# 12meter motion control Acceptance test

#### 24jan11

Test documents and results
Notes from test

# Intro:

The acceptance test for the 12 meter motion control was performed 24jan11 thru 27jan11. The specifications in the test document were met by the system.
Lingo:   SAT: System Acceptance Test

# Mechanical tests (surface accuracy,alignments):

mechanical site acceptance tests (.pdf)
cobham report on how deformations measured in mechanical test affect gain, pointing (.pdf)

# Controller Test Documents and results

The pdf links below point to the 4 sections of the SAT test. These show what is to be done, they do not contain the results. For the  servo performance tests, data was recorded using the CT program provided by Emerson technologies. Plots of this data are provided.
• SAT part 1 Tests in Local Control (.pdf)

• SAT part 2 Remote Interface Functional tests (.pdf)

• SAT part 3 Servo Performance tests (.pdf)
• slew rates (.ps) (.pdf) pages 5-7:
• Measure average velocity for full speed moves between positions.
• Top: move el 88 to 7 deg, Center: move el 7 to 88 deg (az=0)
• compute average velocity between el=[8,87] degrees
• needs to be 1.25 deg/sec or greater
• You can see the max velocity vs elevation restrictions.
• > 75 degrees max velocity is 1.5 deg/sec.
• Ringing in velocity when you start moving down 88 to 7 lasts for about 4 seconds.
• Bottom: move az -90 to 90 with el=10 deg
• Max velocity should be 5 deg/sec
• Ringing in az velocity at max velocity
• accelerations (.ps) (.pdf) pages 8-10.
• Measure average acceleration by stepping velocity between 2 values.
• Linear fit to where the velocity is changing to get the acceleration.
• Values should be >=1.34 deg/sec^2
• Top:el=80, center: el=10. move from -1deg/sec to +1deg/sec in elevation.
• Bottom: move from -5deg/sec to + 5deg/sec in azimuth (el=10 deg)
• motor currents (.ps) (.pdf) pages 11-14.
• Measure motor currents while moving telescope. Record current and % of overload current.
• The vertical dashed lines show the start, end of motion.
• Page 1: El  move 88 to 7 degrees then 7 degrees to 88 deg
• top: motor currents moving 88 to 7 degrees
• 2nd: % of overload current moving 88 to 7 degrees
• 3rd: motor current moving 7 to 88 degrees
• bottom: % of overload current moving 7 to 88
• Page 2: Az moves 180 to -180 then -180 to +180. el=7
• top: motor currents moving 180 to -180 degrees (CCW)
• 2nd: % of overload current moving 180 to -180 degrees (CCW)
• 3rd: motor current moving -180 to 180 degrees (CW)
• bottom: % of overload current moving -180 to 180 (CW)
• static tracking errors (.ps) (.pdf). Pages 15-16.
• sit at el= 10,45, and 80 degrees each for 120 seconds. Measure the tracking error.
• The wind varied from 0 to 15 mph during these tests.
• Top: el=10 deg
• middle: el=45 deg
• bottom: el=80 deg.
• the largest tracking error was 2.4 milli deg in az when el=10deg.
• Tracking errors moving at constant Velocity (.ps) (.pdf). pages 17-19 velocity response.
• Tracking error while moving at constant velocity.
• Black in elevation error, red is azimuth error (units are milli degrees).
• The vertical green lines show the time period where the peak, rms errors were computed. This excludes the error  during the startup acceleration.
• The rfp says that the telescope should track with an rms < 5 millidegrees.
• Page 1: Tracking error while moving in Azimuth. Elevation=45 degrees
• top-> bottom. Azimuth velocities: .001, .01, .1, and .4 deg/sec.
• Page 2: Tracking error while moving in elevation. Elevation=45 degrees
• top-> bottom. Elevation velocities: .001, .01, .1, and .4 deg/sec.
• Page 3: Tracking error while moving in elevation and azimuth simultaneously.  Elevation=45 degrees
• top-> bottom. Elevation,azimuth  velocities: .001, .01, and  .1 deg/sec.
• Step responses (.ps) (.pdf). pages 20-21.
• Move axis by .02, .1, and 2 degrees back and forth.
• Step az and elevation separately.
• Do each step at elevation 10,45 and 80 degrees.
• The telescope should settle to < 5 millidegrees within 5 seconds.
• Each page has 3 frame that correspond to .02, .1 and 2 degree steps.
• Each  frame over plots the results from the 3 elevations:  black 10deg, red 45 deg, green 80 deg.
• The 3 plots of each frame have been aligned to the positive going step.
• Azimuth Steps:
• Page 1: Azimuth steps
• this plots the entire range of the motion
• Page 2: Azimuth step settling error.
• The step end position as been subtracted from the plots to blowup the vertical scale.
• The left curve shows the step up, the right curve shows the step down.
• Elevation Steps:
• Page 3: full position step
• The positions have had the start elevation (10,45,80) removed.
• Page 4: Elevation step settling error.
• The step end position as been subtracted from the plots to blowup the vertical scale.
• The left curve shows the step up, the right curve shows the step down.
• The .02 step (green) at elevation=80 degrees rings for about 5 seconds.
• RA, Dec tracking errors (.ps) (.pdf (replaces page 22,23).
• Track ra=242Deg,dec=-20 (az=231,el=31) and record the az,el tracking errors.
• This tests an normal motion that the telescope would make while tracking a source.
• Clock Test  (.ps) (.pdf)   (not part of the SAT procedure).
• The above plots used the sampling clock provided by a labview program running on a laptop talking to the controller (actually it was the Emerson CT scope program that looks like it was made from labview). The various tests set the sampling rate at 10 or 200 milliseconds. It would then sample the various registers (position, position err, velocity) at this rate.
• The clock test had the labview program read the clock on the controller at a 10 millisecond rate.
• The labview time and the controller time was recorded at each sample
• The controller clock register  is updated every 4 milliseconds on the controller.
• Top: Measured labview clock interval vs controller clock interval
• The measured time between samples was computed for the labview and the controller times.
• Yaxis: the measured labview time interval
• Xaxis: the measured controller time interval.
• Bottom: Histograms of the controller and labview time intervals.
• Discussion:
• the labview time:
•  peaks at 10 and 20 milliseconds. If things were perfect, it would always be 10 milliseconds.
• the values are continuous from 0 to 33 milliseconds.
• The controller time
• Peaks at 0,8,16, 24 milliseconds
• the controller time is discrete. It has steps of about 8 milliseconds. Within the controller it should be discrete at a 4 millisecond rate.
• occasionally the controller clock had a 0 time interval while the labview interval changed by up to 30 millisecs.
• The are timing errors in the labview and reading of the controller clock.
• Previous measurements showed that it took a c -program running flat out about 7.8 milliseconds to read a single register from the controller. This may be what gives the 8 millisecond step in controller time.
• SAT part 4 Tests for Pointing and Tracking accuracy (.pdf)

# Notes from controller test:

• 24jan11:
• the azimuth directional limits use a relay to tell which wrap it is on.
• If you have a limit problem, make sure this relay is operational.
• there is normal and maintenance mode.
• Maintenance mode does not run the patriot application program. It just runs in speed loop.
• You can't read the position of the maintenance,normal switch remotely.
• The elevation has two brakes. Either brake can hold the telescope by itself.
• motor brake
• The brake on/off is sensed by a current through a solenoid that holds the brakes open
• main brake.
• There is a micro switch that tells whether the brake is on or off.
• the brake motion is only about 11 mil, the micro switch hysteresis is about 5 mil so the adjustment of the  switch doesn't have a lot of slop.
• Moving the telescope with the brake on is only a warning. This is from the false readings from the main brake microswitch. The (my) software should be setup to stop motion whenever this warning occurs. running with the brake on down will not trip the motors. running up in elevation with the brake on will probably trip the motors.
• Azimuth has a turn count. The patriot software running in normal mode  keeps track of the wrap when you run.
• If you move to maintenance mode, the turn count is not used so it is possible to move the telescope to a different turn. When you move back to normal mode, there will be a turn error. The solution is to go back to maintenance mode and move the telescope back to the original turn count.
• The controllers have modules with mark's software running in them. If we buy a spare module (later on ) we need to find a way to load mark's software into the module. Cobhan is looking at a means to do this.
• in normal mode we've requested that we can automatically rest the contactor without having to be there.
• In the morning turning the power off then on caused:
• the  UV under voltage error was reset ok. This closed the contactor.
• The controller : voltage trip also reset automatically. this is incorrect. We should have to send  a reset command to clear this. Mark said he'd fix it.

12meter page
home_~phil