[d2n-analysis-talk] BPM Calibration Discrepancy: BigBite vs LHRS

[Diana Parno]

Hello Vince and Brad,

I see an odd discrepancy in the BPM calibration (from the Bull's Eye Scan) 
in the BB and LHRS arms, and I'm hoping one of you can shed some light on 
it. Each arm reads in the BPM data separately, of course, and has a 
separate crate map, but I had expected that the signals in each arm would 
be copies of each other. Instead, I get very different results in each 
arm, leading to very different constants for transforming between BPM and 
Hall coordinates, which doesn't seem as though it should be physical!

Here are the coordinate transforms (4 rotation matrix elements followed by 
2 offset-vector elements) I get for each arm:

BPM A:
BB  : -0.614534 0.65991 0.61804 0.669433 -0.00322684 0.000669938
LHRS: -0.676308 0.695722 0.679473 0.705958 -0.000689187 0.000690562

BPM B:
BB  : -0.816705 0.557122 0.816639 0.559635 -0.00609878 0.00779028 
LHRS: -0.679277 0.675235 0.684231 0.674582 0.000348872 0.00133972

For BPMB, especially, you can see that the differences are quite large (up 
to 20% in the rotation matrix, and much more in the offsets).

The EPICS variables, which I used for the "absolute" coordinates of each 
point of the bull's-eye scan, are the same for the two arms. You can see 
where the difference comes from when you look at the scaler variables, 
e.g. urb.BPMA.rotpos1 or BBurb.BPMB.rotpos2. The numbers from the 
following tables represent the means of Gaussian fits to these scaler 
variables, from each run of the bull's-eye scan. BB run 1072 is equivalent 
to LHRS run 20071, and on up. You can see that the mean reading for a 
given point on the bull's-eye scan differs by two orders of magnitude in 
some cases. Only BPMB.X seems to yield fairly consistent readings across 
the two arms.

LHRS:     BPMA.X	BPMA.Y		BPMB.X		BPMB.Y
20071 	-0.00378398 	0.000246954 	-0.00376672 	-0.00127715
20072 	-0.000942709 	0.00273274 	-0.000640832 	0.00158012
20073 	0.00198571 	6.34531e-05 	0.00225074 	-0.00124078
20074 	-0.00105594 	-0.00282918 	-0.000790473 	-0.00419316
20075 	-0.000993419 	-2.83081e-05 	-0.000745336 	-0.00125726

BB:
1072 	-0.00629947 	0.00207323 	-0.0108113 	-0.00152777
1073 	-0.00301258 	0.00493221 	-0.00864019 	0.00187118
1074 	4.45575e-05 	0.00187431 	-0.00584566 	-0.00146506
1075 	-0.00313696 	-0.000932585 	-0.00873883 	-0.0051056
1076 	-0.00306551 	0.00176921 	-0.00871131 	-0.00148042

I've been assuming that these signals are supposed to be digital copies of 
each other; is that not a safe assumption? The numbers don't seem to 
differ by any constant multiplicative factor, even in the same column. The 
only thing I can think of is that the crate map might be wrong for one of 
these arms (my money would be on BigBite), but for the numbers to make any 
sense at all, it seems as though I'd still have to be reading from 
someplace correlated to the beam position (the raster current, perhaps?) I 
am also not sure where I would start tracking down the correct cratemap, 
if the DB files are wrong.

Any thoughts?

Thanks,
Diana