Re: FW: MA Calibrations update

From: Ramiro Debbe (debbe@sgs1.hirg.bnl.gov)
Date: Fri Jun 28 2002 - 10:39:12 EDT

  • Next message: Peter H. L. Christiansen: "Spectra update" 

    Let me contribute with one hard number about multiplicity resolution;
when we apply the hardware cut on the tiles to make trigger 3 that I'm 
told selects 25% centrality we can see that the shape of the 
distribution is not a sharp edge but rather an almost gaussian shape 
(half of it)
The plot I'm using is:
    When I extract the sigma of that edge I get what I would call the resolution of the hardware trigger of ~7% Maybe one can argue that the final measurement of multiplicity is done with better sampling and it should have better resolution. At least we have a lower limit. Ramiro On Friday, June 28, 2002, at 10:41 AM, Hironori Ito wrote: > Hello. Looking at the multiplicity distribution and particularly > its shoulder, I would say the top 4% is ok. If we really want, the top > 3% > might be ok, but no more. Uncertainty of the top 4% (might also be > 3%) centrality should be the same as that of 5%. > > For anything below 50%, multiplicity is not a good way to > select an event centrality. This should be obvious due to large yields > of low multiplicity events. ZDC is the best candidate for these > centrality selections. However, I am not quite sure how much we > understand the response of ZDC in AA collisions. Without a good event > generator that reasonably describes the response of ZDC, it is hard to > gauge the accuracy of the centrality by ZDC. Of course, if we are as > adventurous as PHENIX, we could ignore these facts and use ZDC. ( I do > not think PHENIX understands ZDC any better than STAR, PHOBOS and we > do.) > > > Hiro > > > On Fri, 28 Jun 2002, Ian Bearden wrote: > >> Hi Steve, >> Thanks for the quick (and clear!) answer. >> I have one more thing, and I am not sure if it is a question or a >> statement. >> Read and decide! >> Can we really look at the 1 or 2% most central? I would guess that >> once we >> are in this regime, we are looking at fluctuations (of detector >> response) >> and not really centrality. What then, is the 'most' central we can >> look at? >> I would say that we can safely (and correctly) look at the most 5%. >> And, I >> would not venture to discuss single events using the spectrometers, as >> ensembles are my friends. >> Anybody else say something different? >> And now, since I am not so conservative, what should X be? 0.5%? >> >> >> >>> -----Original Message----- >>> From: owner-brahms-dev-l@bnl.gov [mailto:owner-brahms-dev-l@bnl.gov]On >>> Behalf Of Stephen J. Sanders >>> Sent: 28. juni 2002 15:08 >>> To: brahms-dev-l@bnl.gov >>> Subject: Re: FW: MA Calibrations update >>> >>> >>> Hi Ian, >>> The binning question should only be an issue if you are looking at the >>> most 1 or 2% most central events-- >>> even by 3% we have sufficient statistics in the min-biased spectra to >>> reasonably describe the >>> functional behavior. >>> >>> Your question invites a misleading answer: For near central events >>> (say 20% central or better) >>> I think your 1% is probably high (that is, a good conservative >>> estimate) >>> for what we might see in variations in the centrality cut from one >>> run to the next. Our main overall uncertainty is in the total >>> min-biased yields which, remember, >>> requires a correction for "missed" low-multiplicity events. >>> >>> HOWEVER, for any given event, a centrality of, say, 5% will have an >>> uncertainty of >>> several percent. You might want to have a look at Fig. 5 in my Park >>> City >>> contribution >>> http://www.sdcc.bnl.gov/brahms/pubs/parkcity.pdf. >>> A picture showing what different centrality "cuts" correspond to in >>> inpact parameter space is found in the analysis note that Hiro and I >>> wrote on the dNdEta analysis: >>> >>> http://www.sdcc.bnl.gov/brahms/private/AnNotes/dNdeta_analysis.pdf >>> >>> On an event-by-event basis it is also possible to have a very long >>> "tail" for the distribution of "measured"-"real" centrality >>> (based on geant, of course...). This is clear by inspection of the >>> figures cited above. >>> >>> Regards, >>> Steve >>> >>> Ian Bearden wrote: >>> >>>> Hi Steve, Hiro, >>>> Have you guys got a good idea of the error on the centrality? >>>> That is, if I select the 5% most central, I really select >>> (5+/-X)%, and I'd >>>> like to know, roughly, what is X? I suppose it is around one, >>> since Steve >>>> says we cannot bin more finely than 2%. Is this supposition correct? >>>> Ian >>>> >>>>> -----Original Message----- >>>>> From: owner-brahms-dev-l@bnl.gov [mailto:owner-brahms-dev- >>>>> l@bnl.gov]On >>>>> Behalf Of Stephen J. Sanders >>>>> Sent: 27. juni 2002 19:12 >>>>> To: brahms-dev-l@bnl.gov >>>>> Subject: Re: FW: MA Calibrations update >>>>> >>>>> >>>>> Hi Claus, >>>>> >>>>> I haven't gotten up to the 5900's yet (I should get there sometime >>>>> tomorrow), >>>>> but I suspect this looks much worst than it is. In order to >>> calibrate the >>>>> centrality I need to start with a min-biased multiplicity >>> spectrum. These >>>>> essentially don't exist for the later runs with any reasonable >>>>> statistics. A very small >>>>> change at the high multiplicity end of the spectrum will be STRONGLY >>>>> magnified when looking at the centrality histogram. For these >>>>> runs I >>>>> would suggest bin sizes of no less than 2% in centrality to avoid >>>>> nightmares of bad calibrations... >>>>> >>>>> Having said this, I am noticing some differences in the centrality >>>>> behavior when I compare the 0-20% range a centrality spectrum based >>>>> on >>>>> min-biased events, and one using trigger 6. >>>>> Hiro and Flemming have also noted this and the behavior is currently >>>>> being explored. The effect of concern is a dip in the centrality >>>>> spectrum for the most central events when >>>>> using trigger 6. Until this latest replay using the reduced event >>>>> files >>>>> I have only been looking at trigger 4 events since these are the >>>>> ones >>>>> needed for the calibrations. >>>>> >>>>> Regards, >>>>> Steve >>>>> Claus O. E. Jorgensen wrote: >>>>> >>>>>> Hi Steve >>>>>> >>>>>> I've looked at the centrality for the high field runs (5901-5983) >>>>>> and >>>>>> I found this: >>>>>> >>>>>> http://www.nbi.dk/~ekman/centTest.gif >>>>>> http://www.nbi.dk/~ekman/centTestZoom.gif >>>>>> >>>>>> which worries me a bit. I should say that to produce these plots >>>>>> I've made a vertex cut (+- 15cm) and requiring good zdc-bb >>>>>> correlation. >>>>>> It looks like the calibrations are not optimal. >>>>>> >>>>>> However, I don't think that there's much to do about it now. >>> The reduced >>>>>> files are produced and I don't think we have time to do more >>>>>> >>>>> calibrations >>>>> >>>>>> and reductions before QM. And I guess that 5% central is more >>>>>> >>>>> or less the >>>>> >>>>>> most central events. >>>>>> >>>>>> Cheers, >>>>>> >>>>>> Claus >>>>>> >>>>>> +-------------------------------------------------------------+ >>>>>> | Claus Jørgensen | >>>>>> | Cand. Scient. Phone : (+45) 33 32 49 49 | >>>>>> | Cell : (+45) 27 28 49 49 | >>>>>> | Niels Bohr Institute, Ta-2, Office : (+45) 35 32 53 07 | >>>>>> | Blegdamsvej 17, DK-2100, E-mail : ekman@nbi.dk | >>>>>> | University of Copenhagen Home : www.nbi.dk/~ekman/ | >>>>>> +-------------------------------------------------------------+ >>>>>> >>>>>> On Thu, 27 Jun 2002, Stephen J. Sanders wrote: >>>>>> >>>>>>> This may come across twice...or not at all. I sent it >>> yesterday but it >>>>>>> never showed up on the server.... >>>>>>> >>>>>>> ---------- >>>>>>> From: "Stephen J. Sanders" <ssanders@ku.edu> >>>>>>> Date: Wed, 26 Jun 2002 13:41:45 -0500 >>>>>>> To: <brahms-l@bnl.gov> >>>>>>> Subject: MA Calibrations update >>>>>>> >>>>>>> Hi, >>>>>>> Now that we have the beautiful reduced event files (thanks >>>>>>> >>>>> Claus, Ian, and >>>>> >>>>>>> whoever else helped to produce these files!!!), I've started >>>>>>> >>>>> to replay the >>>>> >>>>>>> runs to: a) get all of the MA pedestals correct and b) check >>>>>>> >>>>> that we don't >>>>> >>>>>>> have any serious problems with the array calibrations. >>>>>>> >>>>>>> To check the MA calibration, I'm calculating dN/dEta (SiMA) >>>>>>> >>>>> for the 0-5% and >>>>> >>>>>>> 30-40% centrality cuts, with the centrality calculated using >>>>>>> >>>>> the combined >>>>> >>>>>>> TMA and SiMA data as done for our multiplicity papers. The >>>>>>> >>>>> calculations are >>>>> >>>>>>> identical to what was done for the 200 GeV paper. >>>>>>> >>>>>>> Although I still have a number of runs to complete, I think a >>>>>>> fairly >>>>>>> reasonable picture is started to emerge and I wanted to relate >>>>>>> >>>>> this to the >>>>> >>>>>>> Collaboration as people get started on "final" passes. >>>>>>> >>>>>>> A figure of the dN/dEta results vs. run number is at >>>>>>> >>>>>>> http://kuphsx2.phsx.ukans.edu/~sanders/MACalib/dNdEta.jpg >>>>>>> >>>>>>> I've fixed the error bars at +/-4%. >>>>>>> >>>>>>> I have some other figures that might be of interest at >>>>>>> >>>>>>> http://kuphsx2.phsx.ukans.edu/~sanders/MACalib >>>>>>> >>>>>>> In particular, there is pretty clear evidence of a slope to >>>>>>> >>>>> the BB Vertex - >>>>> >>>>>>> TPM1 Tracking Vertex vs. Multiplicity plots for most runs. >>>>>>> >>>>> The same is true >>>>> >>>>>>> with the ZDC vertex. I've been doing these calculations since >>>>>>> >>>>> Monday and so >>>>> >>>>>>> any very recent changes in the DB may not be included in the >>>>>>> present >>>>>>> results. >>>>>>> >>>>>>> I hope to have new pedestals for all of the runs where we have >>>>>>> >>>>> reduced event >>>>> >>>>>>> files by this weekend. Unfortunately, these will be added to >>>>>>> >>>>> brat as ascii >>>>> >>>>>>> calibration files... >>>>>>> >>>>>>> Regards, >>>>>>> Steve >>>>>>> >>>>>>> >>>>>>> >>>>> >>>>> >>>>> >>>> >>> >>> >>> >>> >>> >> >> >> >

    This archive was generated by hypermail 2b30 : Fri Jun 28 2002 - 10:39:53 EDT