Ho :-) A quick reply before I have to run: > - Looking at Fig.4 in the report of Claus and Christian (AnNote#22), > one can see clearly that around the nominal intersection point > there is no significant dependence of Mcut on the vertex (all > histograms and curves are absolutely flat). Therefore, I would > expect that both centrality recipes (JH and NBI) should give > exactly the same results if a narrow cut is imposed on the vertex. > It would be therefore very nice if you could compare the results > (centrality dependence of dN/deta) using both methods under > (for example) a +-5cm vertex cut (statistics should be sufficient, > in my old analysis I even used +-3cm). Not if you look at their figure 5. There you can see that the different selections actually pick out different percentages of the total MinBias. The 0-6% cut in GetCentrality actually looks like it picks close to 10% of the events, even for a narrow vertex cut. Also the 0-50% cut only selects a total of 40% of the events. The NBI cuts also have a few similar problems, but not as severe. I do have enough statsitics to do the check you propose and I can check it pretty quickly once I can sit down with it (there's a lot going on this week, so I'm a bit stressed), but I really don't think that the results are comparable. > - <Npart> does'nt depend linearly on centrality (I am not sure > that this is a few percent effect for all centrality cuts!). > I don't know how this would affect the results but it is clear > that it should be done more carefully. One can calculate <Npart> > directly from Glauber model (or any other geometrical model). I know <Npart> does'nt depend linearly on centrality, and the "real" dependance is there in the numbers we take from our ref. 1 (see below). The only "linear" assumption we make is in converting these numbers, which are integrated from 0-xx%, to our centrality classes which are yy%-xx%. This effect should not be large. I also agree that it is easy to calculate Npart from a glauber calc, but someone must sit down and do it. You have made a calc. using a hard sphere - is this easilly extendable to a Woods-Saxon and glauber model? > In your note, you say that <Npart> should come from simulating > the tiles. This is not really clear to me! All what you need > to calculate <Npart> is the percent of the cross section > for a given centrality cut and a geometrical model (Glauber > for example). Yes, if we assume that our centrality is correct. What I would like to see is a simulation of the Tiles subjected to our centrality cuts, then take the events accepted in each cent-bin and plot Npart (from HIJING or whatever model was used), and then take <Npart> to be the statistical mean of this distribution. This _should_ of course correspond to the simple geometrical <Npart>, but if it doesn't we clearly don't understand our detectors. > Is Reference 1 available somewhere on the WEB ? Yes, look at xxx.lanl.gov/abs/nucl-th/0012025 Ping :-) ------------------------------------------------ Bjorn H. Samset Master-student in Heavy Ion physics Mob: +47 92 05 19 98 Office: +47 22 85 77 62 Adr: Kri 2A709 Sognsveien 218 0864 Oslo
This archive was generated by hypermail 2b29 : Tue Feb 20 2001 - 05:30:35 EST