[Clas_cascades] Minutes Thu Jul 29 posted

Thu Aug 5 00:52:12 EDT 2010

Hi Zhiwen,

Thanks for your comments and questions. See my answers below.

Hovanes.

Zhiwen Zhao wrote:
> Some of my comments on the July 29, 2010 version Phi--(1862) analysis note
> Many of them are certainly my problem of understanding.
>
> 1. p10. It might be good to mention that momentum cut of p and pip are 
> determined by comparison with simulation ( at least this is what I remember how 
> it's done?)
>   
The cut off  for momenta of protons of pions were set so that the 
kinematics does not get distorted
by particles who have very large uncertainty in their momentum 
determination. From data and simulations
we see that there are not that many protons with P<300 MeV and pion 
under 100 MeV. The cut values
were not selected from any special study, I just picked them such that 
there would not be "very large"
energy losses or under uncertainties for their momenta.
> 2. p13. the second last paragraph, should it be "as a function of phi". Also has 
> the momentum correction be applied? I notice it improves the resolution a little 
> bit in Lambda1520 study
>   
The p-dependent and phi dependent corrections in the paragraphs above. 
Since Paul wrote this I am not
absolutely sure, but I think it is correct.
> 3. p15. \tau is the life time in particle rest frame, the actual decay length in 
> lab frame is c\tau\gamma\beta or c\tau*p/m, I see the equation 5 missing the \beta ?
>   
p/m is not really equal (gamma beta), it is (gamma beta c). I think Eq. 
5 is correct.
> 4. p21. The difference between event start time and tagger hit time is 3.2ns. 
> How is it determined? For E,T-counter matching, I remember it's 1-187 Ecounter 
> to 1-19 Tcounter. Tagger cut is not present for both liberal and conservative 
> selection, right?
>   
I picked a shorter range for the tagger E-counters to avoid the lowest 
energy E-counters, which is safer.
In the energy this matches 4.55 to 5.5 GeV. The 3.2 ns was determined by 
studying the effects of this cut.
http://clasweb.jlab.org/rungroups/eg3/wiki/index.php/Study_of_the_tagger_timing_cuts
> 5. p26. In the right plot of fig 15, why those sideband in simulation? Why many 
> beta > 1 particles?
>   
I am not sure what the origin of beta>1 particles is. But I think there 
aren't so many of them.
> 6. p29. If we can determine the ST effect in every mass bin, why can we use the 
> the "actual acceptance with ST requirement" instead of using the "nominal 
> acceptance with a 15+-5% correction?
>   
I could do that, but did not. To do this I have to reprocess all the MC 
files to include the  ST hits
in the cooked BOS files and the ROOT files. This would take a long time 
and very likely would
result in ~15% chenge in acceptance. We would still need to quote some 
uncertainty ~5%. So,
it would be more proper to do acceptance corrections the way you are 
suggesting, and it is
doable, but would take longer time.
> 7. p37. what are the three independent parallel analysis?
>   
I know that Joern, Paul and myself for Phi(1860)-- and did not see it.
> 8. p43. I am not very clear about why the "unknown" background is as it is?
>   
The polynomial mass is just seemed to be  enough to describe the mass 
dependence of the background.
The ctau dependence would be purely Gaussian if there were no detached 
vertices involved in the
background events (events without Xi-). A possibility of detached 
vertices in the background events
would translate is some sort of exponential dependence in the distance 
between the vertices which
would look like and exponential dependence in ctau (of course convoluted 
with Gaussian because of
the vertex resolution). This seemed to work best.

> 9. p51. The final upper limit plots show no sign of Phi--, but with all the 
> fluctuation (particularly those peaking structures), can we conclude that we can 
> detect the Phi-- if it is actually there?
>   
If Phi-- was produced with large enough cross sections we would be able 
to see it. You should remember that the
number is the cross sections which is at 90% confidence is larger than 
the cross section at that mass point.  The variation
of that number is representative of the statistical variation of the 
number of events from one bin to another. We quote
an average value from 1.80 to 1.92 to have a single number, but our  
sensitivity estimate is mass dependent.

> 10. p60. For the last sentence of the statement, the 700pb upper limit safely 
> exclude the models without form factor in Fig 2, but can it really exclude the 
> the model with form factor and coupling g>0, which is almost at or below 1nb.
>   
I think we can exclude g=+1.27, but probably not g=0, and certainly not 
g=-1.27. This needs to be carefully
discussed in the actual paper. 
> Thanks
>
> Zhiwen
>
>
> On 7/29/2010 2:33 PM, Elton Smith wrote:
>   
>> Present: Elton, Stepan, Hovanes, Kijun
>> On the phone: Lewis, Zhiwen<br>
>> Notes by Elton.
>>
>> == Announcements==
>>
>> == Normalization ==
>> # No changes, latest version posted on wiki
>> # All tex and figures will be added to the cvs repository for save
>> keeping.
>>
>> == Comments on PHi-- Note ==
>> # Correction for decays outside of target estimated to be 15+/-5%.
>> # This correction has been included into the upper limits
>> # All documentation has been updated to reflect latest results
>> # Upper limit has increased to 700 pb, with the addition of the 15% loss
>> for decays with 5% uncertainty, as well as updated number number of
>> 10->15% uncertainty in normalization.
>> # Figure 20? in normalization note has been included into the response to
>> the committee. Lewis will send Hovanes the .eps file for best
>>
>> == Draft of Phi-- Paper ==
>> # Hovanes will start a draft of the paper.
>> # The draft will be provided to committee as soon as appropriate, but
>> response will not be delayed.
>> # We discussed how to include comparison of g11/eg3 Delta production.
>> Elton suggests to mention comparison in text but no figure so that CLAS
>> may allow publication to proceed without formal review of that data. In
>> any case the final results depend on it only indirectly as input to the
>> systematic error on the normalization.
>>
>> == Theses ==
>> # Zhiwen and Haiyun should provide final pdf versions of their theses to
>> Reinhard for posting on CLAS thesis page.
>>
>>
>> == Next Meeting Aug 5, 9:00 am ==
>> # Everyone read and comment on Phi-- analysis note and response to
>> committee. Any suggestions should be incorporated before next week.
>>
>>
>> Elton Smith
>> Jefferson Lab MS 12H5
>> 12000 Jefferson Ave
>> Suite # 16
>> Newport News, VA 23606
>> elton at jlab.org
>> (757) 269-7625
>> (757) 269-6331 fax
>> _______________________________________________
>> Clas_cascades mailing list
>> Clas_cascades at jlab.org
>> https://mailman.jlab.org/mailman/listinfo/clas_cascades
>>     
> _______________________________________________
> Clas_cascades mailing list
> Clas_cascades at jlab.org
> https://mailman.jlab.org/mailman/listinfo/clas_cascades
>   