Active Surface -- PTCS Test information
FEM models
- Early days
- The first time the active surface was enabled was March 27, 2002, which resulted in the first good K-band gain curve.
- This resulted in an exchange between Fred Schwab and Don Wells regarding the proper correction for the projection to the surface normal. The original implementation in M&C was wrong, and was fixed sometime after January 2003.
- Later, in Jan 2004, Don Wells revised asAZkit to agree with the change made in the M&C system, as described here.
- Logging
- The first time we started logging the name of the FEM model in the ActiveSurfaceMgr FITS files was in March 2003. It was: "femA59wM95bG.fef" and this as far as I can tell has not changed at all since then. This file is dated Nov 29, 2000, and the "95b" refers to "design symmetric", as opposed to "97f" which signifies "as-built asymmetric".
- All FEM files are kept in /home/gbt/etc/config. There are two types: *.fem, *.fef.
- August 2008: After repairing an LVDT module, the total unstable actuator count dropped to 5, listed in the following table:
- March 2008: According to Pete W., the readback positions of these four actuators seem to oscillate by more than 50 microns. However, he later realized that his mapping to rib and hoop number were wrong. They may be worth inspecting once all the known bad ones have been fixed this suumer. I have added the last two columns to the table below--hopefully they are correct. None of them show up in Jason's service log spreadsheet. All four of these look OK in the Oct03 and Apr04 holography data.
- There are 16 LVDT temperature values recorded in the ActiveSurface-ActiveSurfaceServo-LVDT_temperature log file.
- These are thermometers based on the resistance measurement of the primary LVDT coil (about 100ohms)
- They were calibrated by recording raw data along with a handheld thermometer
- The gains and offsets are stored in gbt/etc/config/LvdtTempGains.conf and LvdtTempOffsets.conf
- The mapping of location with actuator number and hoop and rib is in this spreadsheet.
- Plot of 15 LVDT temperatures from May 15-June 1, 2008
- Plot of the average of 15 LVDT sensors from May 15-June 1, 2008
- Daytime (MJD 0.55-0.85) ambient temperature - LVDT temperature = +0.32 C
- Nighttime (MJD 0.05-0.35) ambient temperature - LVDT temperature = -0.89 C
- Overall ambient temperature - LVDT temperature = -0.18 C
- There are four temperature values recorded in the ActiveSurface-ActiveSurfaceServo-temperature log file.
- These are air temperatures in the actuator room
Mapping of node number with LVDT-module / IIOP / channel / panel
Mapping of hoop and rib to LVDT-module / IIOP / channel
- see ASCII File = /home/gbt/etc/config/asLook_upData.conf
- The five integer columns are: Hoop Rib*10 IIOP LVDT Channel
- see the file: /home/gbt/etc/config/AsZernike.conf which has 6 independent columns: actuator[rib][hoop] x y rho theta phi
- rho = radius from the center of the dish, scaled such that 1.0 = 50m
- theta = angle (in radians) from the y-axis (which originates at the center of the dish and points toward the left edge as viewed from receiver cabin) toward the x-axis (which originates at the vertex and points toward the far edge of the dish).
- Note that: theta = pi/2 + arctan(y/(x-54))
- So, if x'==(x-54), then the x' axis would originate at the center of the dish.
- phi = not an angle, but is the factor for projecting the delta-z values to normal, i.e.:
- Once per scan (into an archivist FITS file)
- Astrid command: SetValues("ActiveSurface", {"writeDataFile": "1"})
- They will show up as a table, e.g. /home/gbtdata/PROJECT_CODE/ActiveSurfaceMgr/2008_05_22_08:53:28.fits
- The size of this file will then be 74880 bytes instead of the normal 2880 bytes.
- Note: The indicated FITS column and parameter are relative to the zero point correction. That is, the raw encoder readback has the zero-point used to bias the command position subtracted out.
- Once (into an ASCII file)
- unix command: /home/groups/ptcs/bin/DumpActiveSurface.tcl (you must first source /home/gbt/gbt.csh)
- Note: The value written is raw encoder feedback (i.e. no subtraction of the zero point.)
Plotting Active Surface FITS tables
- example plots of actuator grouping boundaries: Matlab, python
- Scripts to generate this plot
- Matlab
- /home/groups/ptcs/bin/plotSurface.m (to plot the output of DumpActiveSurface.tcl)
- /home/groups/ptcs/bin/plotSurfaceFITS.m (to plot the ActiveSurfaceMgr FITS Table)
- example: plotSurfaceFITS('2008_08_08_04:54:37.fits')
- python (to run these, be sure that /home/groups/ptcs/bin is in your PYTHONPATH env. variable, as that is where plotSurface.py is located)
- /opt/local/bin/python
- import plotSurface
- plotSurface.plotSurface('filename') to plot the output of DumpActiveSurface.tcl
- cp /home/gbtdata/TPTCSPNT_080808/ActiveSurfaceMgr/2008_08_08_04:54:37.fits ~/
- plotSurface.plotSurfaceFITS('2008_08_08_04:54:37.fits') to plot the ActiveSurfaceMgr FITS Table
- Relevant file: /home/gbt/etc/config/SurfaceActZeros.conf
- It is listed in ActiveSurfaceMgr.conf as "BiasModelFilename := "SurfaceActZeros.conf";
- This is read by /home/gbt2/monctrl/8.4/gbt/devices/antenna/ActiveSurface/manager/Manager/ActiveSurfaceMgr.cc, in the fuction ActiveSurfaceMgr::ActiveSurfaceMgr()
- config.getValue("BiasModelFilename", biasname)
- sprintf(zero_point_file, "%s/etc/config/%s", ygor_telescope, biasname);
- biasModel->loadFromFile(zero_point_file);
- Note that actuatorhomepositions.txt also apparently contains the values, but it is only used by the software to create SurfaceActZeros.conf:
- it is read by ActiveSurfaceMgr.cc
- but only if -createzeropts has been passed as an argument to ActiveSurfaceMgr::ActiveSurfaceMgr() on line 1237
- Procedure for changing it (Note: in the emails below, I have removed the name of the file, because it was incorrect.)
- According to Joe, all we need to do is in the "cleo activesurface" window, turn off the active surface, then turn it on again and click prepare. It is not sufficient to put it into Standby and then take it out of Standby.
Date: Fri, 15 Feb 2008 17:14:08 -0500
From: Amy Shelton
To: 'Fred Schwab'
Cc: 'Todd R. Hunter'
Subject: RE: actuator home position test
Hi Fred,
You should do what you did last time and change the contents of the
file and have the operator restart the Active
Surface Manager. The file lives in /home/gbt/etc/config, which is our
installation directory. The directory you references is our build
directory, so you shouldn't use that one.
You shouldn't have the permissions problem that you had last time because I
asked Wolfgang to add you to the monctrl group. You can change to that
group (if necessary) via the "groups monctrl" Unix command.
I think that you are good to go. Let me know if I can be of additional
help.
Amy
> -----Original Message-----
> From: Fred Schwab [mailto:fschwab@nrao.edu]
> Sent: Friday, February 15, 2008 4:27 PM
> To: Amy L. Shelton
> Cc: Todd R. Hunter
> Subject: actuator home position test
>
> Hi Amy,
>
> I'd like, perhaps next week when Todd has some test time, to do
> another test like the one we did in Oct. 2006, in which we switched
> back and forth between the default actuator home position file,
> and a modified version thereof, during the observations.
>
> I think I just renamed the file appropriately
> each time and had the operator restart the active surface manager,
> but I see that on Oct. 26 you e-mailed me some instructions on using
> an AsConfig script which would ask each time that we supply the
> appropriate file name. I think you had set that up and suggested I
> use it - but then may have phoned me and told me otherwise.
>
> Today, what is the recommended method for such testing?
>
> I remember that at the end of the test I couldn't restore things properly,
> because when I tried to put the default file back in place, the copy of
> the file that I made didn't have the right ownership or privileges -
> something like that - and you had to restore the directory the next day.
>
> It looks like the current working directory is:
> /netapp/code/monctrl/gbt2/monctrl/8.1/gbt/etc/config .
> Is that correct?
> Fred Schwab
-- ToddHunter - 19 Feb 2008
plotSurfaceDiffActual.png 
ActiveSurfaceInformation.xls 
overlayplot.pdf