NRAO Home  >  Green Bank  |  Wiki Topic:    GB > PTCS > InformationForProjectTeam > ServoSystem

Changes | Index | Search | Statistics | Go

Servo System Improvements

• Modification Requests
  • ServoFrictionCompensation
  • MR1C806 (including the shell commands to be entered on gbtccu)
  • ServoInterlock Project
  • MR3C308 (Increase sampling bandwidth on Servo Monitor)
  • MR2C308 (Add Decimation Filter to Sampler2log)
  • MR9C208 (Change Method of Velocity Generation for Pointing Model)
  • MR3C208 (Fix Pointing Update Command Glitches)
• Diagnostic Scans
  • Circular scan around a source at the half-power contour: As suggested by John Ford, I wrote an interactive python script (/home/groups/ptcs/obs/servo/circle2.py) to generate this scan (called circle_r39_s15_p16_5times.fits) and ran it tonight on 3c84 at X-band (9 GHz, radius=39", velocity=15"/sec, scan 63 of TPTCSPNT_070111 = 2007_01_12_04:29:36.fits). The scheduling block is called "CircleAtHalfPower" under the TPTCSPNT database. A comprehensive plot is attached below. The largest following error (about 22" in azimuth) occurs just after the azimuth axis goes through zero velocity, and corresponds to about a 30% flux level. -- ToddHunter - 12 Jan 2006
• Notes on how to change the friction compensation coefficients
  • permanently
  • temporarily:

      belinda> ssh prima
      prima> rlogin gbtccu
## must use floating point format for parameters
# 0 = azimuth, 1 = elevation
-> setFrictionCompensation(0,0.60)
-> setAclDACLimit(0,2.0)
-> setAclDACLimit(1,1.0)
-> setFrictionCompensation(1,0.10)
-> showGains(0)

To start the real-time monitoring of spare channels:
thunter@vortex ~]$ source /home/gbt/gbt.csh
[thunter@vortex ~]$ remoteWindow spares gbtsmu 1

To figure out the name of the spare channels:
[thunter@belinda ~]$ rlogin gbtsmu
VxWorks login: tweadon
Password: 
-> dbmWindows()
Window 0 = mcTime 
Window 1 = sequencer 
Window 2 = VMIC3118 
Window 3 = encoders 
Window 4 = azTorqs 
Window 5 = azTachs 
Window 6 = elTorqs 
Window 7 = elTachs 
Window 8 = summary 
Window 9 = spares 

Spare channels:
spare 1 = current command
spare 3 = rate feedback (from tachometer)
spare 7 = azimuth friction comp.
spare 8 = elevation friction comp.
spare 9 = wheel tilt

6. to run gbtlogview:
  ssh any linux machine
  source /home/gbt/gbt.csh
  cp file from jford's   (.gbtlogdevices.g)
  gbtlogview
  highlight gbtlogs/azel50hz
            gbtlogs/azElCommand
            gbtlogs/Spare50Hz
  click: Last Hour
  click: Fill
  click: Dismiss (when done)
  click: Plot columns...
  for azimuth plots:
  click: "Time" under X, "-Az_El_50Hz_Az" under Y1, "-azElCommand_Az" under Y2
        select Y2+180 under Y2
  click: Plot...
        force Y1min = Y2min   and Y1max = Y2max
  use middle button to define area to zoom, right button to cancel out of zoom

Daisy petal scans

Plotting them

tt=findgen(200)
taur=60.0
azpdratio=4.0
tauaz=taur*azpdratio

omr=2*!pi/taur
oaz=2*!pi/tauaz

xx=sin(omr*tt)*sin(oaz*tt)
yy=sin(omr*tt)*cos(oaz*tt)

plot,xx,yy,/iso

Running them

Astrid provides a standard predefined daisy scan type, documented at
http://wiki.gb.nrao.edu/bin/view/Observing/ScanTypes#Daisy

An example invocation, with parameters I might suppose to be reasonable for your application, is:

Slew("3C123")
Daisy(location="3C123", map_radius=1, radial_osc_period=30, radial_phase=0, scanDuration=90,coordMode='Encoder')

This does a 1' radius map of the form described in my ptcs memo about scan patterns. In particular the ratio of
radial to azimuthal phase is fixed at, I think, pi. I have written another version where that ratio is a free
parameter if you need the generality. Other parametric scan patterns are pretty trivial generalizations as well.

By preceeding the command with a slew to source, and having the trajectory start on source as it does for these
parameters (relevant one is radial_phase=0), I've found that the feedarm doesn't jerk around much. Your mileage
may vary, and I'm not sure that's the only way these days, but so far it's worked for me.

Let me know if you have further questions.

 cheers,
  Brian

Attachment:	Action:	Size:	Date:	Who:	Comment:
circle_xband_posvel.pdf	action	177826	12 Jan 2007 - 06:54	ToddHunter	Circular scan around half-power contour at X-band