[Eg6_analysis] Checking the performance of the extracted drift speed corrections

Nathan Baltzell baltzell at jlab.org
Fri Jan 24 15:24:10 EST 2014 

Hi Mohammad,

I believe there is a misunderstanding.  My delta-theta(RTPC-RTPC) in
that document is the difference between the angles of two separate
tracks in the RTPC in the same event, and does not directly suggest
anything regarding an error or change in the angle.  What the right
side of Fig.1 shows is that the slope seen in left side of Fig.1
doubles when replacing the electron with another RTPC track and using
their angle difference.  This reinforces the hypothesis that this
angular dependence of dz(CLAS-RTPC) is due solely to an error on the
RTPC vertex.

So, you should ignore my delta-theta(RTPC-RTPC), it is just there for
validation of a hypothesis used to interpret the rest of the plots in
that document.  And it cannot be compared to the delta-theta you are
showing.



Rather, the bottom left of Fig. 1 is what I was referring to, and the
reason I brought it up is that your new calibration appears to slightly
increase the strength of our dz(CLAS-RTPC) vs theta dependence.  And
I'm not saying it is wrong, I just had my fingers crossed that your
new drift speeds would make the problem go away:)

We know there is tight correlation between the reconstruction of z
and theta, and the last plot on your wiki page confirms that again.
And that is what why an error on z is concerning: because it implies
a corresponding error on the more important theta those goes directly
into physics results.

Best Regards,
Nathan



On Fri, 24 Jan 2014 10:29:19 -0600, <hattawy at ipno.in2p3.fr> wrote:

> Hello Nathan, All,
>    I have compared the change in z versus the change in theta for both
> 1.206 and 6 GeV data and seen similar behavior. Here you can find them
> (https://clasweb.jlab.org/rungroups/lowq/wiki/index.php/Drift_Speed_Correction:_z/theta_Dependence).
> Also, i tried to make few points to distinguish between what you had in
> z/theta dependence and my delta_z vs. delta_theta. Please let me know
> if i am getting something wrong.
>     It Would be interesting to see your plots with the corrections we
> newly implemented. I will try to build your plots with what we have in
> the 10 test runs. Then, i will go in more details about the
> correlations between p and theta.
>
> Best regards,
> Mohammad.
>
>
>
>> Mohammad,
>>
>> I checked the 1.269 GeV data, and I see a ~40% increase in
>> elastics, similar enough to your 60%.  Where I get zero
>> difference in elastic yield compared to pass1v1 is in 1.206.
>>
>>
>> On your 2nd question, we saw a concerning theta-dependence
>> of dz(RTPC-CLAS) previously (Fig. 1, bottom left, for example):
>>
>> https://userweb.jlab.org/~baltzell/EG6/RTPC/tpcerror.pdf
>>
>> It looks like your Fig. 5 is going to increase it some,
>> but I don't know it means.
>>
>> One observation from your elastic plots is that the change in
>> p and theta due to the new drift speed is highly correlated.
>> We expect z and theta must be correlated, and I showed
>> previously that p and phi are too.  Just an observation ...
>>
>> Regards,
>> Nathan
>>
>>
>> On Thu, 23 Jan 2014 06:36:46 -0600, <hattawy at ipno.in2p3.fr> wrote:
>>
>>> Hell Nathan,
>>>   I have updated the wiki page
>>> (https://clasweb.jlab.org/rungroups/lowq/wiki/index.php/Drift_Speed_Correction:_Results)
>>> with the detailed comparisons for the three data packages (1.2, 6 and
>>> 1.27 GeV). In each section you can find a link to more detailed graphs.
>>>    I am wondering about the 1.27 GeV data, i got 60% more elastic
>>> events.
>>> Do you apply the same selection cuts i used?.
>>>    Also, in 6 and 1.27 GeV data we got a dependence between the change
>>> in
>>> the constructed vertex in RTPC and the polar angle. What do you think
>>> about this.
>>>
>>> Best regards,
>>> Mohammad.
>>>
>>>
>>>> Hi Mohammad,
>>>>
>>>> Looks like the effects on momentum and theta are not too large
>>>> for the elastics you show (~1% and 0.1deg).  But it would be most
>>>> interesting to see how much change in reconstructed quantities
>>>> there is for the 6 GeV runs where the difference in the drift
>>>> speed compared to pass1v1 is much larger.  Maybe we need some
>>>> more 6 GeV files for that.
>>>>
>>>> For the 1.206 GeV data, I actually see almost no change in
>>>> the number of good elastic tracks in the RTPC, but I get 30%
>>>> more good tracks in the dz(CLAS-RTPC) peak for 6 GeV.
>>>>
>>>> Regards,
>>>> Nathan
>>>>
>>>>
>>>> On Wed, 22 Jan 2014 09:42:12 -0600, <hattawy at ipno.in2p3.fr> wrote:
>>>>
>>>>> Hello All,
>>>>>   I updated the wiki page (
>>>>> https://clasweb.jlab.org/rungroups/lowq/wiki/index.php/Drift_Speed_Correction:_Results)
>>>>> with the new plots related to checking the performance of the drift
>>>>> speed corrections.
>>>>>   For the 1.2 GeV data at the beginning, The two cooking versions are
>>>>> consistent at the level of selecting good tracks. While, the newly
>>>>> cooked data have collected 15 % more elastic events. For the 6 GeV
>>>>> data,
>>>>> the same increase percentage has been achieved in collecting good
>>>>> tracks.
>>>>>
>>>>>   Suggestions are greatly appreciated.
>>>>>
>>>>> Best regards,
>>>>> Mohammad.
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>> Hi Mohammad, Hi All,
>>>>>>
>>>>>> After the modification that Mohammad did to initialize the run
>>>>>> number's
>>>>>> variable properly before the new implemented drift path calculation
>>>>>> is
>>>>>> done, I have processed
>>>>>> some runs of 1.2 GeV and 6 GeV data-set. The new 1.2 GeV cooked data
>>>>>> are
>>>>>> located in this directory:
>>>>>> /volatile/clas/claseg6/gempass1_v3/1p2gev/.
>>>>>> However, the
>>>>>> processed 6 GeV data-files, which their cooking is not done yet,  
>>>>>> will
>>>>>> go
>>>>>> eventually to this directory:
>>>>>> /volatile/clas/claseg6/gempass1_v3/6gev/.
>>>>>>
>>>>>> Best regards,
>>>>>>
>>>>>> Lamiaa
>>>>>>
>>>>>>
>>>>>>
>>>>>> On Fri, Jan 17, 2014 at 10:36 AM, <hattawy at ipno.in2p3.fr> wrote:
>>>>>>
>>>>>>> Hello Nathan and All,
>>>>>>>    There was a problem with the previous test cooking. The TDCmax
>>>>>>> calculation needs the run number as input. I implemented it to be
>>>>>>> the
>>>>>>> runNum variable, but it seems not working. Attached two plots, the
>>>>>>> expected TDCmax vs. z from the corrections and the actual one we  
>>>>>>> got
>>>>>>> from the previous test cooking.
>>>>>>>     My corrections include the runs from 61488 to 61960. I kept the
>>>>>>> previous formula for the runs which are not included in this range.
>>>>>>> This formula is: TDCmax = 53.1422 -0.0317579*z +0.00199476*z*z.  
>>>>>>> This
>>>>>>> formula is giving us the concave up distribution of TDCmax which we
>>>>>>> got from cooking, while the corrections have to give concave down
>>>>>>> TDC
>>>>>>> vs. z distribution.
>>>>>>>     I have to investigate the variable runNum and make sure that it
>>>>>>> is
>>>>>>> referring to the run Number. Then i will cook some files to check  
>>>>>>> if
>>>>>>> everything is going well and update the wikipage.
>>>>>>>
>>>>>>> Best regards,
>>>>>>> Mohammad.
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>> > Hi Mohammad, Nathan and All,
>>>>>>> >
>>>>>>> > I have implemented several modified routines from Stepan in our
>>>>>>> software
>>>>>>> > to
>>>>>>> > solve the segmentation fault issue that was occasionally  
>>>>>>> happening
>>>>>>> > while reading the RF TDC time for some data-files, and also to  
>>>>>>> fix
>>>>>>> the
>>>>>>> > false return of a .not. operator whenever applied to a true
>>>>>>> logical
>>>>>>> > expression!
>>>>>>> > I believe these changes won't affect Mohammad's drift path study
>>>>>>> results,
>>>>>>> > but it might affect the particle identification and a time
>>>>>>> determination
>>>>>>> > in CLAS based on Nathan's comparison (this
>>>>>>> > email<
>>>>>>> https://mailman.jlab.org/pipermail/eg6_analysis/2013-December/000357.html
>>>>>>> >),
>>>>>>> > in which he reported some changes in a trigger time and a neutral
>>>>>>> > particles
>>>>>>> > id. between
>>>>>>> > Pass1_v1 and the new OS cooked files. For a quick check, I
>>>>>>> processed
>>>>>>> 10
>>>>>>> > files of a run 61448, which you can find in this directory:
>>>>>>> > /volatile/clas/claseg6/tgemRHL6_v3/1p2gev/. But If you need more
>>>>>>> data
>>>>>>> for
>>>>>>> > this comparison I can process more; either reprocess the same
>>>>>>> files
>>>>>>> > that Mohammad used in his study for both beam energies or others
>>>>>>> based
>>>>>>> on
>>>>>>> > your preferences.
>>>>>>> >
>>>>>>> > Thank you,
>>>>>>> >
>>>>>>> > Lamiaa
>>>>>>> >
>>>>>>> >
>>>>>>> >
>>>>>>> > On Thu, Jan 16, 2014 at 4:18 PM, Nathan Baltzell
>>>>>>> <baltzell at jlab.org>
>>>>>>> > wrote:
>>>>>>> >
>>>>>>> >> Hi Mohammad & EG6,
>>>>>>> >>
>>>>>>> >> I am noticing some strange things with this new data.  While the
>>>>>>> >> yields in dz(CLAS-RTPC) peak are increased like your good track
>>>>>>> >> yields, S/B ratio is smaller than it used to be (but maybe that
>>>>>>> >> is ok).  Next, the width of the peak is larger.  So I looked at
>>>>>>> >> dz vs theta, and that same dependence of the peak position in
>>>>>>> >> pass1v1 is now at least 3x larger.   Also, if you look at it as
>>>>>>> >> function of z, things are quite different:  as you move  
>>>>>>> upstream,
>>>>>>> >> dz peak gets washed out, whereas before it was pretty clear at
>>>>>>> >> all vertices.  Do you see the same?  I don't think any of these
>>>>>>> >> effects are coming from the electron vertex, as a quick  
>>>>>>> event-by-
>>>>>>> >> event scan shows the electron vertex is basically unchanged, and
>>>>>>> >> the beam windows haven't moved.
>>>>>>> >>
>>>>>>> >> Any ideas?  I put a few of these plots on the wiki:
>>>>>>> >>
>>>>>>> https://clasweb.jlab.org/rungroups/lowq/wiki/index.php/Gempass1v2nab
>>>>>>> >>
>>>>>>> >> It might be useful to see some more plots of the change in
>>>>>>> >> reconstructed quantities for the overlap events.  Like
>>>>>>> >> dz/dtheta/dp versus theta/z (you already showed dp vs z).
>>>>>>> >>
>>>>>>> >> And maybe we should go ahead and recook the rest of 1 GeV to
>>>>>>> >> look closer at elastics once 64bit user_ana issues are resolved.
>>>>>>> >>
>>>>>>> >> Regards,
>>>>>>> >> Nathan
>>>>>>> >>
>>>>>>> >>
>>>>>>> >>
>>>>>>> >> On Thu, 16 Jan 2014 11:02:16 -0600, <hattawy at ipno.in2p3.fr>
>>>>>>> wrote:
>>>>>>> >>
>>>>>>> >> > Hello Nathan,
>>>>>>> >> >    For the new region which is around 2 rad in phi, it is not
>>>>>>> totally
>>>>>>> >> new
>>>>>>> >> > region. If you see in the previous cooking it exists but not
>>>>>>> very
>>>>>>> >> > clear.
>>>>>>> >> >
>>>>>>> >> >    About the 60% increase in the 1.2 GeV, yes the TDCmax and
>>>>>>> the
>>>>>>> drift
>>>>>>> >> > speed they are consistent in the two method with a variation  
>>>>>>> of
>>>>>>> 1%.
>>>>>>> It
>>>>>>> >> > is not clear for me why we have this increase. It might be the
>>>>>>> fact
>>>>>>> >> > that we compared the drift speed of the golden elastic events
>>>>>>> to
>>>>>>> the
>>>>>>> >> > one of good tracks. If you can have a look at the two versions
>>>>>>> of
>>>>>>> >> > cooking for 1.2 runs, you will see the distributions of sdist,
>>>>>>> edist
>>>>>>> >> > and the other parameters getting narrower in the new cooking,
>>>>>>> while
>>>>>>> >> > they have been wide before and large number of the tracks they
>>>>>>> were
>>>>>>> >> not
>>>>>>> >> > included in the initial drift speed comparison.
>>>>>>> >> >    The new cooked files are in:
>>>>>>> >> > /volatile/clas/claseg6/gempass1_v2/6gev/
>>>>>>> >> > /volatile/clas/claseg6/gempass1_v2/1p2gev/
>>>>>>> >> >
>>>>>>> >> >    The old cooking are in:
>>>>>>> >> > /volatile/clas/claseg6/gempass1_v1/6gev/
>>>>>>> >> > /volatile/clas/claseg6/gempass1_v1/1p2gev/
>>>>>>> >> >
>>>>>>> >> > Best regards,
>>>>>>> >> > Mohammad.
>>>>>>> >> >
>>>>>>> >> >
>>>>>>> >> >
>>>>>>> >> >> Hi Mohammad,
>>>>>>> >> >>
>>>>>>> >> >> Looks like you also picked up a region of phi that didn't  
>>>>>>> even
>>>>>>> exist
>>>>>>> >> >> before (~2.2 rad)!
>>>>>>> >> >>
>>>>>>> >> >> Is it correct that there is 60% increase even in the 1.2 GeV
>>>>>>> runs
>>>>>>> >> that
>>>>>>> >> >> were being used
>>>>>>> >> >> for the previous calibrations?  I thought old and new drift
>>>>>>> speed
>>>>>>> >> were
>>>>>>> >> >> consistent there?
>>>>>>> >> >>
>>>>>>> >> >> And can you tell me where these newly cooked 10 runs are?
>>>>>>> >> >>
>>>>>>> >> >> Regards,
>>>>>>> >> >> Nathan
>>>>>>> >> >>
>>>>>>> >> >>
>>>>>>> >> >>
>>>>>>> >> >> On Thu, 16 Jan 2014 09:57:42 -0600, Mohammad Hattawy
>>>>>>> >> >> <mohammad.hattawy at gmail.com> wrote:
>>>>>>> >> >>
>>>>>>> >> >>> Dear all, As we have seen from the previous studies, the
>>>>>>> drift
>>>>>>> speed
>>>>>>> >> is
>>>>>>> >> >>> changing during the experiment. I have been working on
>>>>>>> correcting
>>>>>>> >> the
>>>>>>> >> >>> drift speed. New correcting >parameters are extracted and
>>>>>>> >> implemented
>>>>>>> >> >>> in
>>>>>>> >> >>> the cooking package. Selected 10 test runs were recooked  
>>>>>>> with
>>>>>>> these
>>>>>>> >> new
>>>>>>> >> >>> modifications.
>>>>>>> >> >>> I did a study of the performance of these corrections. As a
>>>>>>> result,
>>>>>>> >> we
>>>>>>> >> >>> collect more good tracks in the RTPC. This incease is  
>>>>>>> ranging
>>>>>>> from
>>>>>>> >> 60%
>>>>>>> >> >>> to 126%.  A detailed >study can been found here:
>>>>>>> >> >>>
>>>>>>> >>
>>>>>>> https://clasweb.jlab.org/rungroups/lowq/wiki/index.php/Drift_Speed_Correction:_Results
>>>>>>> >> >>>
>>>>>>> >> >>>  Comments and suggestions are greatly appreciated.
>>>>>>> >> >>>
>>>>>>> >> >>> --Best regards,
>>>>>>> >> >>> Mohammad
>>>>>>> Hattawy._______________________________________________
>>>>>>> >> >> Eg6_analysis mailing list
>>>>>>> >> >> Eg6_analysis at jlab.org
>>>>>>> >> >> https://mailman.jlab.org/mailman/listinfo/eg6_analysis
>>>>>>> >> >>
>>>>>>> >> _______________________________________________
>>>>>>> >> Eg6_analysis mailing list
>>>>>>> >> Eg6_analysis at jlab.org
>>>>>>> >> https://mailman.jlab.org/mailman/listinfo/eg6_analysis
>>>>>>> >>
>>>>>>> > _______________________________________________
>>>>>>> > Eg6_analysis mailing list
>>>>>>> > Eg6_analysis at jlab.org
>>>>>>> > https://mailman.jlab.org/mailman/listinfo/eg6_analysis
>>>>>>> >
>>>>>>>
>>>>>>> _______________________________________________
>>>>>>> Eg6_analysis mailing list
>>>>>>> Eg6_analysis at jlab.org
>>>>>>> https://mailman.jlab.org/mailman/listinfo/eg6_analysis
>>>>>>>
>>>>>>>
>>>>>>
>>>> _______________________________________________
>>>> Eg6_analysis mailing list
>>>> Eg6_analysis at jlab.org
>>>> https://mailman.jlab.org/mailman/listinfo/eg6_analysis
>>>>
>>>
>>>
>>> _______________________________________________
>>> Eg6_analysis mailing list
>>> Eg6_analysis at jlab.org
>>> https://mailman.jlab.org/mailman/listinfo/eg6_analysis
>> _______________________________________________
>> Eg6_analysis mailing list
>> Eg6_analysis at jlab.org
>> https://mailman.jlab.org/mailman/listinfo/eg6_analysis
>>
>
>
> _______________________________________________
> Eg6_analysis mailing list
> Eg6_analysis at jlab.org
> https://mailman.jlab.org/mailman/listinfo/eg6_analysis

More information about the Eg6_analysis mailing list