10:00
Walk-through on ACT modules. Identified a number of (minor) problems to
be fixed:
20:00
ACT: make clean all install of actcal and actcon modules.
13:30
Execution of ccd + act lcu sw of lt5iaa (+GCH)
Test of act + scanning on lt5iab: seems to be ok
Analysis of act panels.
Taken from act/src/ChangeLog
I think that we can use : actcon/src/actconUser.pan as main panel with
possible pop-up panels, when required:
09:00 - 12:00
Working with Lothar Noethe on Active Optics.
We have noticed that under this respect the CCD documentation is very poor, if any, and it seems that some features are not working properly (or we are not able to get them working)
13:00 - 18:00
Working with Lothar Noethe on Active Optics.
09:00 - 12:00
Working with Lothar Noethe on Active Optics.
13:00 - 18:00
Working with Lothar Noethe on Active Optics.
09:00
Discussion with Anders about:
09:30
Looked at attempts to move the telescope, in order to get confidence with
the real system and and with the NTT gui
12:00
Solved problems with access of lt5iaX LCUs from wt5tcs and wa5tcs.
Modified ntt/ENVIRONMENTS/lt5ia/userScript.lt5iaX adding the following
lines:
routeAdd "0.0.0.0","134.171.138.2"
lcubootAutoExec 1,"lqsAddEnvTbl","wt5tcs", "wt5tcs",
"134.171.137.21", 2301
12:30
actcon: renamed some shell scripts to workaround a bug in the vlt Makefile:
when doing make install in the case both an executable and a .sh file with
the same name exist and must be installed, when the executable is installed,
it is overwritten the the .sh. Giorgio said that there is already an SPR,
be that he does not know how to fix it.
14:00
Started test with whole act chaing (WS+LCU): there are still comunication
problems, since @wt5tcs cannot talk to @wa5tcs and vice-versa.
14:30
Fixed comunication problems: the WS environments where not properly configured
and Giorgio has fixed them
12:00
Finished restructuring database for act LCU: now also the WS image is ready
to be installed
14:00
Created shell script to configure scan links on WS for act LCU images
17:00
act: there are still a number of values that are provided as defines but
should be modev to the database, singe they change from telescope to telescope
(mirror radius, CCD pixel size...) This is not urgent but it is anyway
important. It will go on the list of things to be done for the next release.
21:00
Modified WS environment wa5tcs for new act LCU image database. I have saved
the current environment in
$VLTDATA/ENVIRONMENTS/wa5tcs.bak
and regenerated the environment with "make wa5tcs" from the ntt module.
THE SAME OPERATION MUST BE DONE ON wt5tcs.
21:30
act: Now the tables with the results of act image analisys can (and are)
accessed from the WS only wia direct dbREAD from ws to LCU. This could
be replaced by a command, buy is this really necessary? The amount of data
is quite big (more than 8kb, in binary format) and the job that must be
done by the command is much more that the dbREAD. At the same time these
data are usefull only from time to time when something goes wrong or to
save data for historical/test reasons. We think it is not necessary to
write a command and dbREAD is good enought.
22:00
act: completed script actInitScan.tcl to activate scanning and tested.
Now the ntt scripts for the lcus must be configured to call the script.
09:00
There are some (knows) inconsistencies between the setup file alias and
disctionary files in the modules using them.
We have agreed that is is now time of fixing this, even if it will take
approx half a day of work.
We have then decided of:
The following activity requires:
11:00
act: discussion about independency in panels from lcu. The act panels must
have the possibility of setting the lcu to work on, while the actcon panels
and the user panels must use the database values.
11:30
checked out from Gaching: tcs, msw, prs, trkws
14:00
tcs: moved here tcsTCS.alias and tcsESO-VLT-DIC.TCS. Checked merging of
prs, trkws and agws versiones. The files must be installed in INS_ROOT
as TCS.alias and ESO-VLT-DIC.TCS.
15:00
prs: removed prs.alias and prsESO-VLT-DIC.TCS. Now using versions provided
by tcs module
agws: done the same
trkws: done the same
PLEASE CHECK THE ChangeLog files for details
ntt: buildTcsKit is not using any more tssimInstall.sh but its own new file installation program nttInstall.sh. When back to Graching I would have to propagate to tcssimInstall.sh the same changes. I guess that we have also a ycsintInstall.sh, but I have to check.
16:00
trkws: fixing pending trkws problems (working on woff).
20:30
ACT tests on real sky.
Using side B.
Autoguider(TPH+GCH)
10:00
Verification of tccd side B by running ccd sw on:
14:00
Align and focus of lt5agb with Wolgang, lt5iab already ok.
15:00
agws
17:00
ag
Update ntt/ENVIRONMENTS/lt5aga/userScript and lt5agb to add gateway + generation of these 2 env.
No test of ag today
09:00
trkws: working on trkwsMANAGE_LIST::RemTrkTimeEventCB()
10:00
agws: working with Than on handling of focus station. Look at his log for
details
18:00
tif: debugged tif with SUSI. It does not work because tifInitLibs() checks
for the existence of the tif database branch using dbGetDirAddr(), but
this does not work on remote environments. It this case SUSI OS is running
on wemmi, while TCS is on wt5tcs.
Some notes about error handling in SUSI:
For what I have been able to see there are some problems with error
handling. In most cases error conditions are just handled with an errAdd(),
without taking any action, just continuing as nothing happened.
What is worst, quite often a successive function call RESET the error stack
(this should not happen), the only remaining trace of the error is, if
any, an error flag.
As a result, some time after the real error a generic error is logged but
all the significative error messages in the stack are lost forever.
I would suggest an analysis of error handling in the code in order to use better strategies. I cannot say how much this is a peculiarity of the SUSI OS or is due to the implementation of the INS basic libraries used.
21:00
tif: fixed the bug replacing dbGetDirAddr() with dbGetAlias().
Tested with mod test and by hand.
Archived, copied and put in NTTSource
Now it must be compiled and installed on all INTROOTs
Autoguiding (TPH+GCH)
10:00
Scan of "requested" rotator angle, previously patched using the
attribute parallactic angle, now changed to be scanned from trkMonitoredData.adapt
into trkws "Interncoord.adapt". This is a temporary solution.
This scanned data should be part of the trk "Interncoord" db
class.
=> Need to change trkws db class. (add attribute adapt)
11:00
agws: change focus
All these 3 points are simplified by a correct use of the msw table and a better definition of the agws db.(thanks to GCH)
=> Update of TCS.db => Update of agws
17:00
Generation of all env. for wa5tcs
09:00
Presentation on VLT OO soft for La Silla sw people
11:30
Investigating with Anders problems with msw change focus procedure.
See his log for changes in M3 LCU software.
Fixed some problems in msw.
See msw logs for details.
Tuned some timeouts on the database of wa5tcs.
Investigating how to simplify the msw list table.
Fixed problem in mswMOD_LIST_COMMAND::Send.
ATTENTION: it is now necessary to fix it in trkwsTRANSFER_ASM.C, since now the Send() can and will return with an error if it fails and this error must now be recovered by trkws.
16:00
Relaxed conditions on state/substate for executing mode change commands
so that the sysyem does not gets locked any more in case of intermediate
errors.
For example, now STOP is accepted in ONLINE and (STANDBY,ERROR)
Look at msw change log for details.
17:00
msw: The COLD START command is crytical, since there are preconditions
to be checked before it can be executed. For example the hidraulics and
the building must be already online. In order not to make too complex the
MSW module and having at the same time an easy implementation/configurability
of these checks, I suggest of implementing a sequencer check function,
to be called by msw. It has to return "OK" if the preconditions
are fullfilled or a proper error message if not, like:
"ERROR - Please put ONLINE building module"
This error will be reported to the user.
08:30
Discussion on Auto Guiding
Among other things that will be done today (see following sections of the log), there are a couple pending and to be examined:
Other pending points:
09:00
Discussion on PRESET sequence with AWA and JSP.
I will implement it now.
In order to handle the preset sequence when we want NOT to use some modules, they can be dynamically replaced by dummies: just run the dummy, change the target process in the prs DB to the dummy and send INIT to prs. Viceversa to reactivete the module.
Asmall script could be written to make this "automatic" and a button for this purpose could also be added to an engineering PRS gui.
- prs: added interface to PRBGS autoguiding command, that will be on the NTT instead of FINSGS, singe the CCD does not provide now the OBJCNT command.
10:30
- trkws: fixed pending problem with error handling in asm loop
11:00
- agws: Examined how vignetting is implemented.
The keyword is currently available in the commands but not used.
Two values are available in the database to define (in arcmin) the minimum
and maximun search radius. These are actucally used. when msw will handle
instrument parameters, these values or even a more complex implementation
will be available and agws will take care of reading them at INIT time.
For the time beeing we can live with the current implementation.
21:00
Than performed the check for scanning one LCU in two workstations.
THIS IS NOT POSSIBLE.
After discussion with Anders and Jason we have taken the following decision
(I add here some more implementation details). The LCU image branches for
ag and act will have to be on wa5tcs. This should not be a problem anyway,
but panels have to be designed in a way that it is possible to configure
the environment from where to read WS data and LCU image data.
Preferibly the panels should run on wa5tcs, to better distribute the load
but it will be anyway possible to run them on wt5tcs. The panels will accept
2 command line parameters: one for the TCS WS environment (by default wt5tcs)
and one for the LCU WS image environment (by default wa5tcs).
The alternate solution would be to mirror the LCU image on wt5tcs scanning
from wa5tcs, but this would just increase the CPU load without any advantage.
22:00
Some support to Lothar from writing a log program to generate an automatic
report of their measurements.
This sequencer script goes all arount the databases of wt5tcs and of the
act LCU directly reading the requested information. This is a quick and
dirty solution that does very well its job. On the other hand this way
of accessing the database should not be considered legal, since it bypasses
the command interfaces.
This script should not became a "standard" act tools: if this
is the case should be rewritten using tif and other legal interfaces.
22:30
A consideration: we do not have a sequencer TIF interface!!!!
Should we have one? I think so: if you agree, submit a change request.
23:00
To be checked in msw:
To be checked in preset:
08:30 - 14:00
- msw: Clean up of msw table on the database
- msw: mswCHANGE_FOCUS.C (ReadMirrorsPos): Quick and dirty implementation of this function. The DB address of the m3 attribute containing the current position is hard coded. Proper defines and the msw table will have to be used in the final implementation.
15:00
- wt5tcs and ta5tcs environments cleanup. Look at Ander's log
20:00
Configured environment wt5tcs to run the dummy nttGenDummy.
This can be used configuring the PRESET module to emulate not available
modules that are called in the preset sequence
The process is run in the RtapEnvTable.
To run it by hand issue:
evhDummy -n nttGenDummy config/nttGenDummy.cfg
The configuration file nttGenDummy.cfg is in the ntt/config directory and is installed in $VLTTOP/config by "make install"
15:00
prs - fixed problem with abort or error on PRESET to object. In this case
it is necessary to send msw an END FOR ENGINEERING preset, not for ASTRONOMICAL
preset, so that is nows that the system is not tracking but, in the best
case, idle.
17:00
Testing TCS. See AWA logs
21:00
Fixed some configuration error in the msw table and in the msw panel class.
The main problem with the panel class is that it has been necessary to
hard code there some information that should be extracted from the table.
This makes maintenance of the panel a problem, but we do not know how to
do it.
It seems that msw has problems when on the table there are addresses on a remote ws.
To be checked, but the entries for ag seems to have problems.
21:00
Problem 2 on AWA:
Problem with General timeout of timeout of timeout of.. error. Identified timeout value as a define: it is too low. Sent mail to R.Karban asking to move it on the database for tunability.
22:00
When auto-guiding trkws reports sync errors of the order ot 1e-05 It is
possible to set on DB the trkws maximunt time error to fix it (done on
the on-line db but not tested, it needs an init to trkws). It would be
better first to understand if the problem can be removed on the LCU.
24:00
Problem 5 on AWA:
Identified problem when LCU goes off tracking and trkws does not stop the other LCUs (see AWA log).
IT IS A PROBLEM OF CONVENTION WITH STATE NAMES:
09:00
Patched Problem 2 in AWA report from 961209 (timeout on timeout..).
Received confirmation from from RKA : he fill fix it in trk
14:00
msw
trkws
16:00
act (with Lothar and Philip)
24:00
Archived modules: tcs, prs, msw, trkws.
Put a cmmCopy in NTT..../TCS
Problems during the night:
09:00
Read mail of the last 2 days (I was on Hollyday!!!)
09:30
cmmCopied from Graching the last verisions of trk and trkws and compiled
11:00
Analysis of the implementation in trkws of setup commands and rotator commands.
11:30
Discussion with Anders and Jason on preset sequence and handling of
SETUP commands.
A setup command to a new target can contain keywords to set WLENGHT, OFFSET, rotator OFFSET and ADDITIONAL VELOCITY, but these keywords CANNOT be passed in a setup file as isolated keywords, without a targe, since they are intrinsic in a targed definition. If an application wants to pass these as single parameters it HAS to use the corresponding commands (this has been checked and agreed with Erik).
The preset sequence has also been reworked to send commands to trkws when these keywords are present.
It is also important to notice that the offset should be a COMBINED OFFSET, and it should be sent AFTER having started auto guiding. We do not have now a combined offset command, so for the time beeing I have put a normal offset BEFORE starting auto guiding. The OFFSADG command that was used is not correct since it is "combined" but it does not change telescope coordinates at we need.
12:00
Checked out modules prs and tcs from Garching for updates.
14:00
Added parameter Offset Rotator to PRSCOOR command in tif and prs
14:30
Added in prs handling of commands to trkws: RotAngle(), AddVelocity() WaveLength()
and modified NTT preset sequence to read the corresponding keywords from
the setup file and send the commands to trkws.
16:00
Checked out msw module from Garching.
16:30
Added some logs in msw when a basic command is started and completed succesfully.
18:00
Modified the actcon database to allow proper overloading of values in TCS.db
Changed TCS.db, rigenerated environment and rebuilt everything. AO and
AG are not available for preset. I have activated the dummy. Remember to
do it every time you regenerate the environment using the commands:
nttSetDummy -p ao
nttSetDummy -p ag
19:00
The just cmmCopied version of trk was not including the patches for "timeout
on timeout" and additional velocity.
I have patched again the read only version.
09:00
Working on PRS problems.
trk:control.transferTimeout 300
trk:control.offsetTimeout 300
prs:control.stopTimeout 300
The changes have to be put also in the TCS.db. It would also be probably better to add specific timeouts in trkws for these actions, without using the default ones.
14:00
Debugging of actcon problems. See PGH log for details
15:00
Checked out evh from Garching into ~/NTTSourceNOV96/VLTSW_new and debugged
problem.
Made a patch in evh.
When I am in Garching I will have to fix it cleanly.
Now I have done just a patch, so that only the library is changed and
there is no need of recompiling the modules but only of relinking them.
The patch seems to work fine from 20 min tests, but there could still be problems if there are error conditions from the called modules.
Remember to include the module in the list built in VLTSW_new. Tomorrow, if all OK, I will archive it.
18:00
TCS.db for wt5tcs:
added missing pieces in tif branch: I forgot to put here the branch with
temporary attributes and events.
The database needs to be regenerated.
The following attributes are not set now:
ATTRIBUTE DOUBLE airmass 0 ATTRIBUTE DOUBLE parallactic 0 ATTRIBUTE DOUBLE diffVelRA 0 ATTRIBUTE DOUBLE diffVelDec 0 ATTRIBUTE DOUBLE remTimeLim 0 ATTRIBUTE LOGICAL trackingState 0
We should check them. Some are available and would be enought to add some CE formulas to read them, but some are not available.
09:00
Checked logMonitor of last night.
Pretty good! No significative errors logged from TCS. A lot of errors from actcon. I have saved them and I will look at them with Lothar and Philip
10:00
prsAction: starting debugging of prsAction sequence.
There is a core dump (identified yesterday) when trkwsControl fails to preset and returns an error.
11:00
trkwsControl: when one axis fails to preset (for example the given coordinates
are out of limits) the others are not stopped. This should be fixed, since
it does not make any sense to continue presetting.
14:00
prsAction seem OK. I have tested many commands and errors and some setup
commands to verify the sequence.
Archived on garching evh with the fix.
Remember to update the build scripts to include it
Result of prs sequence testing with various setup files:
17:00
- msw: renamed AO permissione request commands as GETPERM/RELPERM
21:00
- trk: identified problem with additional velocity. We set an additional
velocity. If we send a new preset, it seems it is reset, but as soon as
PRESET is completed and the system starts TRACKING it reappears. This could
be related to a previous error with a preset with a position out of range.
According to Jason this could be also the case for a rotator offset. To
be checked
22:00
- act: working on database restructuring for the ws modules.
09:00
prs: added parameter STAR in call to AO command CYCLAO.
10:00
tif: restructured database creating a tifDB_NTT.class.
To give a temporary implementation of missing tif named parameters I have
modified the tif DB temp branch adding calculation engine formulas:
CLASS tifTEMP tifTEMP_NTT
BEGIN
ATTRIBUTE DOUBLE airmass 0
BEGIN
Definition [:Appl_data:TCS:LCU:alt:trk:intCoord.airMass]
END
ATTRIBUTE DOUBLE parallactic 0
BEGIN
Definition [:Appl_data:TCS:LCU:alt:trk:intCoord.parallAngle]
END
ATTRIBUTE DOUBLE diffVelRA 0
BEGIN
Definition [:Appl_data:TCS:LCU:alt:trk:intCoord.alphaVelAccu]
END
ATTRIBUTE DOUBLE diffVelDec 0
BEGIN
Definition [:Appl_data:TCS:LCU:alt:trk:intCoord.deltaVelAccu]
END
ATTRIBUTE DOUBLE remTimeLim 0
BEGIN
Definition [:Appl_data:TCS:trk:data:setup.remTrkTimeLimit]
END
ATTRIBUTE LOGICAL trackingState 0
BEGIN
Definition \
IF([:Appl_data:TCS:msw.substate]=mswSUB_STATE_TRACKING, 1, 0)
END
END
CLASS tifTEMP tifTEMP_NTT
BEGIN
ATTRIBUTE DOUBLE airmass 0
BEGIN
Definition [:Appl_data:TCS:LCU:alt:trk:intCoord.airMass]
END
ATTRIBUTE DOUBLE parallactic 0
BEGIN
Definition [:Appl_data:TCS:LCU:alt:trk:intCoord.parallAngle]
END
ATTRIBUTE DOUBLE diffVelRA 0
BEGIN
Definition [:Appl_data:TCS:LCU:alt:trk:intCoord.alphaVelAccu]
END
ATTRIBUTE DOUBLE diffVelDec 0
BEGIN
Definition [:Appl_data:TCS:LCU:alt:trk:intCoord.deltaVelAccu]
END
ATTRIBUTE DOUBLE remTimeLim 0
BEGIN
Definition [:Appl_data:TCS:trk:data:setup.remTrkTimeLimit]
END
ATTRIBUTE LOGICAL trackingState 0
BEGIN
Definition \
IF([:Appl_data:TCS:msw.substate]=mswSUB_STATE_TRACKING, 1, 0)
END
END
This will be removed when the final implementation is done, but it will be transparent to well behaving programs.
11:00
TCS.db for wt5tca cleaned removing CCD LCU branches and restructuring tif
branch
12:30
msw: When an LCU is rebooted, msw is not informed. In this case the quality
of the state attribute should bo bad and there should be a way for msw
to get informed. Check this and implement what necessary.
13:00
trkws: patched to send a STOP to all tracking units one one fails to preset.
Sent a mail to Graching with details so that they can make the "official"
fix inside the module.
14:00
actcal: going on restructuring the module with Lothar.
16:00
Found the problem of the additional velocity.
This is SOMEHOW the expected behaviour.
When the additional velocity (or the target offset....) are set with a
setup command, this goes in the current setup file. New setup commands
(or preset commands) will just add/replace keywords to the current setup
file, the not replaced keywords remain there and are reused at the next
setup.
IS THIS WHAT WE REALLY WANT? PROBABLY NOT OR NOT EXACTELY!!!
09:00
Check of last night log: ok
10:00
tif: checked out from Garching. Adding loggin off all received and completed
commands. Archived in Garching again.
10:00
actcal: working with Lothar.
16:00
Important discussion with Jason and Paul on how tif handles commands
during calibration when an instrument is not on the light path.
The current implementation (corresponding, for what I know to the TCS/INS nver released ICD) accept a subset of specific commands directed to the adapter/probe /rotator of the corresponding focus.
In a complete system the booking system will take care of blocking commands directed to the prs, trkws and so on control modules when the tif is not the one of the current focus station. Now all commands are accepted, but this is just a matter of configuring the booking system.
The commands acceptd when not on the light path are a specific set of commands. A generic SETUP command will be rejected, since it would be redirected to PRESET and not to an adapter/rotator.
This implies that the procedures to be used for calibration out of focus must be in general not the same used when on the light path.
For example all commands in sky coordinates are no allowed.
It will have to be clarified if there are other needs from the INS side.
09:00
This is my last day here, so I have done a general clean up. You
still find in my directory the ddd debugger and the graphc progam xv, that
could be usefull
11:00
I have done a small test for setting up the scan links from act LCUs to
the WS database.
I have fixed a couple of errors and now it seems to work, although it does
not work for iab (that is probably down, but I have not checked).
To activate the scan links the act module contains the actInitScan tcl
script.
You can put a call per LCU in your ntt scan scripts.
Look at the man page for the details.
Xmp:
To activate scanning from lt5iaa:
actInitScan -l lt5iaa -w iaa
To turn OFF scanning from
lt5iaa: actInitScan -del -l lt5iaa -w iaa
Last modified: Thu Dec 30 15:01:55 MET 1999