Difference between revisions of "CDC 50 50"
Line 39: | Line 39: | ||
Where to start the fit from. | Where to start the fit from. | ||
{| border="0" cellpadding="2" | {| border="0" cellpadding="2" | ||
− | |+ 50/50 gas, prototype horizontal, 4sigma (gray) and 5sigma (blue) hit threshold, fit range | + | |+ 50/50 gas, prototype horizontal, 4sigma (gray) and 5sigma (blue) hit threshold, fit range 6sigma+ |
|[[Image:run_31809_fitmulti.png|thumb|x250px|2090V, Landau fits, log likelihood method, fit range starting at 5, 6, 7, 8 x pedestal sigma]] | |[[Image:run_31809_fitmulti.png|thumb|x250px|2090V, Landau fits, log likelihood method, fit range starting at 5, 6, 7, 8 x pedestal sigma]] | ||
|[[Image:run_31809_fitmulti_z.png|thumb|x250px|2090V, zoomed in]] | |[[Image:run_31809_fitmulti_z.png|thumb|x250px|2090V, zoomed in]] | ||
Line 51: | Line 51: | ||
{| border="0" cellpadding="2" | {| border="0" cellpadding="2" | ||
− | |+ 50/50 gas, prototype horizontal, 4sigma (gray) and 5sigma (blue) hit threshold, fit range | + | |+ 50/50 gas, prototype horizontal, 4sigma (gray) and 5sigma (blue) hit threshold, fit range 5sigma+ |
|[[Image:run_31806_ampfitz.png|thumb|x250px|2050V, lines show 99.7%, 99% and 95% of landau summed 0-4000 (55, 70, 100)]] | |[[Image:run_31806_ampfitz.png|thumb|x250px|2050V, lines show 99.7%, 99% and 95% of landau summed 0-4000 (55, 70, 100)]] | ||
|[[Image:run_31809_ampfitz.png|thumb|x250px|2090V, lines show 99.7%, 99% and 95% of landau summed 0-4000 (55, 73, 111)]] | |[[Image:run_31809_ampfitz.png|thumb|x250px|2090V, lines show 99.7%, 99% and 95% of landau summed 0-4000 (55, 73, 111)]] | ||
Line 76: | Line 76: | ||
|- | |- | ||
| 2075 | | 2075 | ||
− | | | + | | 176 |
− | | | + | | 50 |
− | | 3. | + | | 3.5 |
− | | | + | | 70 |
| 23 | | 23 | ||
|- | |- | ||
| 2090 | | 2090 | ||
− | | | + | | 204 |
− | | | + | | 62 |
− | | 3. | + | | 3.3 |
− | | | + | | 73 |
− | | | + | | 20 |
|- | |- | ||
| 2100 | | 2100 | ||
− | | | + | | 225 |
− | | | + | | 69 |
| 3.3 | | 3.3 | ||
− | | | + | | 80 |
| 18 | | 18 | ||
|} | |} |
Revision as of 12:04, 11 October 2012
The later part of this page is 60/40 being used as a template for the 50/50. New stuff (50/50) is above the line.
50/50 Ar/CO2 and cosmics, prototype horizontal
Where to start the fit from.
The fits started at higher amplitude underestimate the width. Fit range starting at 5pedestalsigma has MPV 204.5, Sigma 61.9, MPV/Sigma 3.3. Fit range starting at 6pedestalsigma has MPV 208.6, Sigma 60.7, MPV/Sigma 3.4. Fit range starting at 7pedestalsigma has MPV 213.7, Sigma 59.5, MPV/Sigma 3.6. Use 5pedestalsigma start.
HV | Fit MPV | Fit sigma (FWHM/4) | MPV/sigma | 99% point | 4045/MPV |
---|---|---|---|---|---|
2075 | 176 | 50 | 3.5 | 70 | 23 |
2090 | 204 | 62 | 3.3 | 73 | 20 |
2100 | 225 | 69 | 3.3 | 80 | 18 |
This does not look nearly as good as 50/50 gas mix. Resolution is calculated as fitted drift distance - measured (interpolated) distance, so it is positive when the measured drift time and distance are too small. In this case they are slightly too small at small drift distance but too large at large drift distance. This could be a problem with the gas mix. For reference, plot res vs distance for tz=5,7,and 0.
Increasing tz_add to 8ns shifts the resolution up at earlier times but it hardly moves at later times. This would probably have been obvious from the tz=0 data (below) where all the measured drift distances are too large. "Measured" drift distances are actually interpolated from a table of distance calculated for a given gas mix and a range of time measurements. So for this data I should use Garfield tables for a gas mix with slower drift velocity, ie less Ar.
Suspect total flow of 60sccm might be too low for the CO2 MFC. Increase to 80sccm and repeat. Also try Garfield for different gas mixes and look at tmax.
Fitted Tmax was 573, edge is from 560 to 600ns. Garfield calcs give 560ns for 60/40, 573ns (ta-daa!) for 59/41, 588ns for 28/42 and 610ns for 57/43.
It's easiest to look at resolution far from the wire to find the best gas mix model.
Now look at timing. Where to take local pedestal. These are plots of change in ADC value vs time relative to the threshold crossing time. Nothing to do with Garfield or tz_add but already have a good default in place (take pedestal at np=13 minisamples before threshold crossing) so small changes here are a fine tuning.
Mean goes through 0.5 and 1 at -16.4 and -13.7 for thr=5sigma, -17.6 and -14.9 for thr=6sigma
Now include previously ignored hits which arrived before tzero+tz_add, assign these to a fixed drift time