Difference between revisions of "Improved timing"
Line 179: | Line 179: | ||
{| border="0" cellpadding="2" | {| border="0" cellpadding="2" | ||
|width="300pt"| | |width="300pt"| | ||
− | [[Image:run_31710_4951_res20_thr.png|thumb|300px|49/51 Ar/CO2 gas mix Resolution ch 20, Hit threshold varied (n x pedestal sigma) ]] | + | [[Image:run_31710_4951_res20_thr.png|thumb|300px|49/51 Ar/CO2 gas mix Resolution ch 20, Hit threshold varied (n x pedestal sigma) , dt=tz]] |
|width="300pt"| | |width="300pt"| | ||
[[Image:run_31710_4951_thr4_res20.png|thumb|300px|49/51 Ar/CO2 gas mix Resolution ch 20, Hit threshold=4sigma]] | [[Image:run_31710_4951_thr4_res20.png|thumb|300px|49/51 Ar/CO2 gas mix Resolution ch 20, Hit threshold=4sigma]] |
Revision as of 15:58, 12 September 2011
- Work in progress - intended more as a logbook than presentation of results*
CDC prototype cosmics data
- 50/50 mixed Ar, CO2
- Recalibrated MFCs
- Outer plenum added to CDC prototype to prevent leaks
- New pre-production HVB
- 2050V
- Modified preamp4 (terminating resistors removed)
- Prototype tilted at 30o to enclosing scintillators (scints trigger the daq)
Fed scintillator signals post-discriminator and coincidence signal through then level converter to ECL, then attenuator, into fADC. Took threshold as half peak value (also tried 1/4 and 3/4 peak but these broadened the timing spectra). Used interpolation algo. (The DAQ is triggered by the coinc signal.) The time offset in the first set is equal to the (uninterpolated threshold crossing bin - 10) x 8 ns.
Drift time histogram (for all straws) and expanded about t0
T0 was taken as the x-intercept of a straight line fitted to the leading edge over 10% to 90% of the maximum value. Here this value is 91.3ns.
Drift time vs fitted drift distance using Garfield calculations for 50/50 Ar/CO2 (left) and 49/51 (right).
Resolution (difference between fitted track radius including and excluding that hit from the fit) for ch 20 (straw close to center of prototype)
Resolution for various time increment (offset) dt added to fitted t0, ie t0 = 91.3 + dt, ignoring events before t0
Optimum is offset dt=11ns for 49/51 (14ns for 50/50 has slightly worse resolution)
Check how the mean varies with drift dist around optimum dt
Use dt=14ns for 50/50 mix, 11ns for 49/51 mix
Re-assign events with time < t0 to a range of early times >= t0
Assign early events to t0+1ns
Resolution with drift distance for 49/51 mix, dt=11ns, early events assigned to t0 + 1ns
Earlier run with same timing setup but larger lower scintillator with angled light-guide gave very slightly worse resolution (optimal for dt=15 ns, early events assigned to t0)
Varied hit threshold. Resolution fitted with 1 gaussian has smallest width for thr=4sigma (previously 5 sigma) dt=10ns tfix=1ns but shows slight curve when plotted vs drift time or distance.
Resolution with drift distance for 49/51 mix, hit threshold 4sigma, dt=10ns, early events assigned to t0 + 1ns
More....
resolution vs drift is slightly curved
resolution does fit better with 2 gaussians but does not help to find minimum, one works well enough
find an easy way to estimate t0 offset dt
does this change with HV?