Problem fixes

• 191126 - loadcell 1 td8 problem was a loose cable in the hoffman box up by the a/d converterr(or signal conditioning board).
  

    The load cells were re calibrated by luis,willie, and osvaldo around 13:00 on 15Nov19. Prior to the calibration:

• td12 loadcells had a difference of about 5 kips (when the avg tension was 20 kips)
• td4 :  ldcell1 had a +10kips offset.
  
• td8 : loadcell 1 looked to be jumping.

td tensions 11-17Nov19

The first plots show the td tensions for 191111 thru 191117 (.pdf)

• the calibration was done 191115 around 13:00
• You can see how the tensions varied before and after the adjustments.

Calibration results table

The  table shows the results of the calibration (.pdf)

• The calibration removes and dc offset, and then sets the gain from a known resistance from the load cell calibration sheet.
  
• The pre column was measured before any adjustments. the pos column was after the adjustments.
• the pre column on the cal off row shows the dc offsets prior to the adjustments.
• Historically we've had large dc offset drifts.
  
• The red boxes show problems.
• td4 ldc1 calon tension:
  
• The calon,off was set to the correct voltage, but the little star did not read the correct kips (.02 kips/cnt)
• This must be a problem between the amplifier output and the a/d results.

Load cell differences before,after calibration

Td tensions and load cell differences for 15Nov19 (.pdf)

• Page 1: ldcell tensions and ldc1-ldc2 differences
  
• top: td tensions before the calibration
• 2nd: ldc1 - ldc2 before the calibration
• 3rd: td tensions after the calibration.
• td8-1 stayed close to 0 after the calibration.
• bottom: ldc1 - ldc2 after the calibration.
• The red lines are linear fits to the ldc1 -ldc2 plots
• Page 2: linear fits to ldc1 - ldc2
• The 0 crossing is the relative dc offset for the 2. the slope is the relative gain error between the 2.
  
• top: td12
• the calibration fixed the 5 kips dc offset
• at 34 kips the difference remained the same -7 kips
• middle: td4
• The 10 kip offset was corrected (ldc1)
• The slope error remains the same (since the on-off voltage remained the same pre and post.
• bottom: td8
• no dc offset pre post,
• the large slope is from ldc1 remaining at 0.

Gain, offset difference moving td's 3 -> 22 inches

    On 19Nov19 (after the load cell calibration) the tiedowns were driven 3 to 22 inches and back at constant velocities

• The ch was at stow and the dome was at 8 deg za
• this should be close to the azimuth arm balance position.
  
• the pointing routine tdcalib was used (helpdt tdcalib).
• it drove the tiedowns for 3 to 22 inches at .06 inches/sec.
• This should have put a uniform load change on the 3 tiedowns.
• It can be used to test the loadcell calibration.

The plots shows the tension change as the td's moved 3->22inches (.ps) (.pdf)

• Top: the  td tension for the motion.
• the colors are td 12(black), td4 (red), td8 (green)
• The solid lines are load cell1. the dashed lines are load cell2.
• You can see that td4 load cell 1 is way off in tension.
  
• middle: plot loadcell 1 - loadcell 2 for each td during the moves vs the average tension in a cable.
  
• the blue lines are linear fits to the differences. The 0 crossing would be the dc offset.
  
• Bottom: linear fit to td tension vs time
• a linear fit was done to each ramp in the top figure.
• The slope is tdTension/second. This can be used to check the gain calibration of the load cell amplifiers.
  
• Since the velocity was constant it was also: tdKips/.06 tdinches (we moved at .06 inches/second)
• Since the rate was the same for all td's, the slopes should have been equal
• The plot shows the slope/.06 to give tension per tdInch. The lope of the downward ramp was multiplied by -1.
  
• black lines: load cell 1
• red lines: load cell 2
• td4 load cell 1 is way off.
  
• the load cell 2 slopes are very close together , the loadcell 1 slopes are jumping around
• On   01dec17 we also measured that the ldcell 2 gain were all the same while the ldcell 1 gains differed
• My guess is that there is something going on with the little star reading the a/d converters.
• Using the average slope for load cell 2:
• Moving the tiedowns 1 inch changes the tension by 5.479 Kips
• There are 1.7 td inches/platform inch.
• We get 9.3 Tdkips/platform inch of motion

Summary:

• The loadcells were recalibrated on 191115.
• Problems that persist
• td4 loadcell 1 is way off.
• When the caldeflection voltage was set correctly , the kips reading was still off.
• voltage dif expected: 2.459 V, voltage dif measured: 2.458,
• kips expected: 80.56, kips measured 60.58
• This is a connection problem ldcell amp to a/d, or an a/d reading problem
• td8 loadcell 1. is intermittently  dropping out (for hours at a time)
• it then jumps back to the correct value
• a constant velocity ramp 3->22inches and back was used to verify the gain settings of the tiedowns.
  
• all of load cell 2's look to be calibrated correctly
• We've seen this back in 01dec17 that loadcell 2's are all the same, but loadcell 1's are off. this may be a little star problem.
• Using loadcell 2 values, we get 9.3 tdKips/PlatformInch.  This is close to what we've gotten historically
• I think the original design spec was 8.4 kips/platform inch (although we've always been a bit off from this).
  
• a vxWorks macro: tcalib   can be used to drive the tiedowns from 3 to 22 inches and back
• probably best to put dome at za=8 deg before starting
• see helpdt tdcalib  for a description.