Transcript ECAL_HCAL_FEB

Status of FEB: Missing boards
•
3 boards are still under repair and test
•
•
•
•
•
 165 back from change of FPGA => to be tested
 270 bad FPGA =>further test => Laudren
 216 short circuit => to be localized
17 boards used for CROC test => liberate 9 boards in 2 weeks?
4 weeks?
Board repair delayed because of more serious problem
discovered on serialiser DS90CR215
January 15th 2008
1
Jacques Lefrancois
DS90CR215 problems
•
•
•
•
•
Problems first seen when we tried to increase the trigger
sensitivity by using a X4 in the trigger calibration constant
One board (out of 13) delatched => we thought that this board
(075) had a problem => back to Orsay
We found that in some case the amplifier noise when present
on trigger output (because of X4) => increase of 500ma in
delatcher current => in some case delatch (closed to limit)
A burnt finger showed us that the excess current was going in
one serialiser DS90CR215 (M24) used for sending trigger data
from TRIGPGA to validation board
ON M24 8bits of Et Max, 5 bits of address, 8 bits of Et sum for
pizero trigger
January 15th 2008
2
Jacques Lefrancois
M 24 problems => test with TESTRAM data
•
•
•
•
•
No problems on other serialisers
Problems with a few bits change with a period of 100ns (no
problem with period of 200ns)
problem located on EtSUM bits
D. Breton suggestion of problem when exceeding 0-3.3 volts
range
DS90CR215 sensitivity measured on test board with removable
TRIGPGA
•
•
Positive overshoot innocuous
Negative overshoot in "protection diode" is the culprit => DC test
•
•
•
•
-0.4 volts 0.7ma
-0.5 volt 4.5 ma
-0.6 volts 9ma
-0.65 volts 13ma limit => excess current in M24 (500 ma) => 80 °C
January 15th 2008
3
Jacques Lefrancois
Pulses on Serialisers
•
Case of no problem
January 15th 2008
4
Jacques Lefrancois
•
Still no problem (2-3 serialiser input in //)
January 15th 2008
5
Jacques Lefrancois
•
Case of problem
January 15th 2008
6
Jacques Lefrancois
Conditions for problem
•
•
•
•
Pulse must be fast (AX vs APA)
Rep rate must be high enough
No series resistor (in some cases 33 ohms in series => no
problem)
Only one serialiser input for a given TRIGPGA output
January 15th 2008
7
Jacques Lefrancois
Consequences
•
•
•
•
•
•
•
•
Chip operate but may die after some time? Delatch in some case
Check case of other users (SPD VFE, PS FE)
Enquire to a contact in Analog Device => suggest use of new,
faster, compatible, chip: DS90CR217
In DS90CR217 specs it is mentioned latchup tolerance 300ma!!!
Few components ordered for test
Other possibility =>solder capacities to slow pulses (8 per
boards, difficult to solder)
No use of X4 on trigger path in 2008 ????
Modification in 2008-2009 shutdown????
January 15th 2008
8
Jacques Lefrancois