A data frame has been specified whose dimensions are orthogonal to those of an already processed frame for this reduction. The most likely cause is that one of the frames has been rotated and the other not. The required frame geometry is to have the orders long axis aligned with the horizontal (X-axis). The solution is to ensure that all frames are consistent and then re-start the reduction. In the case that the first frame specified needs to be rotated then the reduction process must be re-started from the beginning using a new reduction file (delete the current reduction file). In the case that that the first frame specified was OK, then simply correcting the other frame and continuing with the reduction will suffice.
A data frame has been specified whose X-dimension is different from that of the first frame specified during the reduction. All frames must have the same X- and Y-dimensions. The most likely cause is that the frames are from different datasets. If you really want to use this combination of data frames then you must use the facilities of fiGARO to ‘grow’ one of the data frames up to the dimensions of the other. Take care to ensure that any image registration problems are removed before reduction is attempted. In general it will be necessary to re-start the reduction with a new reduction file after any such adjustments have been made.
A data frame has been specified whose Y-dimension is different from that of the first frame specified during the reduction. All frames must have the same X- and Y-dimensions. The most likely cause is that the frames are from different datasets. If you really want to use this combination of data frames then you must use the facilities of fiGARO to ‘grow’ one of the data frames up to the dimensions of the other. Take care to ensure that any image registration problems are removed before reduction is attempted. In general it will be necessary to re-start the reduction with a new reduction file after any such adjustments have been made.
The program has run out of virtual memory space. The system command ‘vmstat’ produces a brief display of system memory status. In particular, note the value which appears under the ‘free’ column. A low value here indicates that the whole system is running low, otherwise it is likely that the problem is specific to your own limits. You may check these with the ‘limit’ command.
The program has been unable to release allocated memory correctly. The most likely cause of this is attempting to continue processing after having run out of virtual memory (Vmem_exhausted). The only solution is to abort the task and re-start.
An attempt has been made to access a reduction object using dimensions different from
those operative at the time of its creation. This will usually be due to the changing of a
tuning (TUNE_???) parameter which is used to specify the dimension, or due to a change in
wavelength scale during scrunching (which changes dimensioning objects NX_REBIN
and NO_OF_BINS). The name of the parameter is shown along with its ‘old’ and current
value.
If a tuning parameter is involved then there are two courses of action available. The first option is to reset the parameter to its previous value and continue with the reduction. The second option is to delete from the reduction file the object which has been created using the ‘old’ value. Such objects must then be re-created using the new parameter value by invoking the relevant processing options. The program will prompt as to which course of action is required. It should be noted that in some cases deleting an object may necessitate the re-running of an earlier option in order to re-create it, and re-calculate its contents. In this case you will be informed as to which step needs repeating.
The task has been unable to find any orders in the frame supplied. The most likely cause is that the frame has been incorrectly specified and contains no orders. Another possibility is that the frame has the wrong orientation (orders long axis horizontal) although at least one order would usually be ‘found’ in such a case. If the frame is correctly specified and orientated then possible causes are as follows:
Cosmic-ray noise, use the ech_decos1 task/ECHMENU Option 1.2 to clean the cosmic rays before re-trying order location.
Very poor S/N data, set the parameter TUNE_AUTLOC=N to specify the positions of the orders manually using a graphics cursor.
The program is unable to determine the slope of the orders across the frame. In this situation the slope
is assumed to be 0. The most likely cause is problems caused by cosmic-ray contamination which can
be avoided by using the ech_decos1 task/ECHMENU Option 1.2 to clean the frame prior to slope
determination.
The slope value is only used when tracing very poor data and it will not usually cause problems if its value is left at zero.
In extreme cases you can set the slope manually using the fiGARO command SETOBJ:
where nn.nn is the increase in y coordinate per X-pixel increment.
The automated polynomial fit in wavelength has generated a wavelength range which is inconsistent with the expected scope of wavelengths (either as specified, or as calculated from previously identified orders). You should investigate the identification of this order manually to verify the situation. In most cases allowing a larger search scope for the order and/or manually rejecting the incorrect identification using the AI option in ech_idwave will lead to a more favorable identification.
The fitted polynomial has been rejected due to an excessive RMS deviation measure. Usually the rejection will be valid and the program will proceed to locate a better set of identifications. However, if looking at the other orders shows that these identifications are likely to be correct, then ech_idwave should be used in its manual mode of operation. Specifying an appropriate search range for the order and using the AI option will let the program relocate the set of identifications it rejected when running automatically. You can then accept the set and interactively improve it by adding/removing individual line identifications.
The program is unable to create an arc line database file. This is usually due to either the destination directory being write-protected, or a disk quota being exceeded.
echomop uses a special arc line information database to optimise its search algorithm. This algorithm
requires sets of approx. 12 lines with which to work. The database consists of lines and their 10 nearest
neighbours on each side (l/r). The arc-line list supplied (.ARC file) does not contain sufficient lines for
the database to be useful. Either more lines must be added to the list, or another program
(such as ECHARC) must be used for the line identification phase. It is possible to add
dummy entries to the line list just to overcome this limitation, however in this case it is
unlikely that automatic identification will work and the program will probably need to be run
interactively.
The program failed to determine the width of the order traces. The algorithm for this process requires
that the intensity falls to less than the fraction specified by TUNE_TWTHR of the intensity at the peak. You
may either adjust the parameter (probably increasing it), or directly specify a value for the trace width
using fiGARO SETOBJ:
where nn is the approximate width in pixels.
A ‘read’ operation on the arc line database has failed. The most likely cause is that an inadvertent database creation has been performed which has resulted in an incomplete copy of the database. Check the database files using
and delete/rename any partial files.
The program was unable to obtain a satisfactory estimate of the position of the ‘sky’ in the region
below the object. The algorithm for determining the sky extent uses the median intensity across the
profile, and the relative peak intensity. The relative peak intensity is multiplied by the value of the
parameter TUNE_SKYHILIM and compared with 1.05 times the median intensity. The lower of the two
values is used as a threshold. When the intensity falls below this threshold, then the program denotes
the pixels from the current pixel to the edge of the dekker as ‘sky’. You may vary TUNE_SKYHILIM to
adjust the threshold used. It is also possible to manually adjust the status of each pixel in the profile by
setting PFL_INTERACT=Y.
The program was unable to obtain a satisfactory estimate of the position of the ‘sky’ in the region
above the object. The algorithm for determining the sky extent uses the median intensity across the
profile, and the relative peak intensity. The relative peak intensity is multiplied by the value of the
parameter TUNE_SKYHILIM and compared with 1.05 times the median intensity. The lower of the two
values is used as a threshold. When the intensity falls below this threshold, then the program denotes
the pixels from the current pixel to the edge of the dekker as ‘sky’. You may vary TUNE_SKYHILIM to
adjust the threshold used. It is also possible to manually adjust the status of each pixel in the profile by
setting PFL_INTERACT=Y.
The program was unable to obtain a satisfactory estimate of the position of the dekker in the region below the trace. The algorithm for determining the dekker position steps out from the peak intensity until the intensity falls below a tunable threshold. If the predicted edge is reached before this occurs then the lower dekker is set to this predicted value (simply calculated as half-way between the orders).
You may vary TUNE_DEKTHR to adjust the threshold used. It is also possible to manually adjust the
status of each pixel in the profile by setting PFL_INTERACT=Y.
The program was unable to obtain a satisfactory estimate of the position of the dekker in the region above the trace. The algorithm for determining the dekker position steps out from the peak intensity until the intensity falls below a tunable threshold. If the predicted edge is reached before this occurs then the lower dekker is set to this predicted value (simply calculated as half-way between the orders).
You may vary TUNE_DEKTHR to adjust the threshold used. It is also possible to manually adjust the
status of each pixel in the profile by setting PFL_INTERACT=Y.
The program has failed to trace the order outwards from the centre of the frame towards the left-hand side (decreasing X). If tracing is being attempted using a cosmic-ray-contaminated frame then the ech_decos1 task/ECHMENU Option 1.2 should be used to clean the frame prior to tracing.
The parameter TUNE_MXBADSMP determines the number of failed increments before the trace is ‘lost’.
The parameter TUNE_MXSMP determines the number of sampled increments across the frame and thus
their separation. In general increasing TUNE_MXSMP will assist the program in tracing difficult orders.
You should also experiment with the various modes of tracing (controlled by TRACE_MODE parameter)
which should allow most cases to be traced. As a last resort the program also provides a manual trace
mode in which points specified using a graphics cursor (ICUR program) can be used as the basis for
the tracing process.
The program has failed to trace the order outwards from the centre of the frame towards the right-hand side (increasing X). If tracing is being attempted using a cosmic-ray-contaminated frame then the ech_decos1 task/ECHMENU Option 1.2 should be used to clean the frame prior to tracing.
The parameter TUNE_MXBADSMP determines the number of failed increments before the trace is ‘lost’.
The parameter TUNE_MXSMP determines the number of sampled increments across the frame and thus
their separation. In general increasing TUNE_MXSMP will assist the program in tracing difficult orders.
You should also experiment with the various modes of tracing (controlled by TRACE_MODE parameter)
which should allow most cases to be traced. As a last resort the program also provides a manual trace
mode in which points specified using a graphics cursor (ICUR program) can be used as the basis for
the tracing process.
The program could not trace the order at all. Check that:
If tracing has to be attempted using a cosmic-ray-contaminated frame, try using the ech_decos1 task/ECHMENU Option 1.2 to clean the frame prior to tracing.
The parameter TUNE_MXSMP determines the number of sampled increments across the frame and thus
their separation. In general, increasing TUNE_MXSMP will assist the program in tracing difficult
orders. It is worth experimenting with the various modes of tracing (set by the TRACE_MODE
parameter). Usually, at least one of the available options will handle difficult or unusual
cases.
As a last resort a manual-trace mode, in which points specified using a graphics cursor (ICUR program) are used as the basis for the tracing process, is available.
The reduction database can not be created at this stage. You must either use ECHMENU Option 1, or task ech_locate, as the first operation. These ensure that the reduction database used by all other task/operations is correctly created and initialised with the data frame dimension information.
The specified file cannot be found and its automatic creation is illegal using the selected option.
Please use the special task ech_ftrdb to perform arc line database creation. This process should
normally be left to the node manager who should place the database files in a commonly accessible
directory (usually the same place as the .ARC files). If you wish to create a private database for your
personally specified line list then use task ech_ftrdb to do so. If you are not trying to create a private
database then the most likely problem is that the environment-variable ARCDIRS has not been used
to prefix the arc database name (e.g., ARCDIRS:THAR), or that the named arc database does not exist
(ask the system manager).
If you wish to include a set of personally specified line list databases in the default search paths used by ECHOMOP, then their location should be added to the list held in environment-variable ARCDIRS. The list of locations is searched in order of their appearance in the environment-variable. The default search path is:
. — the current directory.
ADAM_USER — ADAM user directory.
ECHOMOP_EXEC — echomop base directory.
The specified file exists but is not a reduction database as defined by echomop tasks. The most probable cause is that a data frame file name has been provided by mistake.
The reduction database must be created before the reduction can start. The task ech_locate/ ECHMENU Option 1 is currently the only one authorised to invoke reduction file creation so you should use this if you need a new reduction database.
The program is unable to obtain update access to an array. This may be either a DATA, QUALITY or ERROR array; and the most most probable cause is that the file containing the array is protected against write access in some way. If the file is not owned by you, and in your directory, then it is usually easiest to copy it into a directory to which you have full access rights.
The program is unable to create a reduction file as specified. The most probable cause is due to the directory where the file creation is to be performed being protected against write access, or does not exist at all. Type:
to verify that the file exists, and to see what access rights the current directory has.
If you own the directory in which file creation was attempted and have the required access rights then the most probable cause is the exhaustion of system resources required, most notably disk space. Type:
to see if you have run out of space.
The program is unable to copy reduction database objects from the specified reduction database. The
possible causes are that the object in question does not exist in the reduction file specified (either using
TUNE_CLONE or the < indirection operator in ECHMENU), or that the object is dimensioned differently
from the corresponding object in the primary reduction database. Objects can only be copied when
their dimensionalities are identical (i.e. they have the same number of X-pixels, number of orders
etc.).
To check the dimensions of the object you are trying to clone from, use the fiGARO EXAM
command and the object path as reported by the program. If the only dimensions which
differ are controlled by TUNE_ parameters, then you may adjust the relevant parameter to
allow the object to be successfully copied (Note that this in turn may necessitate re-running
parts of the reduction to re-generate other objects also dimensioned using the same TUNE_
parameter).
The program is unable to obtain UPDATE access to an object in the reduction database. This is probably due to the file being protected against write access. Type:
to check the file protection mask.
The program will continue but please note that it will be unable to write the results of the current step back into the reduction database. You are therefore strongly recommended to copy the file to a location where you have full access to it before proceeding.