• Page 1:
  
• top : turret position vs hour of day
• bottom: velocity command sent from cpu, and velocity feedback
• the velocity feedback is read after the signal conditioning board and before the amp.
• page 2-page9
• turret status bits vs hour of day.
• page 9 has the faults.
• the mot12spdGap error  happens when the velocity reported from the two motor amplifiers differs by more than a defined amount.
• This may just be a calibration problem.
• page 10: motor torques for the master and slave motor.

torque oscillations (.ps) (.pdf)

• page 1: status 0 to 16 hours
• top: turret position
  
• 2nd motor torques. black master, red  slave.
• 3rd; motor1,2 speed gap error.
  
• This happens most times the turret moves (except 6:55am)
• 4th: motor speeds. master and slave. This shows when the turret moved.
• bottom: master speed= slave speed.
• This shows the speed different between the 2 motors as read back from the amplifiers.
• Page 2: blowup 8 to 10:30 am
• The position move around 10:00 was when the turret was jerking (at least someone was there to see it.. it may have also occurred at other times).
  
• You can see the torque oscillation 9-10am
• There is a speed gap error for most of the 10am move.
• Page 3: 9.3 to 9.4 hours. turret sitting at lbw
• The position is oscillating by 1 a/d level (about 2 millidegrees)
• the motor torques are oscillating because the turret is moving (by only 1 count).
• the torques are getting up to about 8 foot-lbs (in master)
• Page 4: the position torques at 10:00 when the floor was jerking.
• Page 5: 9.3 to 9.95 hours.. turret sitting at lbw.
• Top: position oscillation with drift
• 2nd : motor torque
• bottom: spectrum of the torque oscillations
• I used 9.3 to 9.95 hours
• the period is .0136 hz .. or about 73 second period  (see purple lines in frame 1,2)

• I've plotted 3.0 to 5.0 hours. the turret was sitting at 26 degrees (alfa in focus)
• There are oscillation at 3.25 to 3.5 and 4.4 to 4.7 hours.
• this corresponds to za  14-13 deg and za 5-6 degrees.
• The turret is level at za=10 degrees.
  
• You might expect things to be unbalanced at the highest (20 deg) or lowest (1-2) degrees.
• We are seeing oscillations in intermediate za .. so it may not be associated with the floor tilt.

Summary:

• When the turret is commanded to sit at a position, the position  and torque oscillate with a period of about 73 seconds.
• The position is only moving by about 1 or 2 a/d counts
• the torque (in master or slave) can be changing by up to 8 ft-lbs.
• This is  coming from the PI loop.
  
• I wonder if the amplitude of the torques needed for the motion is  because the friction (stiction) might be large.
• It may also be a problem in the anti backlash board adjustments
  
• It might be interesting to see if the amplitude of the torque is a function of where it is sitting?
  
• this might tell us something about bad spots on the rail
• unfortunately we normally only sit at discrete locations (where the receiver is that we are using).
•