Performance_jw

advertisement
Performance issues
Parameter
Extreme
Operation
conditions
Accuracy of
individual
measurement
Online:
TGC
Comments
MDT
Comments
RPC
Comments
MMGAS
Total rate 14 KHz/cm2,
Charged particles : 300 Hz cm^2 (from CSC measurement 10Hz/cm^2 at 3*10^33)
Photons:
Neutrons3.105 cm2/s ?
Using ToT
~100-150
um ?
<100 um for
20o
Layer to L.
alignment??
Can be
compliment
ary to MDT
Online track
segment
accuracy
<1 mrad
<60 um
Numbers
depends on
the angle???
1mrad
~60 um?
Timing
RMS 7ns
97% in 25
ns
~50 ns
100% (3BX)
for offline
~3ns from
offline t0
100% in 1 BX
~100 ns
~5 ns
Offline:
Dead time
Comments
~160 um
at Max
occupancy
Is it enough
to provide
60 um
accuracy~
If the same
approach
used as
proposed
for MDT
upgrade
~300 um
online with 2
mm strip pitch
Estimated
No data?
~ 1 mrad
~150 um
(standalone)
~300 um
<150 um
~1 mrad
60 um
RMS ~1 ns
with mean
timer.
100% 3BX
sliding
window
<100ns/
-From Strip
pitch
-From micro
TPC fit to the
strip number
and the time
With the 300
it is difficult to
get to 60 ?????
Off line 1 BX
in average
particle/chip
Dead time and
efficiency drop
induced by
sparks
Not an issue
Spark is
not an
issue
Spark is not an
issue.
Minimum
number of
measurements
fro trigger
6 =2*3(out
of 4)
Redundancy
at high rate?
Using Pads,
max pad
occupancy
~2%
Is it enough
For hit
efficiency of
80%?
Can be complimented by
MDT trigger logic
?
4=2*2(out of 3) 100-1%
Redundancy ?
Can be complimentary to
the other trigger systems
Can be complimented by MDT
trigger logic
Decision is based only on
MMG detectors
Failures
1. Sparking will lead to an
exclusion one plane.
2. One link-one plane
Sparking leads to
exclusion of 2(?) layers.
Very unlike event.
Expected fake
rate
No study
yet
No study
yet
1. Sparks not a problem. Mainly
mechanical defects (1 layer)
2. One link-one tower
(redundancy?)
No study yet
1. Sparking will lead to an
exclusion one plane.
2. One link-one planeredunducy?
No study
Fake rate can
yet
be emulated on
the basis this
System
Redundancy
Generally the
same but
depends on
the area
affected by
the voltage
drop
6 out of 8
Is it enough
for occupancy
chip
occupancy
(22%) at mzx.
rate.?
Redundancy ?
year data by
overlap
method
Maximum
Occupancy
Pad ~2%
Strips in the
inner region
~0.6%
Efficiency at
maximum rate
per layer.
85%?
-3% (25ns)
-1% (dead)
= 81%
Dead area
1%
Rate for
inner Pads
700 KHz
Rate for
inner strips
250 KHz
Seems to be
confirmed
by latest
results
~23%
Between
chambers?
~3%
Rate 1
MHz/tube
Bandwidth?
?
70%
400 um
tube walls
Between
chambers
?
Rate 50*2cm
strip is ~1.4
MHz
Occupacy 3.5%
99%
Voltage drop
affects efficiency
will require
additional 200V
of operating
voltage
1%
Area affected
Not an issue
by signal/spark
Non an
issue
Not an issue
Critical rate of
sparks
Not an
issue
Not an issue
Not an issue
0.5%
occupancy for
5ns
~22% per
chip per BX
Seems OK
since margin is
> 1000V
~77% per plane (22%
occupancy + 1% pillar+…)
can be compensated adding
more a few more plans in the
high occupancy region if
needed.
Between
chambers?
1%
Area
affected by
spark ~ few
cm is dead
for 100 us?
>1000
Hz/cm2 ??
3.2*50
cm2*14kHz*10
0 ns (drift)
Between
chambers?
At 3 Hz/strip
inefficiency
0.003%!?
Seems large
safety margin
Experience
Large chambers, Large
system
Large chambers, Large
system
Trigger
demonstration
Main
performance
concern
Advanced but still under
development
Trigger efficiency ????
Single layer efficiency at
high rate 81% at high
rate? (85% rate effect, 97%
in 25ns, 1% dead area +
area between chambers in
one layer???)
Beginning of
developments
1. Is occupancy an issue?
(for 8 tubes only 1-2
have background hits)
- Seems no participial
problem for 7*1035
But is it enough to
provide <60um accuracy
standalone
(160um/sqrt8= 60 um if
everything is OK)
2. Is Bandwidth enough
for reading out the inner
ring with tubes running
at >1 MHz??? Margin?
Large chambers, Large system,
But significant change (cathode
material, gap)
Beginning of developments
Not yet in the stage of
operating large scale
prototype
Beginning of developments
1. Can operate without mean
timer?
Very good time resolution but
not clear if it is really needed?
1. Trigger efficeincy:
at high rate 77% sigle pane
efficeny (issues -> number of
layers, Fake rates? Processing
time? Bandwidth)
2. Space resolution for inclined
tracks using uTPC concept (still
not demonstrated) but assuming
160 um then same concern as
for the MDT
2. Is 300 um is achievable at the
trigger level?
The efficiency at 10 KHz/cm^2
has not been measured and it is
not clear that the detector can
stand this very high rate
3. Is number of layers sufficient?
Can not compliment MDT space
point accuracy
4. What fraction of the detector
is read out??? Bandwidth
issues???
Download