[Hps-ecal] pedestals and pileup

[Andrea Celentano]

I agree,
a running average is a even better choice, N has to be tuned according 
to the two effects:

- Single event noise on the first 4 samples (the larger N, the smaller 
the pedestal indetermination)
- Long-term fluctuations (the smaller N, the better determination of 
these fluctuations).

Andrea
On 01/11/2015 03:22 PM, Nathan Baltzell wrote:
> Hi Andrea,
>
> I think you are right:  we should use the pedestal from mode-7 data.  And there is a 3rd
> option I mentioned on page 3 that I think may be better than event-by-event or run-averaged:
> keep a running average of the pedestal over the last N seconds/events.
>
> -Nathan
>
>
> On Jan 11, 2015, at 6:48 AM, Andrea Celentano <andrea.celentano at ge.infn.it> wrote:
>
>> Hi Nathan,
>> this is a very interesting study. I'd like to add a comment to the discussion (something we already mentioned some time ago at one of the Ecal meetings)
>>
>> - In FADC mode 7, for every event and for every channell over threshold, we record 4 numbers:
>> * Energy (in FADC counts, i.e the sum of the samples around the signal peak)
>> * Time
>> * Pedestal (average of the first 4 samples, in FADC counts)
>> * Signal amplitude (in FADC counts, i.e. the max signal height)
>>
>> We can think about 2 ways of subtracting the pedestal using these data
>>
>> 1) Event by event, as energy - pedestal * fadc_integration_window_width
>> 2) On average, computing the average pedestal for a given channel over the run, and then subtracting it as before
>>
>> If we expect the pedestal to be roughly constant during a run (if the beam current is constant), then I expect the second solution to be better than the first one.
>> With the first solution, event by event the pedestal noise affects the energy resolution.
>>
>>
>> Andrea
>>
>> On 01/10/2015 10:48 PM, Nathan Baltzell wrote:
>>> Hi Everyone,
>>>
>>> Here is a recent look at some pedestal studies.  It appears they have an effect from pileup.
>>> It’s not significant for many channels, but for high rate channels near the beam I think we
>>> are going to want to correct it.  It would be interesting to see if Luca’s gain calibration sees
>>> any current-dependence coming from incorrect pedestals.
>>>
>>> -Nathan
>>>
>>>
>>>
>>>
>>>
>>> _______________________________________________
>>> Hps-ecal mailing list
>>>
>>> Hps-ecal at jlab.org
>>> https://mailman.jlab.org/mailman/listinfo/hps-ecal