Moderator of this Wiki page: Juan Penalver [JP]
- Revisions of Antenna Mount Drive (antMD) software
- How to switch to fallback version
- Test Version 2011-01-04
- Development Version 2010-09-03
- Current Version 2008-10-24
- Intermediate Revision 2008-05-31
- Previous Revision 2008-05-07
- Previous Revision 2007-06-15
- Previous Revision 2007-04-26
- Revision 2007-03-16
- Revision: 2006-09-25
- Revision: 2006-07-11
- Revision: 2006-06-29
- Revision: 2006-06-06
- Revision: 2006-05-30
- Revision: 2006-04-05
- 2005, November to 2006-03-06
Revisions of Antenna Mount Drive (antMD) software
How to switch to fallback version
- We should not switch back to a version before 20070316
- to change versions of antenna MD software
- telnet as "root" to vac1
ln -nsf /home/master/antenna.20080507 antenna
- to point to the "current" version at that time
ln -nsf /home/master/antenna.20070615 antenna
- to point to any different (fallback) version
NOTE: To be sure that there is no version mixup, these commands should only be entered after a reboot of vac1 and before loading any antenna control software.
Test Version 2011-01-04
- This is a copy of the Version 2010-09-03. Test carried out with this version are:
- Modification of the pointingCorrection.c to get the display of the refraction correction with dmpObservation (JP, with AP)
Development Version 2010-09-03
- This Version started as a copy of the Current Version 2008-10-24 when getting operative in 2010-09-03.
- 8-Sep-2010. The stateFlags of the antenna slow trace is set to 1 each time the antenna is pointing to a distance smaller than 1.14 deg from the sun avoidance circle. It is re-calculated each time it is sent at the defined slow loop rate. A stateFlags constantly set to 1 indicates a target inside the sun avoidance circle or a drive problem to solve rapidly (AP). Pending to be tested (JP)
- The debug variables have been changed by AP on 22-Mar-2011
Current Version 2008-10-24
antenna -> /home/master/antenna.20081024/
- Current Version since 3-Sep-2010. The major changes are summarized below
tests of this version and its integration into the NCS are ongoing
- The debug variables have been changed by AP on 5-Apr-2011.
- Slow Rate adjustable to 1, 2, 4 or 8 (previous versions had only the option 1). Use the utility setSlowRate. The slow rate is now written to the antenna datastreams and then to the IMBFits files, in order to make the necessary adjustments to the data when switching between slow rates.
- Correction of the acceleration estimation in the transition trajectory calculation in onePPS.c 2008-09-10.
- onePPS.c renamed itSlow.c.
- Interpolation of the reference position in preset mode as it is done in tracking mode. This change avoids an azimuth position behavior in "saw teeth" when the tracking was not permitted at high elevation (2008-10-20).
- The NCS-message send every slow rate has been extended to include the fast and slow rates and max., min and RMS of the 'tracking errors'.
- Implemented Lissajous trajectories. Use the utility setNextSegmentLissajous.
Suppressed the small oscillations, below 1 arcsec amplitude, at low elevation speed (servo loop parameters changed el.cXKp (2560 -> 3075) and el.cXKi (4 ->6). Implemented on 6-Sep-2008. Applied in 20080507 version too (CK: this was implemented in the current version in 12/08.)
- All the variables like dUT, xp, yp, tdk, pmb, rh are now read from the PICO shared memory area in the endless loop of slaParams, i.e. every 10s. Any change is right away applied. Previously, the dUT value was read only once, at slaParams task start, and was corrected by the pointing model parameters. Correction in 20080507 version too.
- Used the astronomical coordinates at the antenna azimuth axis with an estimated error below 1 meter. Previously we used the astronomical coordinates of the geodesic vertex 19.55 meters out the azimuth axis. Implemented on 5-May-2009. Applied in 20080507 version too
- Implemented alarm if too high tracking error in tracking mode and after backOnTrack. Use the utility setXDelta2Alarm (2008-09-06).
- successfully tested with pointing scans up to 88.1 deg. elevation after changed the parameter xDeltaMax from 150 to 600 arcsec. Tested on 9-Sep-2009. To get smooth transition trajectories for position changes below xDeltaMax, the slowRate should also be increased to (4 or) 8.
- observed problems
In one occasion moving the antenna manually at 0 deg. elevation, after reaching the prelimit down, the antenna moved up at +50% velocity. It occurred three times until vac1 was rebooted. (Q by HU: was this problem with the development version or the current version, see below? => A by JP: it occurred on 17-Sep-2009, project T06-09, using the development version 2008-10-24)
Intermediate Revision 2008-05-31
- implements Lissajous curves and rate for the slow loop of up to 8Hz
Lissajous curves were extensively tested only in this revision
- this revision was used for the GISMO tests in 2008-10-21 to 28. JP: After that, this revision has been frozen as a working revision. Later changes will be implemented in the development version 2008-10-24.
- known restrictions and issues
- fastest Lissajous curves only possible up to a limited elevation
- at least 4 scans with trackFlag -1 for "unset fast traces" (during GISMO tests)
Previous Revision 2008-05-07
[root@vac1 /root]# ls -l /control/ lrwxrwxrwx 1 root root 29 Jul 29 17:14 antenna -> /home/master/antenna.20080507
- Used from beginning of June 2008 as default
- AP: The 40 elements of the array debug of AntennaTraceFastS are used in order to trace all variables involved in the cascade controller.
- AS: basisSystem GALACTIC introduced and tested
- dUT bug fixed in 'slaParams' 2009-05-06
- observed problems:
- CK: strong oscillations leading to tacho failture occuring for unknown reasons on 1.6.09, 26.8.09.
Suppressed the small oscillations, below 1 arcsec amplitude, at low elevation speed (servo loop parameters changed el.cXKp (2560 -> 3075) and el.cXKi (4 ->6). (CK: this was implemented in the current version 2008-05-07 in 12/08.)
- Used the astronomical coordinates at the antenna azimuth axis with an estimated error below 1 meter. Previously we used the astronomical coordinates of the geodesic vertex 19.55 meters out the azimuth axis. Implemented on 5-May-2009
Previous Revision 2007-06-15
antenna -> /home/master/antenna.20070615/
- Used from 2007-10-22 as default
- major change: revised start of OTF subscans: antenna goes to start offsets
Previous Revision 2007-04-26
antenna -> /home/master/antenna.20070426/
- minor change: preset to -1 of fast trace data. AP: "I modified onePPS.c on mrt-peg2:/home/master/antenna/src to preset to -1 the variables azTrackFlag and elTrackFlag in the active fast trace bloc."
antenna -> /home/master/antenna.20070316/
- major change: revised start of TRACK subscans: antenna goes to offsets before counting the time per subscan
- minor changes
- this also includes revision early 2007 of handling of (fast) interrupts (GP)
antenna -> /home/master/antenna.25Sep06
- this also includes revision early 2007 of handling of (fast) interrupts (GP)
This revision implements the correction (software patch) for the "bell" effect bug, it has been applied directly to pointingCorrection() on vlab1 without any change to sunLimit().
Major revision by Alain Perrigouard. See: RevisionsAntennaMountDrive20060629
This revision includes writing debug information to the antmd datastreams.
This revision includes the possibility to write debug information into the antmd datastreams.
Starting at about 17:00 local time we went to a modified version of ant MD see the notes AP and GP below! We did some simple tests (cal, pointing, focus, onoff, otfmap) after this change and we are staying with this version for observations. (we do not expect any major changes for the execution of observing modes from this change!) -- HU
... we have now in the share memory area the elements int xAccel for az and el. az.xAccel and el.xAccel are used to set Ax, the acceleration step, in onePPS.c. They are set to their new default values, 20480, with the new configuration file config.30M. dmpAntenna prints these new variables. -- Alain Perrigouard
add some code to evItSlow on vac1 that will send the antenna trace data as network messages in addition to the current file write. The overhead should be negligible. This way I can work on the receiving end without being noticed. -- Gabriel
During the TT NCS we did final tests of the "3rd version" of the Antenna Mount Drive (antMD) software. (1st, 2nd, and 3rd version are explained in an email by AP 2006-02-08)
This version implements the "smooth transition trajectories" with a limit (maximum) on the acceleration.
The configuration parameter az.xDeltaMax was set to correspond to 600", the maximum acceleration was set (in the source code) to 0.05deg/s^2.
This version has been used for observations from after the TT NCS on 2006-03-06.
2005, November to 2006-03-06
"1st version" (email by AP 2006-02-08).
Note that this version had a bug, and was causing high acceleration and a (anti)backlash-problem.
This version was used for observations from November 2005 to the beginning of the TT NCS on 2006-03-06.
Other versions and configurations of antMD were tested in TT NCS 2006, January to March.