Transcript transparencies

CMS Tracker:
(A contribution to:)
Detector Control Units
&
Tracker Monitoring
Fatima Kajout
My Summer Student project
11th of August 2003
Student Session 2003
The CMS Tracker:
Front End Control and Monitoring
Silicon detectors
~10 000000 detectors strips
=> A lot of temperature monitoring to do!
Front End Readout,
Control and Monitoring
Student Session 2003
Units
DCU
Detector Control Unit
Measures and monitors slowly varying analogue signals;
i.e.: Temperatures, low voltages, detector currents,…
2*2 mm2 chip; power consumption  50mW
Contains an 8-channel, 12 bit ADC,…
Provides 2 bias currents for powering temperature
probes
Will be mounted on every module hybrid where it
will monitor: - Internal temperature probes, external temperature probes,
voltages and currents.
Student Session 2003
Agenda
Project Outline
Hardware and Objectives
Software and Objectives (1)
Software and Objectives (2)
Architecture Validation
Conclusions
Student Session 2003
Project Outline



The project aimed to build the tools needed to
collect DCU data, store them into a permanent
database, retrieve all or part of the data, and
finally display and analyze them. An evaluation of
the temperature and voltage measurement spread
was also made.
A prototype system with 12 DCU ASICs has been
installed and used for tests.
Software technologies used are: Java, C++, SQL,
mySQL and PAW.
Student Session 2003
Hardware and Objectives

Prototype tests: no system of such size (“of the
CMS Tracker”) can be assembled without
exhaustive prototype work, followed by scalability
studies that should validate the proposed
architecture even when the number of modules
grows by orders of magnitude, as in our case. I
have set up a system with 12 DCU chips as a
starting point. Performance measurements are
foreseen.
Student Session 2003
Measurements and calibration of 12 DCUs
We used a set of 8 very-well measured resistors which simulated one
thermistor at different temperatures (from –40 to 30 0C).
DCUcts
Calibration linear formula:
Rin = offset + scale*DCUcts
Rin
Steinhart-Hart equation:
1/T = a*ln(Rin)³
+b*ln(Rin)² + c*ln(Rin) +d
Temperature
Converting from DCU counts to Ohm
The
present plots show the offset distribution and the scale
factor distribution for the 12 DCU under test. The formula used
is, of course, Counts = Offset + Scale * Resist

 Although the spread is smaller, a nonClearly, there is a non-negligible spread in the
offset values. It is still to be seen whether this
uniformity in this calibration constant is as
influences the temperature measurement to the dangerous as one in the offset.
point of requiring a set of separate calibration
constants for each DCU.
At –30 and +300C, the RMS is ~ 1.2% of the mean value, i.e. for the same resistance value the
measurements would vary ~ 1.2% (which is ~.5 0C at +300C and less at lower temperatures.)
Test temperature distribution


This plot shows, for all 12 DCU
chips tested, the reconstructed
temperature for the 8 resistors used
to simulate thermistors. It is clear
from this plot that with increasing
temperatures the measurements
become less accurate. This is due to
the non-linearity of the thermistor
characteristic: the higher the
temperature, the less the value of
the resistance changes for an equal
temperature step.
Note that we will operate the CMS
Tracker at –20° C => nice!!!
Student Session 2003
Software and Objectives (1)

Database design: a proper definition of the
tables and their relationships is mandatory
in a project of this size. Therefore, I have
studied the problem and derived tables in
the third normal form, as recommended by
IT literature.
Student Session 2003
Database design
Measurement
1
1
N
N
1
Value
1
Timestamp
1
ref
Sensor Id
1
ref
Location Id
Each measurement entity comprises:
• a Value (an integer number produced by the analogue to digital
conversion process);
• a Timestamp (an integer number encoding the date and time at
which the measurement was taken);
• a Sensor Id (a reference to a Sensor entity in a separate table);
• a Location Id (a reference to a Location entity in a separate
table);
Student Session 2003
Software and Objectives (2)

Interface design: in order to be of any use, a database
system must offer interfaces that are easy to use and
effective. Given the present database structure, the very
first interface needed will be the one that allows to read the
list of existing sensors and add to it.
Sensor-Type Table
Thermistor 1
Thermistor 2
Voltage 1
We can:
• Consult
Current 1
Description
a=0.005;b=0.002;c=0.014;d=0.014
Algorithm
Steinhart-hart equation
Student Session 2003
Architecture Validation

The DCU prototype system will be also
used to validate the proposed general
scheme for the monitoring of Tracker
environmental parameters. In particular, one
should establish whether individual DCU
chips need calibration constants, or a single
set can be used without excessive loss of
accuracy.
Student Session 2003
Conclusions
What we wanted to do:
 Gather DCU DATA and
 Store
 Retrieve
 Display
 Analyze
How we accomplished it:
 Prototyping a system with 12 DCUs and
 Performed measurements and calibration (used
predevelopped software for the CMS Trackers)
 Designed a new, improved database
 Designed a graphical interface
Next steps:
 Performance evaluation
 Architecture validation
 Finalize database
Student Session 2003
Thank you!
Any question?
Student Session 2003