|
Rods that arrive from Helsinki are stored in a clean cabinet. They
still reside in their aluminium shell. There is nothing like a
reception test, the big crate is being emptied and the contents will
be stored in the cabinet. The ShockLog is read out by Antti and the data are
analyzed to reconstruct the history of the shipment.
|
|
|
Here are all the components being stored (Interconnect bus,
interconnect cards, Optohybrids, CCUMs, cables, ....).
|
|
For all parts and before mounting, a visual inspection is done to ensure the quality of
the components.
|
|
|
All components are registered in the database. The mounting is guided by a rod
assembly plugin to the Tracker
Construction Database Browser that has been created by Matteo. It ensures
the correct mounting of the rod, i.e. it only allows to register the correct
types of optohybrids with the given rod frame. It also tells how to order the
fibres in the fibre holder.
|
|
After the components have been verified, the assembly starts with
inserting the bus into the frame (see picture). Then the interconnect
cards are mounted and the HV cables connected. Afterwards, an
electrical test will be performed, assuring pin-to-pin connections and
detecting shorts. Then the CCUM and the Optohybrids are mounted.
|
|
|
The rod is then moved to the functional test
setup. There, hybrids are attached to the ICC cards and the rod is
being connected to a CMS-like readout system that comprises one TSC, one FEC,
one FED, two O-FED's, one KMUX with three MUX cards, a CMS Tracker
Test power supply, one FEC2CCUM card as well as cables and fibres...
|
|
In the functional test, the I2C connectivity and the laser output of the
Optohybrids is being verified, as well as some other basic checks that cannot
be performed by the electrical test. If the test fails, components need to be
replaced and the test must be repeated. If the test succeeds, the rod will be
stored.
|
|
|
After 40 rods have been cabled, a shipment to UCSB or Fermilab will be
done. There, modules will be mounted and the rod will be tested heavily as a
whole. You can look at a movie showing how a module
is built, tested and mounted on a rod. The tested rod is then shipped back
to CERN. After a reception test, it will be installed in the TOB wheel.
|
The production test tool is based on the program called XROD that was
developped by Juan Valls Ferrer. This program was mainly a debug tool and not
suitable for a production test. This is why I had to heavily modify it to make
it a real production test. Since the
last version Juan supplied was not even compiling, you can download a fixed version (1.7) containing
small modifications to the program (but not containing the full production
test).
XROD depends heavily on other installed software, please take a look at the
page from Juan.
If you want to compile the production test yourself, you have to download and
install the software package for the
Karlsruhe multiplexer.
After downloading the software package, install it via
cd ${HOME}
tar xvzf xrod_v17.tar.gz
and put the file xrod_v17/xrod_new somewhere into your path.
Please be aware that there is a nasty undocumented feature in the program:
If your user name is not xdaq, you need to change the file xrod.opt before you
can use xrod - it is a binary file, you need to preserve the exact byte offset
in that file. Use e.g. emacs or vi.
Once you have installed all software necessary for the operation of XROD (see
paragraph "A foreword about XROD" above), you
can compile and run the latest version of the production test program. If you
are working as user xdaq on pctobrod2, you will find
all current code in the directory
/home/xdaq/mkw/ProdTest/ProdTest. The code is maintained in CVS,
so the recommended way to install it is to setup cvs with export
CVSROOT=/home/xdaq/cvs and chen check out the code with cvs co
ProdTest. The whole directory tree /home is being backed
up daily with the help of the TSM™ backup
client, so in case of a harddisk failure you can always recover the source
code and all rod measurements.
A "documentation" of the source code is available if you run
doxygen in the src/ subdirectory. After doxygen completes (takes
a while), a directory named html/ appears. A good starting point
for browsing is html/files.html.
I have written a draft CMS internal note
describing the production test (also known as "functional
test"). It also describes what results you will find after the test has
completed and a FAQ.
The production test program is running on the account
xdaq@pctobrod2. The main directory is ~/ProdTest,
where all the subdirectories containing the diffent files can be found.
The DCU calibration files are residing the directory
~/ProdTest/dcu. The files can be copied there from any UNIX
account with help of the scp command:
scp *.xml xdaq@pctobrod2:./ProdTest/dcu/
(It is assumed that the calibration files reside in the current working
directory on the machine where the scp command is
executed). After this has been done, the calibration files are accesible for
the production test program.
On each rod of type 1 (first rod in a control ring), a HMX2000 humidity sensor
from Hygrometrixis installed to
allow the humidity measurement during CMS operation. I have evaluated the
calibration data from the company, derived calibration constants and supplied
source code for the computation of the
calibration constants as well as the computation of humidity readings from
sensor outputs. Together with Matteo Risoldi I have established a procedure on
how the sensor calibration data are stored in the Tracker Construction Database,
and Matteo has modified the rod
assembly plugin to ensure the correct mounting and data upload of the
humidity sensors.
We have written a CMS internal note
(draft!) describing the calibration procedure and rod assembly plugin
interface. |