the yymmdd.map file and the irtfWE.pat pattern file and comparing star coordinates.
d) When restarting the sky map, use the -l and -a options of the skymap command, so that the proper resumption point in the pattern file is used and the data are appended to the yymmdd.map file. Use the same file names as used in the interrupted session, for example:
skymap -l 46 -a irtfWE.pat 940623.map
This command will resume the sky map at Line 46 of the pattern file irtfWE.pat and append the pointing data to the existing map data file 940623.map. An error occurs if the map file doesn't already exist (you probably spelled it wrong).
I recommend starting with the last observed star in the interrupted pattern. The pointing error should be close to that recorded for this star the previous time. If there is a large difference, that means there may have been a shift in sky coordinate reference between runs, such as a PB5. You may then want to consider adjusting all star errors for this run by the difference, later on when you analyze the data set. A graphic display of the data set (see Data Plotting below) will readily show a shift in coordinate reference. Right now, the only way to adjust numbers in the data set is with an editor and hand calculator.
The output from the skymap program is displayed on the screen, so you can see the progress of the sky map. The output may contain useful information for future use, and it would be nice to be able to save it. There is a way to do this with the command snapshot, which should be available with the right mouse button on the workstation desktop. See a telescope operator.
Use the following command format, whereby the verbose mode is used and the -t option sends the coefficients to the TCS computer:
skyfit -v -t yymmdd.map yymmdd.data > yymmdd.out
Choose the same prefix name for the .data file and .out file as was used for the .map file, for consistency. The prefix should be in the date format as discussed above.
If for some reason there already was a file of the same name yymmdd.data, you are asked if you want to overwrite it. You probably want to say yes. The .data file contains the table of coefficients derived from the sky map. The old .data file was probably created during a previous execution of skyfit using the same .map file as input, possibly with or without the -t option. It doesn't hurt to overlay the old .data file with the same data calculated again by skyfit. If you say no to the request, the program is aborted.
skyfit will take about 5 seconds to compute the coefficients and 15 seconds to send the data set to the LSI-11/23 TCS computer. You are shown the progress of the sending of the coefficients.
At completion, the new set of polynomial coefficients will be stored in the TCS LSI-11/23 computer system's memory, but not on diskette. A section below discusses saving to diskette.
Inspect the output file *.out on the Sun workstation by printing it or displaying it with an editor. To print it, type-
nenscript -2rG <filename>.out
where <filename> is the prefix of your file name. You might need to establish the name of the printer, with the -Pirsummitpr option in the command.
The *.out file contains the following items:
Look for obviously bad data points in item 3 above, whereby the residual error is substantially greater than, say, 10 arcseconds in one or especially both axes. If such data points exist, make a duplicate of the .map file with a different name (cp yymmdd.map yymmddbak.map) and delete the lines containing the bad data points from it with an editor. Refer to the ID number to locate the proper line to delete in the .map file.
Re-run skyfit, using the renamed file. The RMS error (item 4) should be significantly smaller as a result of the deletion. If not, the deleted data point should probably be restored.
Inspecting a plot of the data set is a good way to visually identify bad data points. See Data Plotting.
Save the set of pointing coefficients onto the diskette of the TCS computer system. If they are not already mounted, mount the TCS diskettes in both drives and type on the TCS: PUT-CASSEG for a Cassegrain data set, or PUT-COUDE for a Coude data set. Mark the diskette #1 with today's date in the "pointing" section of the label.
You may get a "disk error" condition reported in the TCS window when trying to do the above operation. If so, the LSI-11/23 is having trouble writing to the diskette in drive 1. Make sure there is a diskette in the drive and the drive is powered up. Try the operation a couple more times. If it is a hard error, use another pair of diskettes and put a sticker on the bad diskette identifying it as bad.
It is a good idea to check that the recently derived pointing polynomial coefficient data set got transferred to the LSI-11/23 memory correctly. It is fairly easy to compare the pointing data set in the TCS with the data set in the yymmdd.out data file. Use the printed file or the equivalent screen display. The last section in the .out file is the set of polynomial coefficents.
Display the pointing data set resident in the LSI-11/23's memory by typing ?COEFFS in the TCS window.
The entire set of coefficients is displayed, most off the top of the TCS screen. Scroll back and compare the first and last C coefficients for each axis in the TCS with the values in the .out file, and the second and last Alpha coefficients. (The first Alpha coefficient is 1, so ignore it.)
A Benchmark before and after the sky map run should show the improvement in pointing. Note that to do a Pushbutton 5, you will have to re-enable the pushbuttons with -POINTING and then +POINTING after the PB5, the mode required by the Benchmark operation.
The .map and .tpt output files are now complete. Type +POINTING on the LSI-11/23 to restore normal setup of the MCC pushbuttons.
This completes the sky map procedures. See below for plotting the sky map data set.
Plotting the data set gives a more positive identification of bad data points than attempting to decide on magnitude alone of the residual error. Plotting the raw errors shows when a data point is much different than its neighbors in its direction or magnitude. A plot of the residual errors should should show a more or less random scatter of error magnitudes and directions. A systematic pattern indicates something wrong. (A "residual" is the difference between the raw measured error and the computed error as computed from the polynomial set.) If additional data points need to be deleted, do so with an editor and return to "Skyfit" above.
Be conservative when deleting data points. Only delete those that seem obviously bad, due perhaps to mis-identifying a star or making a mistake in procedure during the sky map. Each data point is important to define the error surface in the sky area in the vicinity of the measured star. If it truly was a bad data point, re-computing the polynomial set after deleting the data point should produce a noticeable improvement in the RMS error over the data set. If there is no significant improvement in the RMS, consider putting that data point back into the data set.
Frequently, one or both of the axes shows a large bias in the average raw error, sometimes amounting to over a hundred arcseconds. Although a large bias like this has no effect on the accuracy of fit, since it is just a constant offset from some arbitrary zero point, it does distort the plot of the errors for both axes, from the point of view of perceiving the relative contribution of the HA and declination error components.
If you ran skyfit with the -n option, the file *.out will have normalized data and you can skip the rest of this section.
You can see the amount of the bias in the file *.out. It is the signed average raw error for both axes, listed right after the sky map stars in Item 2 of the .out file. If there were no bias, the average would be zero, as much accumulated error in the positive direction as in the negative direction. If the data set needs to be normalized, run skyfit with the -n option:
skyfit -n -v yymmdd.map > yymmdd.out
Say "y" to the overwrite question; we don't care that the default coefficients file (AandC.data) is being overwritten. For plotting, all we want is the .out file.
If you inspect the output file after specifying normalization, the values of the raw errors in the part of the file where the raw errors and residuals are listed (Item 3) will show variation about zero. Item 1, the list of stars and raw slew errors, will show the unadjusted data set. Item 2, the average of the raw errors over the data set, will have been set to zero by the normalization process.
Attached (to hard copy versions of this manual) are examples of vector plots of a pointing data set. The plot examples are in PostScript form in my sky map directory (~harwood/skymap/) on the IRTF mountain network. The files are named cajan94err.ps and cajan94rsd.ps. (The names reflect Cassegrain pointing run and raw error and residual plots; this is somewhat different from the naming convention previously described.) If you have a PostScript viewer (Ghostview), you can view the example plots directly; otherwise, print them as follows-
lpr cajan94err.ps or lpr -P<printername> cajan94err.ps
The plots consist of vectors representing ha/dec pointing errors at the ha/dec coordinates (y/x axes). The plotted vectors can be either raw errors or residuals.
The vector plot is done using the sm program within the Openwin (X) environment. Enter that environment by typing openwin at the Unix prompt. (You may already be in Openwin, especially if you are a telescope operator on max.)
You specify to the program the name of the data file and (as an option) if you want residuals plotted; raw error is the default. The plot is then drawn on the screen. A command can cause the output device to be a laser printer, so the plot will be printed the next time the plot command is given.
The first thing we have to do before we call the plotting program is to condition the .out data file for plotting either raw setting errors at each star or the residuals after fitting, both of which are carried in the .out file. A program prep is executed at the Unix prompt to take the appropriate data from .out and make either a setting error data file or a residuals data file from it. After these plot files are made, we can enter the sm plotting program.
The X display system needs to know to which display screen to send the output if you are working at a remote workstation, a usual situation. Assume you have logged in remotely into herschel to execute the plot program. Set herschel's display environment as so:
setenv DISPLAY [local]:0
where [local] is the name of your workstation.
It may be that your local machine is not set up to establish an X connection to herschel. If you get an error about not recognizing a host, tell your computer about herschel as so:
xhost herschel
All the summit and major Manoa IRTF workstations should already be xhosted for herschel as well as each other.
Choose the skymap output file that you want to plot. This will be a yymmdd.out file. You probably want to plot both the raw error and the residuals. The raw error is the absolute pointing error for each star in the sky map; the residuals are the differences between the computed error and actual pointing error for each star.
Prepare files to plot from by executing the filter program prep as so:
For raw errors:
prep yymmdd.out yymmdd.err
For residuals:
prep -r yymmdd.out yymmdd.rsd
The input file to prep is yymmdd.out; yymmdd.err or yymmdd.rsd is the resulting output data file to be plotted. If no output file is specified, stdout (the screen) is used. You will need an output file to use as a source for plotting.
prep selects lines from its input file containing the desired sky map data and creates a new file containing only the relevant data to be plotted.
(If the sm command is not recognized, you need to include the directory on herschel: /usr /local/bin in the path.)
Run the sm plot program by typing sm. A new window opens up, which is to eventually receive the plot. Before doing the plot, you need to load in two macro files. Type at the sm prompt (":") -
macro read macro.sm macro read vplot.mac
[If for some reason you want to plot the benchmark data set, not usually useful, use macro read bmplot.mac instead, and substitute bmplot for vplot in the following instructions.]
Now at the sm prompt, plot the data file by typing-
vplot yymmdd.err
where yymmdd.err (or yymmdd.rsd) is the name of your data file. You will see the data plotted in the previously opened window.
You will need to manually erase the plotting window before plotting another data set.
To print the plot, you have to specify the printer as the output device and then make the call to vplot again. Type exactly as given, assuming your printer's name is irsummitpr, with the appropriate name of the file to be plotted in place of <name>.err -
dev postscript irsummitpr vplot <name>.err
After a minute or so, the ":" prompt returns. It takes a couple of minutes to create the plot for the printer. It won't be redrawn on the screen. To cause the plot to print, you will probably have to go back to screen plot mode to force the print operation (next step), since the way sm works, switching output devices causes output to happen to the former device.
To go back to displaying the plot on the screen, switch device names and plot:
dev x11 vplot yymmdd.err
To create a PostScript file named "yymmdd.ps" of the plot: (case sensitive)
dev postscript :SY :OF=yymmdderr.ps (or :OF=yymmddrsd.ps) vplot yymmdd.err (or yymmdd.rsd)
You may need to return to screen plotting at this point in order to finally generate the PostScript file. It seems like with sm, nothing happens until you change devices.
To exit the sm program, type quit at the sm prompt.
The procedures for producing a surface plot of the pointing data set are currently being worked out. [Forget it, never done.]
The following ASCII files are used in the sky map and fitting programs:
Input file to skymap program.
Consists of a table of sky map coordinates. Entries are hour angle and declination in decimal hours and degrees. Organized as groups of declinations for each hour angle scan line. These sky map points are the targets for which the nearest FK5 star is found.
The hour angles are expanded by the factor cosine(declination) in order to eliminate crowding of sky map points near the pole, i.e., make the distribution of data points uniform over the celestial sphere.
Output file from skymap program.
Input file to skyfit program.
Contains records for each sky map data point consisting of the coordinate from the pattern file, the actual catalog star coordinates (normalized to 6h = 90d = 1.00), and the pointing error for each axis in arcseconds. The first field of each record is an identification number representing the line number in the pattern file. NOTE: A future version of skymap will produce a *.map file with star coordinates given in hours and degrees, not normalized.
Another output file from skymap program.
Input file to the TPOINT pointing data analysis program. [Not standard TPOINT format; the entire record should have been on one line. See Andrew Pickles for a converter utility.]
Contains records starting with an asterisk on the first line, the FK5 catalog number on the second line, the final (adjusted) apparent coordinates on the 3rd line, and the sidereal time as hours and decimal minutes (hh:mm.mmmm) on the 4th line. There are then 2 blank lines.
Output file from skyfit program.
Contains the hour angle and declination "C" coefficents and the set of "ALPHA" coefficients common to both axes.
This file is compilable into a c-language program. It currently is not used this way, but can be when the TCS is converted to c.
Note that this file is not transmitted to the TCS. The data it contains are transmitted in a simpler format at the same time this file is being generated.
The standard output from skymap contains the comment lines from the .pat file and the RA and declination coordinates of the selected FK5 star for each sky map data point. If you want to save this information, re-direct the standard output to a file with .dat extension with the > symbol at the end of the skymap command line. Note that you will then lose the current information on the screen of what the skymap program is doing, since the information will be going to a file instead. I recommend not creating a .dat file; it is much more useful to see the information on the screen in real time. Also, it may be possible to dump the displayed window contents to a file or the printer.
The standard output from skyfit contains the computed error for each of the sky map stars and the residuals, for each axis, where the residual is the difference between the measured and computed pointing error. I find that it is valuable to keep this data in a file rather than let it just go to the screen as standard output, so I re-direct the output to a file .out as so:
skyfit 940615.map > 940615.out (for example)
This catalog currently resides in my directory ~harwood/skymap/fk5 and /benchmark/skymap/fk5 and is used by the skymap program to look up candidate stars for the sky map. The FK5 catalog has been re-ordered in right ascension, rather than catalog number (as originally provided).
There is nothing the operator needs to know about this catalog, other than its existence.
Performs the sky map by selecting stars from the FK5 catalog corresponding to the input pattern file (.pat) and sending their coordinates to the TCS for slewing. Receives from the TCS the slew error at each star, which is inserted into the output file (.map).
skymap [-l <line#>] [-a] [-b] irtfWE.pat yymmdd.map [> yymmdd.dat]
Performs the sky map operation. Looks up stars in the catalog corresponding to the pattern coordinates for each sky map data point and sends the coordinates to the TCS as a slew request.
Accepts from the TCS the error in arcseconds on each axis.
Option -l <line#>
You can start at any specified ID number in the pattern file. This is useful if a sky map was stopped early due to clouds or something, and you want to resume in the pattern file where you left off.
Look at a printout of the pattern file to determine ID numbers. The ID number of a pattern entry is its line number from the start of the file, starting with 0.
If you are restarting a sky map, use the -a option to append the data so that your previous data are not overwritten (see below).
Option -a
Allows for appending pointing data to an already existing .map file (specified in the command). Use this option to continue from an interrupted sky map.
If the named file doesn't exist, an error message is displayed and the program exits. Check the spelling of the file name by listing all map files with the command: ls *.map.
Option -b
Causes the LSI-11/23 TCS computer to return the guide error for each sky map star instead of the absolute slew error. This is used in doing a "Benchmark" quick pointing check.
File irtfWE.pat
The input pattern file.
File yymmdd.map
The output map file, used as input to the skyfit program.
Reduces the sky map data set file obtained from skymap (yymmdd.map). Produces a file of coefficients (yymmdd.data) that can be transmitted to the TCS and also stored on the local computer. Also computes and outputs a table of residual errors for the input data set, which are the differences between the computed and measured pointing errors at each star, and the RMS residual error over the data set.
skyfit [-n] [-v] [-t] yymmdd.map] [yymmdd.data] [> yymmdd.out]
Performs the calculations for creating a set of polynomial coefficients from an input file of sky map pointing error.
Option -n
Causes the in-memory error data to be normalized about zero. Sometimes, for some reason the whole data set for one or both axes has a large bias component. This has no effect on the calulation of the polynomial coefficients, but distorts the data plots. Normalizing subtracts the average value for the position error from each data point, so there are as many negative errors as positive ones.
This normalization is done only on the in-memory data, but will show up on the output. The .map file is not affected.
Option -v
Verbose output (recommended). Displays to standard output a table of the sky map data points with the measured errors. Without the -v option, only the computed errors and residuals are displayed.
Option -t
Causes the computed polynomial coefficients to be transmitted to the TCS over the RS-232 communications link. The TCS must be running on-line with the current Forth TCS control system. The communications daemon must be running on the Sun computer. If a communications error occurs, the program tries two more times before giving up.
If option -t is accidentally left out of the command, simply repeat the command using it.
File *.map
The input file to the skyfit program. This file was generated by the skymap program. If not present, the program will look for the file "skymap.map".
File *.data
The output file from skyfit consisting of alpha and C polynomial coefficients. If not present, the program will write the file to "AandC.data".
File *.out
The output file from skyfit that is redirected from stdout (the screen). Consists of among other things sky map coordinates together with measured and computed errors and the RMS for both axes. It is recommended to specify this file rather than letting the output data scroll off the screen.