[Halld-cal] BCAL sampling simulation

[David Lawrence]

Hi Andrei,

   Thanks for the clarifications. Here is the modified list of 
parameters based on the info you gave:

1. Total energy of incident photon that created the shower
2. Theta angle of the incident photon
3. Distance of incident photon at point of entry into BCAL to center of 
cell in azimuthal direction
4. Energy deposited in whole cell (2 bins 1 for E<50MeV and one for E>50MeV)
5. Cell row within module (1-8 with rows 7,8 being summed cells)

So are the summed outer cells representing 4 cells being summed or just 
the 2?

Let me know if the above list seems correct now.

For me personally, I think the single ROOT file would be more 
convenient. It has built-in compression, keeps multiple tables in a 
single file, and it makes it easy to plot the parameters in 2-D color 
plots like you showed at the meeting. (In fact, I'd assumed you had 
those in ROOT format already based on those plots.) Regardless of the 
storage format, the tables will be read into in-memory tables so the 
access speed will be the same. If ASCII files are a lot easier for you 
to produce, then just give me those along with a detailed specification.

Regards,
-David

On 1/28/13 8:10 PM, semenov at jlab.org wrote:
> David:
>
> Your list of input parameters is not quite correct:
>
> Parameter #5: At the moment, our GEANT model of BCAL module contains 8
> "effective readout" rows because of the real (2x2 cm^2) outer-cells rows
> #7 and #8 are combined together; the same is for the real rows #9 and #10.
> If parameterization in real individual cells is desired, the
> decision/request should be made as soon as possible. (Note that the
> threshold of 272 keV was applied on combined "effective" cells in the
> model.)
>
> Parameters #3 and #6: Instead of these 2 parameters, I used only one,
> which is the distance between the photon hit in the module and the center
> of the cell of interest.
>
> Parameter #4: Actually, that parameterization is pretty coarse; I just
> provide one set of coefficients for the events with the energy deposited
> in the whole cell < 50 MeV, and another set of coefficients for the rest
> of the events.
>
> By the way, why we need to keep the coefficients in the root histograms?
> What's wrong with the simple ASCII tables? (Less space in memory, works
> faster etc.)
>
> Thank you,
> Andrei
>
>
>
>> Hi Andrei,
>>
>>     I promised at the meeting on Tuesday to send you an e-mail indicating
>> the format of the the resolution tables that I would prefer. I'm CC-ing
>> the calorimetry mailing list in case anyone else would like to comment.
>>
>> First, let me verify what the input parameters are:
>>
>> 1. Total energy of incident photon that created the shower
>> 2. Theta angle of the incident photon
>> 3. Distance of incident photon at point of entry into BCAL from center
>> of module in azimuthal direction
>> 4. Energy deposited in cell (whole cell, not just fibers)
>> 5. Cell row within module (1-10)
>> 6. Cell column within module (1-4)
>>
>> I assume that depth of shower is already taken into account by index 5
>> and so is not needed as separate input parameter. Let me know if this is
>> not correct.
>>
>> Also, please remind me how the value for parameter 4. is used. Is this
>> just a cut on Ecell<50MeV or Ecell>50MeV to decide which parameter set
>> to use, or are there more than 2 bins in this parameter?
>>
>> What you will provide:
>>     Effectively, a six dimensional table that has 2 values, A.) the
>> sampling fraction and B.) the sampling fluctuation coefficient.
>>
>> I would suggest providing these in the form of a ROOT file containing a
>> set of 3-dimensional (TH3F) histograms.
>>    - indices x,y,z of those would correspond to input parameters 1,2,3
>> listed above
>>    - The names of the histograms would contain indices 4,5,6.
>>       e.g.
>>                 SampFrac_0_50MeV_row2_col4
>>                 FluctuationCoeff_0_50MeV_row2_col4
>>
>> If my understanding is correct, this will lead to:
>>    2*10*4 = 80 histograms for each of the 2 values or 160 histograms total
>>
>> Each histogram would be 8*12*32 = 3072 bins
>>
>> So for single precision, this would be about a 2MB file, not accounting
>> for any compression.
>>
>> Please let me know if anything is incorrect or you have another idea
>> regarding format.
>>
>> On a side note, I would like to confirm the 272keV threshold you used is
>> correct. To make it clear, this is the threshold hdgeant uses to decided
>> whether to write the hit for a single SiPM to the output file, or
>> discard it. The 272keV comes from requiring 1MeV deposition on the far
>> end which would attenuate to 272keV on the opposite end. Electronics
>> thresholds are much higher, but applied later after additional effects
>> (e.g. dark hits) are included.
>>
>> Regards,
>> -David
>>
>