[Halld-physics] Polarization values

[Michael Dugger]

Sasha,

I think I understand. You are talking about the use of the PS energies 
when performing the CBSA for the determination of polarization. I am not 
qualified to comment on the energy resolution required to obtain a 
reliable CBSA result for polarization. My comments regarding the PS and 
tagger energies are only to be taken within the context of TPOL 
measurements of polarization (without regard to CBSA).

Take care,
Michael


On Wed, 20 Sep 2017, Alexander Somov wrote:

>
> Hello Michael,
>
> TPOL measurements are clearly more effected by the
> PS reconstruction
>
> My point is that for the photon spectrum measurement
> with the PS, the shape difference between PARA and
> PERP even at the edge will be small and can be
> accounted for when you use the PS energy
> (for the spectrum measurement, one effectively
> integrates over phi, and the difference of events
> between PARA and PERP at the coherent edge in the
> vertical/horizontal planes is about 20%).
>
>
> Cheers,
>      Sasha
>
>
> On Tue, 19 Sep 2017, Michael Dugger wrote:
>
>> 
>> Sasha,
>> 
>> The results from TPOL when using PS energies is nearly the same as when 
>> using the tagger energy, except near the coherent edge. My hope was that I 
>> could neglect the tagger so as to get better statistics, but it turned out 
>> that near the coherent edge, the TPOL results using PS energies deviated 
>> too much from the case when tagger energies were used. This meant that near 
>> the coherent edge, there was no choice but to use the tagger energies. For 
>> the sake of consistency, we are now using the tagger energies for all TPOL 
>> energy binning.
>> 
>> Take care,
>> Michael
>> 
>> 
>> On Tue, 19 Sep 2017, Alexander Somov wrote:
>> 
>>> 
>>> Hello Richard,
>>> 
>>> A couple of things:
>>> 
>>> 1. Based on Mike's tool, for the full run set
>>> of 2017 (if I use the tool correctly, v2),
>>> error bars don't allow to conclude much about the
>>> edge of the coherent peak. Shapes look  identical
>>> for 0, 45, 90 and 135 deg. within error bars. The
>>> energy resolution of even 100 MeV will not impact
>>> the shapes much.
>>> 
>>> 2. The energy spread, which Mike presented seems
>>> to be relatively large (I have to check my Geant
>>> simulations). For the typical electron angle of
>>> about 0.1-0.2 mrad, the energy spread (sigma) in
>>> the horizontal plane should be better than
>>> 25-30 MeV  for the near converter (station 1).
>>> 
>>> For the far converter, the flight distance is
>>> larger, resulting in the (slightly) worse energy
>>> resolution due to the intrinsic PS angle. Originally
>>> we wanted to place the TPOL to the station 1. I think
>>> that we have to move it (at some point); the PS magnet
>>> fringe field should be Ok.
>>> 
>>> 3. A concern of using the tagger spectrum for
>>> high-statistic polarization measurements (without
>>> TPOL) is non-uniform and poorly known efficiencies
>>> (especially in TAGM), which we don't need for
>>> the tagged flux calculation but have to use
>>> to correct the shape.
>>> 
>>> Cheers,
>>>      Sasha
>>> 
>>> 
>>> 
>>> On Tue, 19 Sep 2017, Richard Jones wrote:
>>> 
>>>> Hello Mike,
>>>> 
>>>> It might be better to plot the same quantity on the x axis across the
>>>> entire plot, even if the statistical errors turn out to be somewhat 
>>>> larger.
>>>> Whatever the acceptance of the TAGH might be, it IS what defines the 
>>>> tagged
>>>> beam at energies above the TAGM. Unless someone has a physics channel for
>>>> which they do not need to know the energy of the beam photon, it is only
>>>> the tagged flux that is of any interest to GlueX analyses.
>>>> 
>>>> in response to Sascha,
>>>> 
>>>> The PS energy contains a polarization bias. If one is in a flat part of 
>>>> the
>>>> spectrum then this will average out, as you say. But if you are sitting 
>>>> in
>>>> a steep location, say just above the coherent edge then the PS has a
>>>> different energy resolution depending on the polarization of the photon. 
>>>> If
>>>> the pair comes out in the horizontal plane then it has an energy spread 
>>>> on
>>>> the order of 100 MeV, as Mike showed. The resolution for vertical-plane
>>>> pairs is much better. This would be expected to produce a different
>>>> steepness of the coherent edge in PARA and PERP runs. This is what I mean
>>>> by artifacts. But Mike is using the TAGM energy in the region of the
>>>> coherent edge, so that is good.
>>>> 
>>>> -Richard Jones
>>>> 
>>>> 
>>>> On Tue, Sep 19, 2017 at 12:04 PM, Michael Dugger <dugger at jlab.org> wrote:
>>>> 
>>>>> 
>>>>> Richard,
>>>>> 
>>>>> In the coherent peak I use the tagger energy. For energies outside the
>>>>> coherent peak, I use the PS. I use the PS outside the coherent peak 
>>>>> because
>>>>> of statistics.
>>>>> 
>>>>> For the Spring 16 priority set, I use the tagger energy for energies
>>>>> bewteen 8.4 and 9.4.
>>>>> 
>>>>> For the Spring 17 runs, I use the tagger energy for energies between 8.2
>>>>> and 9.3.
>>>>> 
>>>>> As long as people are within the energy ranges given above, the tagger
>>>>> energy is being used to bin the data.
>>>>> 
>>>>> If you don't like the mixing of PS and tagger for the energy binning, I
>>>>> can easily modify the code so that only tagger information is used. For 
>>>>> the
>>>>> most part, it will just pump up the error bars for the energy regions
>>>>> outside the coherent peak.
>>>>> 
>>>>> Take care,
>>>>> Michael
>>>>> 
>>>>> 
>>>>> On Tue, 19 Sep 2017, Richard Jones wrote:
>>>>> 
>>>>> Mike,
>>>>>> 
>>>>>> These are very nice results, good to see that the spread between the
>>>>>> different polarization orientations is now much less than we were 
>>>>>> seeing
>>>>>> before, esp. 2017.
>>>>>> 
>>>>>> One question about the energy scale. Are you using the tagger energy to
>>>>>> decide which bin each event should go into? Now that we know the PS 
>>>>>> energy
>>>>>> is entangled with the polarization, to produce a spectrum free of
>>>>>> artifacts
>>>>>> we should only be using the tagger energy to make this spectrum.
>>>>>> 
>>>>>> -Richard
>>>>>> 
>>>>>> On Tue, Sep 19, 2017 at 10:08 AM, Michael Dugger <dugger at jlab.org> 
>>>>>> wrote:
>>>>>> 
>>>>>> 
>>>>>>> Hi,
>>>>>>> 
>>>>>>> I have code that creates the polarizations for
>>>>>>> Spring 16 (priority set 1) and Spring 17 data
>>>>>>> at
>>>>>>> 
>>>>>>> https://userweb.jlab.org/~dugger/triPol/makePolValsV2.tar
>>>>>>> 
>>>>>>> Once you have un-tarred the file, go through the steps in
>>>>>>> the README file to generate the polarizations.
>>>>>>> 
>>>>>>> Take care,
>>>>>>> Michael
>>>>>>> 
>>>>>>> _______________________________________________
>>>>>>> Halld-physics mailing list
>>>>>>> Halld-physics at jlab.org
>>>>>>> https://mailman.jlab.org/mailman/listinfo/halld-physics
>>>>>>> 
>>>>>>> 
>>>>>> 
>>>> 
>>> 
>> 
>