[Halld-cal] updated study of Fcal response in MC

[Richard Jones]

Hello Ilya,

I agree, although I want to add the light is generated separately in the
lead glass blocks and the lucite light guides, in the MC simulation. These
two are combined into a single light output in mcsmear, in a treatment that
is the same for both hdgeant and hdgeant4.

-Richard J.

On Wed, Mar 13, 2019 at 12:58 PM Ilya Larin <ilarin at jlab.org> wrote:

>
> Hi Liping,
>
>
> I guess final linearity corrections might be involved in
>
> mcsmear and reconstruction to explain the observed
>
> positive deviation in reconstructed energy at high energy end.
> I don't think it is an effect of light guides or limited
> transparency of the blocks.
>
>  Regards,
> Ilya
>
> ------------------------------
> *От:* Gan, Liping <ganl at uncw.edu>
> *Отправлено:* 13 марта 2019 г. 12:50
> *Кому:* Richard Jones; Ilya Larin
> *Копия:* Liping Gan; Matthew Shepherd; halld-cal at jlab.org
> *Тема:* Re: updated study of Fcal response in MC
>
>
> Hi, Richard,
>
>
> In your result, the energy deviation for the reconstructed photon in FCAL
> reaches more than 100 MeV. I would like to ask Ilya to comment on this
> since he has done a lot of simulations on calorimeter.
>
>
> Liping
>
>
> ------------------------------
> *From:* Richard Jones <richard.t.jones at uconn.edu>
> *Sent:* Tuesday, March 12, 2019 2:16 PM
> *To:* Gan, Liping
> *Cc:* Liping Gan; Matthew Shepherd; halld-cal at jlab.org
> *Subject:* Re: updated study of Fcal response in MC
>
> Hello Liping,
>
> The response of the calorimeter itself in the MC does not have these two
> inflection points that you see in these plots. Those are coming from the
> interaction between the non-linear response in the simulation coupled with
> the (non-equivalent) non-linear corrections being applied in mcsmear and
> the cluster energy reconstruction. This plot shows the net effect of all 3
> (MC + mcsmear + recon).
>
> -Richard
>
> On Tue, Mar 12, 2019 at 1:13 PM Gan, Liping <ganl at uncw.edu> wrote:
>
> Hi, Richard,
>
>
> Thanks for the excellent result. I just wonder if there is any explanation
> for the small bump at 1.5-2.5 GeV in you plots? Thanks!
>
>
> Liping
> ------------------------------
> *From:* Richard Jones <richard.t.jones at uconn.edu>
> *Sent:* Tuesday, March 12, 2019 10:37:29 AM
> *To:* Liping Gan; Matthew Shepherd
> *Cc:* halld-cal at jlab.org
> *Subject:* updated study of Fcal response in MC
>
> Hello Liping and all,
>
> At your suggestion, I went back and redid my study of reconstructed
> showers in the FCAL from MC, comparing hdgeant and hdgeant4, this time
> covering a larger energy range up to 8 GeV. For the results, see the latest
> document uploaded to docdb.
>
> https://halldweb.jlab.org/doc-private/DocDB/ShowDocument?docid=3852
> <https://gcc01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fhalldweb.jlab.org%2Fdoc-private%2FDocDB%2FShowDocument%3Fdocid%3D3852&data=02%7C01%7Challd-cal%40jlab.org%7C2314165c877b442e598d08d6a7dab03b%7Cb4d7ee1f4fb34f0690372b5b522042ab%7C1%7C0%7C636880955031313611&sdata=03Fr%2FH%2Fr3RugCH51zVVHbqQInuICxC3%2Bx9rd3zRb7Eg%3D&reserved=0>
>
> Here is a summary of what I found.
>
>    1. I found a small bug in my fitting code that shifted the
>    Ereconstructed down by 16 MeV relative to the generated in my earlier
>    studies. This made the negative y-intercept in the Ereconstructed vs
>    Egenerated appear worse than it actually is.
>    2. I adjusted the output from shower particles in hdgeant4 relative to
>    hdgeant down by 0.8% as we discussed a the last calorimeter meeting. This
>    change is visible in the new results. I now claim that the agreement in the
>    shower energy response between hdgeant and hdgeant4 is very good.
>    3. With the new larger simulated energy range, my statistics of 100K
>    events total is limiting my statistical accuracy in studying the
>    non-linearity. Since there is no visible variation with polar angle seen in
>    these individual plots, I added at the end a plot with the full statistics
>    over the range 3-11 degrees, which has less statistical scatter and reveals
>    the nonlinear behavior better.
>    4. Above 4 GeV there is a very large rise in the reconstructed energy.
>    Overall the nonlinear correction needs work at both the low and high energy
>    ends of the Fcal spectrum.
>
> -Richard Jones
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mailman.jlab.org/pipermail/halld-cal/attachments/20190313/d99d8d96/attachment.html>