Repaired version of the awk script for Himali: forHim.awk
ME+4/TP2: Document UR-0003 seems to have a different numbering from UR-0059. Unfortunately it is not possible to compare very many points: the targets on the reference DCOPS are present in both documents, as are the targets on the reference laser mount. The first set of measurements for the targets has 14101, while the second doesn't. The first uses a 5mm target while the second (and UR-0059 ?) uses a 20mm one for a difference of 15mm.
As near as I can tell, the point represented by 14201 in UR-0059 is represented by 14203 in UR-0003.
The B prefix means Bench measurements (first set), while the others are from taking the UR-0059 and transforming them into the nominal Cocoa transfer plate coordinate system.
Then I flip the Y and Z signs for the bench test numbers, so that the coordinate systems will be comparable, and subtract off the position of 14209 (close to a DCOPS dowel) in the appropriate coordinate system.
Notice that the only point close to 01 in the Bench system is B03. The local X and Y agree quite well, but the Z is rather far off. 01 has a Z of -30.93, which needs to be corrected by the 15mm difference in targets to give -15.93, to be compared with the -14.51 for B03. This represents a change of 1.42mm (over a distance of 288.8mm from the reference center), or 4.92mrad.
B01=[-91.770000; -76.810000; -14.460000] B02=[-208.79000; -116.32000; -14.530000] B03=[-288.80000; -21.290000; -14.510000] B04=[-185.56000; 4.6400000; -109.60000] B05=[-214.49000; 28.610000; -109.40000] B06=[-243.38000; 4.5800000; -109.45000] B10=[29.870000; 100.05000; -37.980000] 01=[-288.66192; -21.280000; -30.932808] DCOPS points B07=[-57.840000; 0.040000000; -0.13000000] 07=[-57.779560; -0.43000000; 0.035665769] B08=[-28.850000; -0.17000000; -24.100000] 08=[-28.899958; -0.57000000; -24.091022] B09=[0.E-26; 0.E-26; 0.E-27] 09=[0.E-28; 0.E-27; 0.E-29] 11=[-51.305101; 1.0600000; -101.32339] 12=[-18.002134; 0.74000000; -55.004556] 13=[10.828431; 0.71000000; -101.80101]
For ME+4/TP3 I have a similar situation, except that the labels seem to agree this time.
B01=[34.130000; -76.780000; -14.690000] B02=[151.15000; -116.28000; -14.740000] B04=[128.13000; 4.8100000; -109.81000] B05=[157.06000; -19.160000; -109.73000] B06=[185.95000; 4.8700000; -109.69000] B10=[-29.960000; 100.08000; -37.980000] B03=[231.14000; -21.250000; -14.720000] 03=[231.13199; -20.630000; -30.313668] B07=[-57.830000; 0.090000000; 0.16000000] 07=[-57.919116; 0.12000000; 0.021213203] B08=[-28.970000; 0.010000000; -23.910000] 08=[-29.019662; 0.040000000; -24.055773] B09=[0.E-26; 0.E-26; 0.E-27] 09=[0.E-28; 0.E-27; 0.E-28] 11=[-54.168484; 0.22000000; -75.047995] 12=[-21.528435; 0.15000000; -104.60506] 13=[14.272381; 0.22000000; -74.263107]
Here the Y coordinates differ by .62mm, and the Z coordinate by .59mm instead of the 1.42mm above.
For ME+4/TP1 I have for point 14103
B01=[51.240000; -70.410000; -14.910000] 01=[50.739670; -70.230000; -31.801207] B03=[466.24000; -21.290000; -14.820000] 03=[466.40984; -21.280000; -29.199190]
Which gives a difference of -0.62mm over a distance of 466.3mm.
I tentatively conclude, given the spread of values, that the Z shift does not represent a systematic tilt--the model is correct even if the mounting is not perfect. Unfortunately I'm not sure how well I can reconstruct the tilt, since I often don't have many points at large enough lever arm--sometimes only one. Notice that the point 14101 has quite a large Z shift: 1.9mm; and yet it is close to the reference point. This makes me think of target seating errors.
The log file for an event after power cycling the ME-3 low voltage: 1267557834.log
A 1267557834.dat Cocoa input file
And the profiles:
Power cycling the low voltage does not help. The lasers work fine when run by hand, but not always under program control. This suggests that possibly contention is causing some delay in the command reaching the laser controller. It is mediated via the DIM server, which is sometimes CPU-bound. In fact the DIM server is the same machine the DAQ runs on, which sometimes is running at 98% CPU and uses the network heavily in bursts.
Modified 02-March-2010 at 14:43
http://hep.physics.wisc.edu/~jnb/cms/02Mar2010
| Previous notes | Next notes | Main slide directory |