AIPS HELP file for CONFI in 31DEC24
As of Thu Apr 25 4:42:44 2024
CONFI: Optimize array configuration by minimum side lobes
INPUTS
INNAME UV file name.
INCLASS Input UV file name (class)
INSEQ 0.0 9999.0 Input UV file name (seq. #)
INDISK Input UV file disk unit #
The UV file is needed
only to attach PL table
INFILE Filename of the data
of initial configuration.
Blank => the initial
configuration is calculated
IN2FILE Filename of the topography
mask.
Include character sequence
'FITS' in the file name
if the IN2FILE is in
FITS format.
Used also as fixed positions
of the two and three circle
configurations (see HELP)
CELL Size of the cell at mask,
meters (x, y); 0 => 10m
XPARM BLC and TRC of selected part
of the mask relatively
the initial mask
1. X of the bottom left
corner of the mask,
in meters
2. Y of the bottom left
corner of the mask,
in meters
3. X of the top right
corner of the mask,
in meters
4. Y of the top right
corner of the mask,
in meters
5. ALPHA the degree to
(1/DIST) at the
dependence of the gain
on the distance
ALPHA > 0 (see explain
on APARM(2))
0 => 1
OUTFILE Filename of the final
configuration
'' => no record a file
NPLOTS Number of plots of beam's
pattern at the page.
0 => 20.
DOTV -1.0 1.0 > 0 Do plot on the TV, else
make a plot file
APARM 1. Number of iterations.
0 => 1
2. gain in the iterat. loop
0 => 0.001
<0 => GAIN = ABS(APARM(2)*
(1/DIST)**ALPHA
ALPHA=XPARM(5) > 0
3. Number of elements at the
array. 0 => 64
4. Outer radius of optimizat.
circle at the sky
in main lobe size; 0=>20
5. Step of the beam calculat.
in main lobe size 0 => 0.2
6. 0 => display configuration
1 => not display it;
7. Size of the array in km
8. 0 => The configuration had
been shifted
1 => The configuration had
not been shifted
9. Inner radius of optimizat.
circle at the sky in main
lobe size; 0 => 1.2
10.Sector of optimization
in degrees. < 180 deg.
0 => 180 deg.
BPARM 1. not used
2. Total number of plots of
the beam pattern; 0 => 20
3. 0 => auto scale beam plots
otherwise: Min = 0
Max = BPARM(3)
4. X-shift of configuration
relatively BLC of mask, m
5. Y-shift of configuration
relatively BLC of mask, m
6. Turn of configuration
clockwise, degrees
7. Number of steps in topogr.
fitting, X, 0 => 50
8. Number of steps in topogr.
fitting, Y, 0 => 50
9. Number of steps in topogr.
fitting, Rotation,
0 => 360/STEP
10. Step in topography
fitting of rot., degrees
0 => 5
see explain of BPARM(4-10)
CPARM Data to prepare the output
file at UVCON format
1. 0 => CONFI format
1 => UVCON format
The following CPARMs used
only if CPARM(1) = 1
2. The site geodetic
latitude, degrees
3. The site geodetic
longitude, degrees
4. Antenna diameter, meters
0 => 12
5. Ant. efficiency, 0 => 0.5
6. Antenna system temperature
0 => 50K
7. Number of levels of
digitizer; 0 => 2 level
8. 0 => no beam correction
by the primary beam
1 => yes beam correction
by the primary beam
The diameter of the
antenna should be
given at CPARM(4)
9. The site height above
the geoid, in meters
DPARM 1. Initial configuration:
0 => homogeneous
circular array
2 => two circles
3 => three circles
4 => hexagon tile
2. Number of fixed antennas
3. Constraints:
0 => no constraint
1 => constraint of topogr.
2 => constraint of 2 circ.
3 => constraint of donuts.
4 => constraint topography
and donuts
5 => constraint topography
and 2 circus
6 => inside of elliptic
outer border;see help
4. 0 => input configuration
is normalized to 1
1 => input configuration
is in meters
5. 0 => Yes side lobes optim.
No topography fit
1 => No side lobes optim.
Yes topography fit.
2 => Yes side lobes optim.
Yes topography fit.
number of steps at
fitting to topography
BPARM(7,8,9) are
forced to 1
6. Output file mode:
0 => normalized
1 => in meters
7. Minimum distance between
elements, meters
8. ratio of the ellipse axis;
for elliptic outer border
if DPARM(3)=6; see help
Inner radius of donut, in
the outer diameter parts.
radius of the 2nd circle
if DPARM(1) =2, or 3.
>0 => the circle is
concentric
<0 => the circle is inside
tangent to the outer
circle.
9. radius of the third
(innerest) circle
10. number of additional
elements at the outer
circle. see explain.
HELP SECTION
CONFI
Type: Task
Use: This task optimizes the array configuration minimizing side lobes
Adverbs:
INNAME.....Input UV file name (name). Standard defaults.
INCLASS....Input UV file name (class). Standard defaults.
INSEQ......Input UV file name (seq. #). 0 => highest.
INDISK.....Disk drive # of input UV file. 0 => any.
The input UV file actually does not participate
in the calculating process. It is used only if DOTV<0
as a file to attach the plot file. It is a good idea
to use the task TASAV to create the UV file with only
one visibility.
INFILE.....The name of a file where the initial array configuration
is stored.
X and Y of the array elements are given in decimal form.
Twenty characters field is reserved for X and Y.
Maximum 9(10) characters are given for the data part before
decimal point and maximum nine character are given for
the data part after decimal point (F20.9).
For example:
1234567890.123456789 -123456789.123456789
-23.65 123.74
2345789.564327 12345.896
The file has not to have empty lines.
If INFILE = '', then initial configuration is calculated
as a homogeneous configuration on the circumference;
as a configuration on the circumference with progressively
increasing of the spacing;
as a homogeneous configuration on the two circumferences; or
as a homogeneous configuration on the three circumference;
The calculated configuration is selected under control of
DPARM(2).
IN2FILE....Filename of the mask - restricted positions
If the IN2FILE includes character sequence 'FITS' then
the IN2FILE is interpreted as FITS format file
else the file is interpreted as ASCII format file:
Three columns at the file:
1. Pixel number at X;
2. Pixel number at Y;
Y changes faster than X (Scan by Y for each new X)
3. IND, restricted (IND=1) or not restricted (IND=0);
Position of the given pixel in meters is determined
by the two dimensional array CELL
X, Y, IND should be located at one line and
given at the free format
^^^^^^^^^^^
If topography constraint are not used (DPARM(5)=0) then
this file is used as positions of fixed elements
at the outer circle when two and three circle configurations
are used (DPARM(1)=2 or =3). INFILE is not used in this case.
The position of the fixed elements at the outer circle are taken
from IN2FILE (in meters). FORMAT is 2F20.9
The rest positions in the outer circle and inner circles
fill evenly the spaces between the fixed elements at the
outer circle and in the inner circles.
This is useful when the inner circle of A-configuration
is the outer circle of B-configuration and so on.
Then the pads of the A-configuration's inner circle
can be used as part of the B-configuration's outer
circle. So during optimization of B-configuration's
these positions should be fixed.
Positions of the fixed elements should be given in
normalized form. If the bigger configuration is given in
meters it has to be normalized:
Run CONFI with the following parameters to provide
the normalization: INFILE - the bigger configuration
in meters. APARM(7) diameter of the inner circle
of the bigger configuration (or that the same
the outer circle of the smaller configuration).
DPARM(4) = 1.
APARM(1) = 1. OUTFILE - the bigger configuration
in normalized form.
Now extract the outer circle from the found configuration.
It is the fixed part of the smaller configuration's
optimization. Use it with DPARM(4) = 0.
If you want to have fixed element in general case including
topography, you have to:
1. Use INFILE as input configuration.
2. Consider the first DPARM(2) lines of INFILE
as the fixed antennas.
3. Use IN2FILE as the topography mask if you want
to include the topography constraint (DPARM(5).NE.0)
CELL.......Size of the cell at mask in X and Y, meters; 0 => 10 meter.
XPARM......BLC and TRC of the selected portion of the mask relatively
the initial mask
1. X of the bottom left corner of the mask, in meters
2. Y of the bottom left corner of the mask, in meters
3. X of the top right corner of the mask, in meters
4. Y of the top right corner of the mask, in meters
5. ALPHA the degree to (1/DIST) at the dependence of
the gain on the distance of the beam maximum from
the beam center.
ALPHA > 0 (see explain on APARM(2))
0 => 1
OUTFILE....Filename of the final configuration. The configuration
corresponding to the lowest found side lobes is
recorded to the file. The format of the file is
identical with the format of the INFILE.
So the OUTFILE can be used as INFILE for the following
optimization
Blank => no record to a file
NPLOTS.....Number of plots of the beam pattern at the page. 0 => 20.
DOTV.......> 0 Do plot on the TV, else make a plot file
APARM .....1. Number of iterations. 0 => 1
2. gain in the iteration loop 0 => 0.001
The required gain depends on the DIST - distance of the
found beam maximum from the beam center.
The more distance the less gain.
Using negative APARM(2) you can introduce this dependence
GAIN = ABS(APARM(2)) * (1/DIST)**ALPHA
ALPHA=XPARM(5) > 0
GAIN = ABS(APARM(2)) for DIST=1 for any ALPHA
The recomended ALPHA is between 0.5 and 1.0
3. Number of elements at the array
Used if INFILE = BLANK. 0 => 64
4. Radius of optimization circle at the sky, measured
in main lobe size; 0=>20
5. Step of the beam measured in main lobe size. 0 => 0.2
6. 0 => display configuration; 1 => not display
7. Size of the array in km; 0 => 1
----------------------------------------------
|The task uses the -SIN projection of the sky |
| So the value of one corresponds to THETA=90 |
| The following inequality should be valid |
| to exclude ABS(SIN(THETA)) > 1 |
| APARM(4) * LAMBDA/APARM(7) < 1 |
| LAMBDA is the wavelength |
| (LAMBDA and APARM(7) at the same units) |
----------------------------------------------
8. 0 => The configuration had been shifted.
The shifted configuration is created during
fitting in topography.
1 => The configuration had not been shifted.
APARM(8) does not have a sense if configuration
is given in the normalized form.
9. Inner radius of optimization circle at the sky
in main lobe size; Beam maximum is found as maximum
outside of circle radius of APARM(9).
To exclude main lobe and include nearest side lobes,
put APARM(9)=0 => =1.2
10. Sector of optimization in degrees. < 180 deg.
0 => 180 deg.
BPARM....... 1. not used
2. Total number of plots of the beam pattern. 0 => 20
Each plot represents one cross section of the two
dimensional beam at the range (-APARM(4), +APARM(4)).
The maximum total number is APARM(4) / APARM(6).
BPARM(2) allows to plot only some part of the all
available cross sections of the two dimensional beam.
3. 0 => auto scale of the beam pattern plots
otherwise: Min = 0; Max = BPARM(3)
4. X-shift of configuration relatively BLC of mask, m
5. Y-shift of configuration relatively BLC of mask, m
6. Turn of configuration clockwise, degrees
7. Number of steps in topography fitting, X, 0 => 50
8. Number of steps in topography fitting, Y, 0 => 50
9. Number of steps in topography fitting, Rotation,
0 => 360/STEP
!!!!! BPARM(7,8,9) are used onlyduring initial search
of the configuration to fit it to the topography
Use BPARM(7,8,9)=1 when you optimize the
configuration with topography constrain.
10. Step in topography fitting of rot., degrees 0 => 5
Step's value in X and Y is determined by CELL
Look below the explanation of BPARM(4-10)
CPARM.......Data to prepare the output file at UVCON format
Format: CONFI/UVCON?
1. 0 => CONFI format;
1 => UVCON format.
The following CPARMs used only if CPARM(1) = 1
2. The site geodetic latitude in degrees
3. The site geodetic longitude in degrees
4. Antenna diameter, meters; 0 => 12
5. Antenna efficiency; 0 => 0.5
6. Antenna system temperature; 0 => 50K
7. Number of levels of digitizer; 0 => 2 level
8. make the beam correction by the primary beam?
0 => no
1 => yes
The diameter of the antenna should be given at CPARM(4)
9. The site height above the geoid, in meters
DPARM.......1. Initial configuration:
0 => The elements are distributed homogeneously
on the circumference
2 => The elements are distributed homogeneously
on two circumferences.
The radius of the outer circumference is 0.5;
The radius of the inner circumference is DPARM(8);
The number of elements on the circumferences
is proportional its radius.
See DPARM(10) for difference.
3 => The elements are distributed homogeneously
on three circumferences.
The radius of the outer circumference is 0.5;
The radius of the intermediate circumference
is DPARM(8);
The radius of the inner circumference is DPARM(9);
The number of elements on the circumferences
is proportional its radius.
See DPARM(10) for difference.
4 => hexagon tile. The number of the array antennas
given at APARM(3) should be equal:
7,19,37,61..= 3*NC^2 + 3*NC + 1; NC = 1,2,3...
2. The first DPARM(2) antennas are fixed duaring optimization
3. Constraints:
0 => no constraint of the elements location
In the case of a constraint initial configuration has
to satisfy to the constrains: the elements should be
in good topography points or inside the donuts...
1 => constraint of topography
2 => constraint of two circles.
3 => constraint of donuts. The initial configuration
may be two circles also (DPARM(1) =2)
-----------------------------------------------
|USE THIS CONSTRAINT WITH DPARM(8)=0, WHEN YOU |
|WANT TO HAVE ALL ANTENNAS INSIDE OF THE CIRCLE |
|DIAMETER=ARSISE=APARM(7) |
-----------------------------------------------
4 => constraint both topography and donuts
5 => constraint both topography and 2 circus
6 => inside of elliptic outer border
-----------------------------------------------
|USE THIS CONSTRAINT , WHEN YOU WANT TO |
|HAVE ALL ANTENNAS INSIDE OF ELLIPTIC AREA: |
| MAJOR AXES=ARSISE=APARM(7) |
| MINOR AXES=DPARM(8)*ARSIZE |
-----------------------------------------------
4. 0 => input configuration is normalized to 1
1 => input configuration is in meters
5. 0 => Yes side lobes optimization. No topography fit.
1 => No side lobes optimization. Yes topography fit.
2 => Yes side lobes optimization. Yes topography fit.
In this case number of steps at the searching
of configuration fitting to topography is
forced to 1
6. 0 => record normalized configuration in output file
1 => record configuration in meters.
7. Minimum allowed distance between elements, meters.
Initial configuration should satisfy the given minimum
spacing. It is not simple in the case of some antennas
being fixed. The routine OUTER provides insertion
of the rest antennas in the outer circle. It inserts
number of element between the neighbor fixed elements
proportional to the relevant arc length. Such an algorithm
does not follow the minimum spacing condition. Of course,
it is simple to satisfy the minimum spacing condition
using the step .GE. the given spacing. But in this case
it can happen that all inserted elements will appear at
the only interval what is not good. For small size array
the routine outer add one element to the interval
where the minimum spacing allows. If still there is a
problem of minimum spacing, then it is necessary to
diminish number of fixed elements and number of elements
at the outer circle (DPARM(10)). It is useful to plot the
initial configuration with very small spacing (DPARM(7)=1,
APARM(1)=1)
8. Inner radius of the donuts, in parts of the outer diameter
If DPARM(1)=2 or 3, it is the radius of the second
circle, starting from the outer circle.
>0 => the circle is concentric;
<0 => the circle is inside tangent to the outer circle.
Ratio of the ellipse axes of the elliptic outer border,
If DPARM(3)=6. (horizontal:vertical)
The ellipse vertical axis is equal 1 (ARSIZE=APARM(7))
If the ratio is <1, then the major axis is vertical;
If the ratio is >1, then the major axis is horizontal;
9. Radius of the third (most inner) circle, if DPARM(1)=3
10. number of additional elements at the outer circle.
Used if DPARM(1) = 2,3
Useful for D MMA's configuration.
It can be both positive or negative.
The number of elements in the circumferences are
calculated proportional to its radius.
Then the number in the outer circumference is
increased by DPARM(10);
For the three circle configuration (DPARM(1)=3)
the number of elements in the intermediate
circumference is decreased by (DPARM(10) + 1), and
the additional element is added at the center.
For the two circle configuration (DPARM(1)=2)
the number of elements in the inner circumference
is decreased by DPARM(10)
EXPLAIN SECTION
CONFI: Optimizes the array configuration minimizing side lobes.
Documenters: L.R. Kogan
The task finds the array configuration which has the minimum worst
side lobes at the given circular area at the sky.
The initial configuration is taken from the file 'INFILE'. The data in
the file is arranged as two columns (2F20.10) for X and Y. They
can be normalized to unit maximum baseline or given in meters.
It is anticipated by DPARM(4).
The number of the array elements is determined by the reading
end of file. That is why the empty lines are not allowed in the file.
If INFILE = '', then initial configuration is calculated
as a homogeneous configuration on the circumference;
as a configuration on the circumference with progressively
increasing of the spacing;
as a homogeneous configuration on the two circumferences; or
as a homogeneous configuration on the three circumference;
The calculated configuration is selected under control of
DPARM(2).
The last iteration is not necessary is the best one. That is why
the best configuration is stored and recorded to the output file
'OUTFILE' in normalized form (DPARM(6)=0) or in meters (DPARM(6) = 1).
If you want to plot(display) the found best configuration and relevant
beam's crossections, you need repeating the task with one iteration
(APARM(1)=1) and INFILE selected as the output file in the previous
iteration process.
The configuration is plotted (displayed) by "diamonds" and the relevant
UV coverage by dots.
The process of finding the configuration is an iterative process.
At each iteration the position and value of the worst side lobe in the
given sky area is found. Then a small correction of each element position
is provided to minimize the value of the worst side lobe.
Side lobes' value both before and after optimization for each iteration
are printed for each iteration.
The worst side lobe for the given iteration will be suppressed,
but you'll see another side lobe at different position, which is definitely
bigger than minimized value of the previous iteration. Usually the value
of the new worst side lobe is less than this value of the previous iteration.
But sometimes it is not true especially if you select the gain (APARM(2))
two big.
The value and the direction of the correction is calculated in accordance
of the theory (L. Kogan, MMA memo # 171 ). The value is determined by the
product of the calculated shift and the gain (APARM(2)).
The value of the gain should be found empirically but has to be rather small.
If it is two small the task run too long. If it is too big the found new
worst side lobes will jump from iteration to iteration and convergence can
be missed. The compromise should be found empirically. For example if
the array elements number = 36, 0.0001 < APARM(2) < 0.01.
FITTING TO THE TOPOGRAPHY.
First of all the topography file IN2FILE should exist. The file have to
have the three columns:
1. Pixel number at X; I4
2. Pixel number at Y; I5
3. Restricted (0) or not restricted (1); I4
The spacing between near pixels is determined
by the two dimensional array CELL in meters.
The fitting consists of the three stage:
STAGE 1. Shifting and rotation of the configuration having found
without constraints.
It is clear that the side lobes property in the given circular area
does not change if the configuration is shifted or rotated.
The initial guess for the shifting/rotation is determined by BPARM(4-6).
Steps in shifting is determined by CELL(2). Step in rotation is BPARM(10).
Number of steps in the searching is BPARM(7-9).
APARM(7) = the array's size.
You need to fix DPARM(3)=1; DPARM(5)=1 to invoke this option.
Use as INFILE the configuration found in a optimization without constraints.
The task provides the output file which includes X,Y coordinates
of the best fitted configuration.
If you are lucky to fit all elements of the array the following message
will be added at the top of the file:
"Configuration fitted topography is found"
If not, the number of unfitted elements and their coordinates will be added
at the top of the file.
The plot of the topography together with the best fitted configuration
is provided also.
To go to the stage #2, edit the output file shifting manually the unfitted
element to the nearest good topography place; do not forget to remove the
top messages.
Now having had the configuration which fits the configuration and
reasonably differ from the original optimized configuration you are
ready to optimize the configuration with the topography constraints.
STAGE #2 Optimization with the topography constraint.
Use as INFILE the edited OUTFILE found in stage #1.
BPARM(3), BPARM(4) should be equaled to the the found shifts.
BPARM(5) = 0. APARM(7) = the array's size. (the same as in stage #1)
DPARM(3) = 1; DPARM(5) = 2.
The greeting message:
"Configuration fitted topography is found" is appearing at the beginning,
because you started with the fitted configuration.
The process of the optimization and outputs will be identical to the
optimization under other constraints.
Stage #3. Check the result.
Use the OUTFILE of the stage #2 as INFILE.
DPARM(3) = 1; DPARM(5) = 1.
BPARM(3), BPARM(4) should be equaled to the the found shifts.
BPARM(5) = 0. APARM(7) = the array's size. (the same as in stage #2).
The task provides the plot of the topography with the found fitted and
optimized configuration.
All questions send to Leonid Kogan: lkogan@aoc.nrao.edu