Release Notes for Sked version 2022-03-28
John Gipson

Most of the recent changes to sked have been in the nature of bug fixes or enhancements that do  not affect the overall user experience. This note summarizes some important updates to sked. For more detail look at the file change_log.txt in the sked and skdrut directories.

UNIFIED SOURCE CODE for SKDRUT

Sked and drudg use many of the same utilities. These are contained in the skdrut directory and skdrut.a library.  Because the development for sked and drudg is done on different machines these libraries diverge over time.  The current version of skdrut is that which is used in the FS10.1.

NEW SLEW MODELS GGAO12M and ONSALA TWINS

Sked uses a simple slewing model which assumes constant acceleration upto the terminal velocity, and then constant deceleration when the antenna stops. This is sufficient for most antennas. 

It failed at GGAO12M where there is a large blockage due to the SLR antenna.  The niave slewing model assumed the antenna traveled right throught the forbidden region. The FS knew about the forbidden region and would cause the antenna to travel around it. This resulted in some observations being late on source. The subroutine slew_ggao.f calculates the correct slewing for GGAO. This has been used to schedule VGOS sessions since ~June 2021.

The motion of the Onsala Twin telescopes is also more complicated. The velocity in az and el depends both on the direction the antenna is moving and the region. The antenna slows down as it reaches the end of the limits.  Eskil provied a table of these values and I coded  themup in a rouinte clalled slew_oe_ow.f90.

FASTER SCHEDULING, LESS OUTPUT
There were several items that sked would calculate every time it needed them.  I modified sked to compute many of these only once for each session. The next time it needs this information it would find it in a lookup table.

I also modified sked to reduce the output during routine scheduling. The amount of sked output can be controlled by the parameter VERBOSE_LEVEL. This number varies between 0 and 10. For most purposes a VERBOSE_LEVEL of 0 is appropriate. The higher this number, the more information is displayed.

NEW EXPERIMENTAL RANDOM COMMAND
There is a new command 'random' which will insert one or more random scans.  It uses the same syntax as the "automatic" scheduling command. It will select a random scan from the scans that pass the major options.  An example of the syntax
?  random
will insert one random scan at the current place
?  random _ 2330  will insert random scans until we get to 2330

You can use the RANDOM command to randomize the starting point of schedules.

NEW MAJOR OPTION. EL_NOISE
There is a new majore option EL_NOISE.  (This option is similar add_ps.) This only makes sense if you are using covariance optimzation. It will add elevation dependent noise to all of the observation using the formula: 
     sigma^2=sigma_calc^2+  (el_noise/sin(el1))^2+ (el_noise/sin(el2))^2
sigma_calc is the calculated sigma based on the SEFDS and the source flux models. 

EASIER TO ADD NEW CORRELATORS
The list of valid correlators was hard coded in sked. It still is, but the correlators are now kept in an include file in the sked directory called 'valid_correlator.ftni'. If your favorite correlator is not in this file free to add it and recomple. 


As usual, if you find any bugs please let me know and I will endeavor to fix.

John 





