• Main INDEX
  • Monthly INDEX
  • PREV
  • NEXT

    User name D.J. Mack

    Log entry time 12:51:53 on June 30,1999

    Entry number 304

    keyword=systematic errors in ssxptar, ssyptar

    Here I summarize our present systematic reconstruction errors in xptar and yptar for the SOS. I only mention errors that are in the same phase space as our good Fpi data.

    Good Fpi data are in the range ssdelta between [-15,+15]. There are no ssxptar or ssyptar cuts at this time, but the sieve slits properly cover more than the collimated angle range so I have to give the nominal collimated ranges: ssxptar roughly +-35 mrad, and ssyptar roughly +- 55 mrad.

    ssxptar:

    1. (Figure 2 of Fpilog 299) ssytar vs ssxptar shows a large correlation which SHOULD BE REMOVED. A correction of the form ssxptar*=ssxptar + a*ssytar*ssxptar might work. This will improve the overall ssxptar resolution (unimportant) and will prevent this error from becoming correlated with ssyptar in our sieve mask (important!). The warping of the sieve slit plot due to this effect can be seen in Figure 2.


    ssyptar:

    First let me say that our miserable ssyptar resolution combined with our large ssyptar acceptance makes it difficult to see problems at the 1 mrad level. In the one case where I think there is a few mrad systematic error (at large -abs(ssyptar) and large -abs(ssdelta) for ssytar =0 in a plot Jochen provided ) the problem simply isn't there in Ketevi's plots of ssdelta vs ssysli (Figure 2 from Fpilog 76). Ketevi's plots combine data from ALL the ssytar positions, which is fine since at all our settings the SOS sees an extended target. The problem I saw in Jochen's plot presumably washes out over an extended target. Based on Ketevi's plot I'd say that correlations between ssyptar and ssdelta are less than 1 mrad, averaged over the target. Hence, DO NOTHING.