# Run 9 Calibr & QA

There are several things collected w/ this entry:

**********************************************************

I. 1st DRAFT OF BPRS HV SETTINGS

**********************************************************

Recall Jan's note summarizing his work on BPRS MIP response some
months ago:

http://www.star.bnl.gov/HyperNews-star/get/emc2/2895.html

An outgrowth of that was to try and figure out how to set the
HV for run 9. I dumped Jan's tube-by-tube response numbers into
a spreadsheet and came up with (as a draft for discussion) the
following (e.g., this blog entry and attached _version_2_ of the
the new columns in the spreadsheet are the ones with the bold labels
at the top within "boxes". Some notes are:

0) first columns is voltages set previously to "equalize the grass"
(these are read off HV setting files generated for the LeCroy
supplies)

1) power law HV extrapolation (column "H")

used power law from Julie/MIT EEMC QA (via Jan) for gain
gain_1/gain_2 = (1HV_1/HV_2)**x, with x=8.3

Use to put average MIP response of tubes in desired chn (here
chosen to be 15 ... selection of chn 20 results in column "J").

2) One RDO crate (boxes 39 to 53 and yellow shaded in spreadsheet)
seems quite low in response by factor ~ 8/14.4 = 0.56
(eyeball averaged)
If assume "resolved "-> multiply av MIP adc by fac =1.8 & then
compute HV needed (columns "I")

3) One way of choosing "final" values in column "K" is following:
a) If MIP adc is > 15, retain original voltage, if not use power
law HV set for adc chn 15
b) If too few pixels (generally means tube replaced), set voltage
to 985 (start slightly higher than average)
c) For low crate region (boxes 39 to 53), if "adjusted chns" power
law HV is less than original, use original ... otherwise use
associated "adjusted chns" power law HV values
d) Limit tube voltage to ~1050 (vendor suggested max is 1000)

4) ok, there are some more notes and items on the spreadsheet.

This is obviously all for discussion: how to set (e.g., how to go about
items in "3)" above) and how to check as we start running.

****************************************************************

II. BTOW TIMING SUMS w/ "FLAT TOP FIT""

****************************************************************

Timing curves for BTOW were taken and analyzed by Alice B.:
http://drupal.star.bnl.gov/STAR/blog-entry/aliceb/2009/mar/08/btow-timing-runs-10065031-45
followed by discussion/preliminary conclusions drawn by Steve T. in thread:
http://www.star.bnl.gov/HyperNews-star/get/emc2/3045/2/1/2.html

I believe one finally sees evidence for a real flat top region
(expected!) in the "sums spectra" for the timing curves  -> to indicate this
I did something simple with the "line function" on the .pdf file and saved
results in the attached file below: "btow_timing_flat_top.pdf".

While not particularly optimized, I uniformly (_all_ crates) took a
region ~ 32 to 42 ns (delay) as a flat eyeball/calc average of relevant (4)
points. I indicate this average with a horizontal line covering the relevant
delay region ...  and then just connected with a straight line at either
side to indicate the falloff. I think this describes things fairly nicely
(e.g., most crates quite well, with a few not as good). It seems that the
"Gaussian fit" (also shown on same plots, has a mean systematically low
w.r.t. flat region of these new curves I have sketched. [Note: the Gauss
fits that Steve refers to in above thread are for the _slopes_ which
seem to be much more "washed" out and don't exhibit this marked flat
region)

As to where one wants to set timing in the above "flat top" picture ...
clearly one wants to back off a safe distance (~ 3 ns jitter?, etc.) from
relevant (large delay) edge as well as account for apparent crate to crate
shifts or few ns?

Here is an update of the "DRAFT OF BPRS HV SETTINGS" posted above.
The updated spreadsheet is attached below as:

The main difference regards the treatment of boxes 39-53. It was pointed
http://www.star.bnl.gov/HyperNews-star/get/emc2/3041/1/1/1.html
that the boxes supplied by the 2nd LeCroy unit were set to -900 V instead
of the (higher) box specific voltages as per the 1st supply.

Indeed a check from "hvlog" for 2008:

/export/home/users/sysuser/psd/hvlog/2008
[sc3.starp.bnl.gov]:~/psd/hvlog/2008>ls -l
-rw-r--r--   1 sysuser  sysuser    13515 Jan  2  2008
psd_hv1_2008.Jan.02-10.02.27.tex
-rw-r--r--   1 sysuser  sysuser    13515 Jan 28  2008
psd_hv1_2008.Jan.28-14.28.09.tex
-rw-r--r--   1 sysuser  sysuser     2785 Jan  2  2008
psd_hv2_2008.Jan.02-10.02.31.tex
-rw-r--r--   1 sysuser  sysuser     2785 Jan 28  2008
psd_hv2_2008.Jan.28-14.28.43.tex

-> For hv1 the "wwj" box specific HV are loaded starting on 2 Jan, 2008
[boxes 1-38; 54-60]

-> For hv2 a _constant_ -900 V is applied starting on 2 Jan, 2008
[boxes 39-53]

Since Jan's analysis encompasses days 43-70, 2008 -> the reduced settings
for supply #2 are the relevant ones.

NOTE: this also explains the "low response" for the RDO that reads out
these _same_ boxes. It is not therefore bad timing, bad crate or any of the
other speculations.

So, updated spreadsheet has "900 V" as starting voltage for boxes 39-53
(still the yellow highlighted region). The column artificially increasing
the response for this crate is _removed_ as is the particular selection
criteria for final voltage for this region (the final select criteria is
now uniform for all boxes).

A further notational change is that the column "map of HV chn to box
position" specifically indicates the box tube position coupled to the
to the right give indeed the affected channels (e.g., Jan has already
the swap incorporated). So the numbers to input to the HV files for loading
into the LeCroy are indeed just as indicated in order in the column "FINAL
HV", _except_ for box #25 where two cables were swapped during pre-run9
maintenance.

There are some updates also to the spreadsheet notes to (hopefully) clarify.

**********************************************************

IV. ETOW TIMING SETTINGS

**********************************************************

Attached (e.g., tower-crate-1_2_set.pdf, etc) are annotated crate-by-crate timng

scan plots from Alice B:

http://drupal.star.bnl.gov/STAR/blog-entry/aliceb/2009/mar/11/preliminary-eemc-timing-curves-crate

indicating the proposed settings for ETOW in run 9. The numbers in terns of TCD delay

(up a ~ 1 ns from run 8) are:

crate 1   56

crate 2   56

crate 3   43

crate 4   43

crate 5   42

crate 6   43

Here are detailed notes and final Run 9 box delays (and Run 8 for reference):

*****************
RUN 9 ETOW TIMING
*****************

TCD        box                      leave    new        config
scan        scan                    phase    box        file
peak       effective               at 21    delay      HEX
run 9      delay

#1   56       0X5A ->      -16.6   -21     18.4   #1   0X12
(90-106.6)
#2   56       0X50 ->      -26.6   -21      8.4    #2   0X8
(80-106.6)
#3   43       0X42 ->       66       -21     88     #3   0X58

#4   43       0X3C ->       60       -21     82     #4   0X52

#5   42       0X17 ->       23       -21     44     #5   0X2C

#6   43       0XB  ->        11       -21     33     #6   0X21

########################################################
RUN 8 SETTINGS     TCD    BOX SCAN       PHASE   SET

#1  0X12   (18)    |   55      0X5A ->   -16    -21      18
|             (90-106)
#2  0X8    (08)     |   55      0X50 ->   -26    -21        8
|             (80-106)
#3  0X57   (87)    |   42      0X42 ->     66    -21      87
|
#4  0X52   (82)    |   43      0X3C ->     60    -21      82
|
#5  0X2B   (43)    |   41      0X17 ->     23    -21      43
|
#6  0X1F   (31)    |   41      0XB  ->      11    -21      31