E1031 alpha calibration

Here's what the real measured energies from the triple alpha source look like in TUDA as used in July.
geometry:
upstream LEDA is 45 mm from the target
downstream LEDA is 150 mm "
(downstream) S2 is 170 mm "
dead layers: 0.33 um for LEDA, 0.39 um for S2: explained in the previous entry: these are the values I needed to reproduce the SRIM results for 0.4 and 0.8 um respectively.



(A note about the figure: the blue points are the actual values for individual alpha events. The pink points are a clever way of using those values in the code to automatically calculate average energies and angles for each strip...except that as you can see they don't always give the right values for the average energies. I have absolutely no idea why this is so, and particularly why it's only the S2 energy where this problem shows up. However, the LEDA values seem to agree well with the actual averages for each strip, so I've used those automatically-calculated averages below to calculate E(alpha) as a function of strip number.)

For the downstream Leda and the S2, the angle dependence isn't too bad, but for the upstream Leda it is. I suggest using the following linear fits:
For downstream LEDA:
"5.155 MeV" energy = -0.001144 * strip number + 5.0871
"5.486 MeV" energy = -0.001047 * strip number + 5.4211
"5.806 MeV" energy = -0.000988 * strip number + 5.7440
For upstream LEDA:
"5.155 MeV" energy = -0.006499 * strip number + 5.0563
"5.486 MeV" energy = -0.006187 * strip number + 5.3923
"5.806 MeV" energy = -0.005906 * strip number + 5.7168
and for S2:
"5.155 MeV" energy = 5.0089
"5.486 MeV" energy = 5.3475
"5.806 MeV" energy = 5.6741

A note on LEDA strip numbering: the strip numbers are calculated here as
DLstrip=(int)Math.floor((rDL-50)/5);
that is, the numbering goes from the inner radius to the outer radius. I'm pretty sure this is the opposite from the numbering in the actual data, but it'll be easy to convert.
Update: weird. It turns out the numbering is the same in the simulation as in the data. How'd that happen? I must have checked the data analysis code when I was writing the simulation code, or something crazy like that. Anyway, in the simulation the distance from the beamline is defined as
DLrad=5*DLstrip+52.5;
whereas in the data analysis code it's
leda1_y = 5.25 + ledastrip1*0.5 (plus a randomizing factor)
which is the same, except for the change of units from mm to cm.