Slides 25-September-2002

• OK, some joy for a change. The reason the chi-squares were looking so rubbishy was that the old hit collections were being used instead of the new (better alignment ones) which had just been created. Thanks to Ken for the quick work-around while I rewrote modules.
• From the old stubfinder I find better looking chi-squares and the 3-parameter t0 fit shows signs of the other buckets for a change. First page top is of the fit t0's, requiring nhits=4. The bottom is of the chi-squared: left is nhits=3, right is nhits=4. The mean shouldn't be so high, of course.

  The second page shows the number of 3- and 4-hit stubs, and the distance of closest approach to the beamline of the stub projection; upper right for 3-hit stubs and lower left for 4-hit stubs. In the lower right you find the distribution of the reconstructed z for the stub. I don't know why it is sometimes 0. The spike at just under 600 is the center of the chamber--the default if there was no good reconstruction of Z for any hit. The tall spike is the rear of the chamber, where we get toasted.

  The third page shows the numbers of 3- and 4-hit stubs when the abs(z) is greater than 750 cm (the rear of the chamber) at the upper left, and the upper right shows the same when the z is forward of the hot rear. Notice that 3- and 4-hit stubs are found pretty equally in the rear, but at a ratio of 3 to 1 in the rest of the chamber. The bottom plots show the abs(z) when the number of hits is 3 (left) and 4 (right)--no surprises.

  The last page shows the chi-squared again when the stub in entirely in a single stack (self=0; upper left) and crossing two stacks (self=1; upper right). The lower left shows the self attribute when the number of hits is 3, and the lower right when the number of stubs is 4.

• The results from the new algorithm stubfinder are also better, though I still haven't addressed the odd chi-squared issue. The first page shows all chi-squares (upper left), those for nhits=3 (upper right), nhits=4 and all in same stack (lower right), and nhits=4 and crossing stacks (lower left).

  The second page shows the abs(z) distributions, and below that shows that the west to east ratio doesn't change when you require rear hits (left) or non-rear (right).

• I look at the time differences (OK, scaled by the signal propogation speed) for hits in jumpered cells. The main peak is at about 650 cm. I have a cut in the code on how far away from the peak the difference can be and still be considered a good pair of hits, which left the cut at about 700 cm. Eyeballing the plot suggests that 750 might be a better spot. The plots above were done with 730 cm as the cut.

Modified 25-Sep-2002 at 14:37

http://hep.physics.wisc.edu/~jnb/imu/25Sep2002

Previous meeting notes

Next meeting notes

Main slide directory

Please contact jnbt@hep.physics.wisc.edu if you have trouble accessing the information on this page.