Torques at various turret max velocities

• maxvelocity = 3 deg/second (normal running)
• maxvelocity = 2 deg/second
• maxvelocity = 1 deg/second
• maxVelocity =3 and the turret was run with a a rate of 1 deg/second
• The feedback values are   read at a 1 Hz rate. Any oscillations greater the .5 Hz will be aliased in the data.
• Looking down on the turret
• CCW motion increases the turret angle (26 to 340)
• CW motion decreases the turret angle (340 to 26).
  

    In the first 3 runs, the floor will go as fast as possible (while staying within the maxvel limitation). There  is not much feedback on the velocity since it is being clipped at the maxvel. For the final test, the 1deg/sec rate is within the maxvel rating so the pid loop will  be able to feed back the error into the velocity.

The tests were done with the dome at 11deg za. This has a floor tilt of 1 deg (the floor is level at za=10deg).
For most tests the turret was moved from 26 to 340 degrees (alfa to 327 rcvr).
The maxvel=1deg/sec run only had a CW rotation ( i ran out of time).

The plots are:

• max velocity of 3 deg/sec (.ps) (.pdf)
• max velocity of 2 deg/sec (.ps) (.pdf)
• max velocity of 1 deg/sec (.ps) (.pdf)
• max velocity of 3 deg/sec and rate of 1 deg/sec (.ps) (.pdf)

The 3 pages of each plot have:

• Page 1: position, encoder velocity and torques vs hour for both motions.
• Page 2: CCW motion
• position, velocity, and motor torques vs hour,
• motor torques vs position
• Spectrum of the total motor torques
• I just added the master and slave motor torques.
• the recorded torques are only positive numbers. If the backlash compensation is active for slow motions, one of the torques may be small but negative.
  
• Page 3: CW motion

Discussion:

• Torques :
  
• CCW: (26 to 340 degrees)
  
• The master motor applies most of the the torque. The slave motor does not do much
• CW: (340 to 26 degrees)
• The slave motor applies more force, but the master value is almost the same.
• position variation
• 145 to 26 degrees needs a larger torque with more variation. especially in the CW (150 to25 direction)
• This is more obvious in the 2deg/sec and 1deg/sec motions
• Velocity:
• computed the velocities from the change in encoder position.
• The measured velocities did not match the maximum velocities.

• MaxVel requested	Velocity measured	measured/requested
  3 deg/sec	2.09 deg/sec	.66
  2 deg/sec	.84 deg/sec	.42
  1 deg/sec	.42 deg/sec	.42

• The calibration needs to be checked.
• This is  not seen in the 1deg/sec rate (because the pid loop had enough extra velocity to correct any errors).
• The CW rotations (340 to 26) velocities had a lot more variation (oscillation)
• It was especially strong in the 3deg maxvel run (this may be an aliased value).
  

Summary:

• The turret velocity does not match the maximum velocity.
• we need to check the calibrations.
• The variation in velocity in the CW direction (340 to 26) has a lot more variation than the CCW motion.
• The torques needed for 145 to 26 degrees (CW) are larger than other parts of the motion.
• this corresponds to when alfa is pointed a little uphill from the door (120 degrees from focus) to when alfa is in focus).
• since the floor  angle was 1 degree, i doubt if this is a gravity problem (and it didn't show in the other side of focus).
• it may a problem with the floor support structure.
• The spectrum of the velocity shows some spikes, They are stronger in the CW rotations (decreasing  turret angle)
• The sampling was done at 1Hz so any frequencies above .5 hz would be aliased.