here is the e-mail/comments for this post -- wwj

Ciao --

I was looking more at Willie Leight's most recent update to the BSMD
calibration studies he posted at:
http://drupal.star.bnl.gov/STAR/blog/wleight/2010/aug/29/bsmd-calibration-version-2

We discussed on the EMC² phone the early turnover from linearity and
what might be responsible for the behavior above ~ 5-6 GeV. BTW, that
region by itself looks to be pretty linear ... so I went ahead and "drew"
a straight line there calculated by

a) selecting the initial "turnover" point in measured electron
momentum vs. cluster ADC sum (picked by eye)

b) assuming the central strip begins ADC saturation there!! (e.g.,
central strip is at 1024 for 10 bit ADC)

c) using the Willie determined slopes (ADC/GeV) for each eta region
for the continued behavior of the other (non-saturated) strips

d) picking a reasonable value for the energy fraction of the
central strip (e.g., ~ 0.60-0.65)

Ok, the resulting curves (sorry I had to use the "line" function of
Foxit to draw on the original .pdf file) are posted below in the attached
(and modified full .pdf) ... see pages 13 and 14.

While my method is admittedly a bit crude, the message is that the
_dominate_ behavior of the region above the turnover point in Willie's
plots seems _very_ consistent with central strip ADC saturation!

If this is a correct interpretation, then single strip ADC saturation is
occurring at ~ 3.25-3.5 GeV, quite a bit lower than previously "guessed"
and as in the DB. The corresponding calibration would be
~ 3.25[3.5]/(1024-130) = 3.4[3.9] MeV/chn compared to the value ~8.01
MeV/chn in DB at eta = 0.

Comments welcome.
Cheers,
-- Will j.

PS: it is not crazy as I seem to recall a factor of 2 or more uncertainty
was associated with the initial calibration estimates.

**** here are the first order straight line parameters/calc ****

BSMDE e_low cluster willie %cent e_high calc*
GeV sum low slope strip GeV sum high

<.25 5 900 201 65% 15 1600

.25->.5 5 950 204 65% 15 1660

.5->.75 5.5 1000 206 65% 15 1690

>.75 6.0 1350 238 65% 15 2100

BSMDE e_low cluster willie %cent e_high calc*
GeV sum low slope strip GeV sum high

<.25 5.5 800 140 65% 15 1265

.25->.5 5.5 800 146 65% 15 1285

.5->.75 6.0 950 162 60% 15 1535

>.75 7.0 1000 139 60% 15 1445

* calc_sum_high = sum_low + (e_high - e_low)*(1-%cent/100)*slope