Hi Eugene,
I think you have shown that with ELOSS=off the average difference of pml,r
and pwl,r is consistent with zero. Moreover, pml,r (ELOSS=on) and pml,r
(ELOSS=off) should be identical since the eloss happens only along
tracking. Therefore, I thought it should be fine to just consider
pml,r-pwl,r for ELOSS=on to parameterize the momentum losses (as a
function of pwl,r).
You are right that it is a good idea to check if the same momentum loss
parametrization that you've already given also holds for the large bgrid.
On the shift of the elastic peak in MASCARAD-W: According to Chi, Mascarad
only calculates the radiation tail, but not the combination of unradiated
and radiated events. The radiation tail only starts off for radiated
photons of >10 MeV, therefore Mascard-tail is always shifted by at least
10 MeV. Convoluting the tail with finite resolution makes the observed
shifts even bigger. There seems to be a remaining issue on how to properly
combine the unradiated Epel with the radtail from Mascarad.
Regards,
Michael
On Sun, 26 Mar 2006, Eugene J. Geis wrote:
> Quoting Michael Kohl <kohlm@mit.edu>:
>
>> Eugene, Adrian:
>> -keep MASCARAD off
>> -compare pml,r-pwl,r with ELOSS=on to determine momentum loss for
>> protons and electrons (only significant for protons).
>
> I have to do this with ELOSS on AND off to determine momentum loss for
> protons. which is already done.
>
>
>> -correct measured data for momentum losses (relevant only for protons)
>
> I sent out 4th degree polynomials of this momentum loss as a function of
> p and theta a week ago. I am attaching a plot that shows the v16 (without
> extended Bgrid - v17 plot will be available tomorrow) theta_e(theta_p) and
> vice versa. I think we might be back to trusting theta based on these plots.
> There are 8 of them, they are 2D with Profiles on top. The title of each is
> given as:
> T(e/p)_T(e/p)T(p/e)_vs_T(e/p)_(0/1)
> which can be translated as:
> theta_{e/p} - theta_{e/p}(theta_{p/e}) vs. theta_{e/p}
> and the particle's sector (the first term) is 0=L, 1=R
>
> From the profiles it can be seen that the discrepancy is never more than
> +0.2 degrees and it stays positive. Since theta_e(theta_p) and it's inverse
> are always decreasing functions, any overestimation of theta_e/p will give
> the positive discrepancy. This is great because it shows that with our errors
> theta_e + dte - theta_e(theta_p+dtp) at its greatest possible propagated
> error gives us dte ~ 0.14 degrees.
>
> I have been plotting 2D Chi^2 distributions of Beam Energy as a function of
> changes in p_e and theta_e or p_p and theta_p and I'm getting a valley that
> shows great variation as a function of theta but minimal variation as a
> function of p. This seems to be the case up to the point that I'm pushing
> the BeamEnergy sigma(p) outside of a reasonable value.
>
>> -determine residual dp's and dth's by imposing elastic kinematics.
>
> I think: dp(theta_e,theta_p) and 0>dth>-0.15 degrees.
>
> I still don't understand the shift in W from Mascarad, but I'll
> try to look into it.
>
>
> -eugene
>
>
>
>> In case of electrons, this will "effectively" account for the shift
>> (from rad+resol) and for the real "residual" kinematic correction in
>> delta_p.
>> -for all subsequent data analyses, use momentum loss correction +
>> residual corrections from above to correct reconstructed data.
>>
>> Regards,
>>
>> Michael
>>
>>
>> On Sat, 25 Mar 2006, Chi Zhang wrote:
>>
>>>
>>> Hi Michael
>>>
>>> I can not agree with your reasoning.
>>>
>>> The reasoning is only valid if we generate the delta peak and the
>> tail. The
>>> MASCARAD channel in DGen only produces the tail. There is a channel
>> that
>>> mixes epel and MASCARAD but there is a problem of what fraction of
>> events are
>>> generated in either channel. I do not think we have resolved the
>> problem of
>>> how to make the admixture of pure elastic and radiate elastic yet.
>>>
>>> The bottom line is: without the resolutions, i.e. with the original
>> pml/r
>>> variables, there shall NOT have any shift. This is NOT the case if one
>>
>>> generates with MASCARAD channel alone in DGen. This will be the case
>> if the
>>> mixed channel of epel and mascarad is used but I do not know how to
>> interpret
>>> the results, both in terms of the cross section and asymmetries.
>>>
>>> we are comparing two cases:
>>> no shift originally + resolution = shift (Bill T's case)
>>> significant shift originally + resolution = more shift
>>> I am not convinced that we did not put ourself in the second case.
>>>
>>> Since we are trying to derive the kinematic corrections, which
>> themselves are
>>> at the order of couple tens of MeV, a 10MeV shift in the ORIGINALLY
>> TOSSED
>>> events is significant.
>>>
>>> Chi
>>>
>>> On Thu, 23 Mar 2006, Michael Kohl wrote:
>>>
>>>> Chi,
>>>>
>>>> as long as the cutoff energy (10MeV) is much smaller than the
>> resolution
>>>> (30MeV) the resulting Mascarad-W is very well approximating the real
>> one.
>>>> Therefore, if we see a shift in the Mascarad centroid, that's
>> expected!!
>>>>
>>>> Imagine the Blast resolution was 1 MeV, then one would have to choose
>> a
>>>> cutoff energy of significantly less than 1 MeV and hence the centroid
>> would
>>>> be shifted significantly less. This is exactly what we had discussed
>>
>>>> several times, remember BillT reporting at the January collaboration
>>
>>>> meeting:
>>>> "The higher the resolution with which we looked at the scattered
>> electron
>>>> energy, the less of a shift we actually saw ... however the centroid
>> does
>>>> appear shifted because the radiation tail is convoluted with the
>> resolution
>>>> ... and the amount of shift is resolution-dependent"
>>>>
>>>>
>>>>>> -What we really need is pml,r(before RAD+ELOSS) - pwl,r(after
>> RAD+ELOSS)
>>>>>
>>>>> This is impossible. the best you can do is:
>>>>> pwl,r(before RAD+ELOSS) - pwl,r(after RAD+ELOSS)
>>>>> after binning and averaging pwl,r
>>>>
>>>> Can we use the "delta peak" for pml,r(before RAD+ELOSS)?
>>>> Is pml,r with MASCARAD just the tail, or is it an admixture of the
>> "delta
>>>> peak" and the Mascarad-tail?
>>>>
>>>>
>>>> Regards,
>>>>
>>>> Michael
>>>>
>>>>
>>>>
>>>>
>>>>
>>>> On Thu, 23 Mar 2006, Chi Zhang wrote:
>>>>
>>>>>
>>>>>
>>>>>> -I'm still doubtful about the electron spectrum with and without
>>>>>> Mascarad.
>>>>>> In Adrian's plot "Wcorr.gif" from 2006/03/01 (see meeting minutes)
>> the
>>>>>> radiated elastic peak maximum differs from the unradiated one by
>> ~10-15
>>>>>> MeV. I don't see such a difference in your comparison of MASCARAD
>> on/off.
>>>>>> I have a suspicion why. MASCARAD is used in the generator. Is it
>>>>>> perhaps that pml,r is the generated momentum *after* the event has
>> been
>>>>>> radiated? (Experts may want to comment!!) Then, of course, both
>> pml,r
>>>>>> and pwl,r are shifted by the same amount, and the differences
>> pml,r-pwl,r
>>>>>> must look the same for both MASCARAD on and off.
>>>>>
>>>>> This is the case. when comparing pml-pwl, there is no point to do
>>>>> with/without MASCARAD. The non gaussian shape is a result of the
>> "impact
>>>>> angle" dependent resolution that is simulated in the blastmc. The
>>>>> resolution is worse at very forward angle as you can see form the
>>>>> tossed1.ps. The wiggling, if I remember correctly, is an artifact
>> due to
>>>>> the non-exact reversion of the cubic t2d/d2t relation used in
>>>>> recon/blastmc. It is not present when one uses the linear t2d/d2t.
>> but the
>>>>> size of the wiggles is only 5 MeV anyway.
>>>>>
>>>>>> -What we really need is pml,r(before RAD+ELOSS) - pwl,r(after
>> RAD+ELOSS)
>>>>>
>>>>> This is impossible. the best you can do is:
>>>>> pwl,r(before RAD+ELOSS) - pwl,r(after RAD+ELOSS)
>>>>> after binning and averaging pwl,r
>>>>>
>>>>> I would like to repeat a point I made before. I think we should be
>> careful
>>>>> reading into Adrian's W plot. The mascarad has a minimum photon
>> energy of
>>>>> 10MeV!!!!
>>>>>
>>>>> I repeat, minimum photon energy 10MeV!!!!
>>>>>
>>>>> This means every single event MASCARAD generates emits a photon with
>>
>>>>> energy
>>>>>> =10MeV. so none of them belongs to the original delta peak. but we
>> know
>>>>>> the
>>>>> delta peak persists and is only reduced by about 20%. i.e.
>> significant
>>>>> fraction of the events do not radiate. So all MASCARAD generates is
>> some
>>>>> kind of tail. Then ofcourse the peak is shifted by about 10MeV but
>> this is
>>>>> merely an artifact. if you do kinematic correction based on this
>> artifact
>>>>> ... ...
>>>>>
>>>>> On the other hand, it is not clear to me after a few attmept to
>> understand
>>>>> it, how to reproduce the delta part with MASCARAD. I think MASCARAD
>>
>>>>> experts should gather and give a better view of what is being
>> generated
>>>>> and what it means.
>>>>>
>>>>> Chi
>>>>>
>>>>
>>>>
>>>> +-------------------------------------+--------------------------+
>>>> | Office: | Home: |
>>>> |-------------------------------------|--------------------------|
>>>> | Dr. Michael Kohl | Michael Kohl |
>>>> | Laboratory for Nuclear Science | 5 Ibbetson Street |
>>>> | MIT-Bates Linear Accelerator Center | Somerville, MA 02143 |
>>>> | Middleton, MA 01949 | U.S.A. |
>>>> | U.S.A. | |
>>>> | - - - - - - - - - - - - | - - - - - - - - -|
>>>> | Email: kohlm@mit.edu | K.Michael.Kohl@gmx.de |
>>>> | Work: +1-617-253-9207 | Home: +1-617-629-3147 |
>>>> | Fax: +1-617-253-9599 | Mobile: +1-978-580-4190 |
>>>> | http://blast.lns.mit.edu | |
>>>> +-------------------------------------+--------------------------+
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>
>>
>>
>> +-------------------------------------+--------------------------+
>> | Office: | Home: |
>> |-------------------------------------|--------------------------|
>> | Dr. Michael Kohl | Michael Kohl |
>> | Laboratory for Nuclear Science | 5 Ibbetson Street |
>> | MIT-Bates Linear Accelerator Center | Somerville, MA 02143 |
>> | Middleton, MA 01949 | U.S.A. |
>> | U.S.A. | |
>> | - - - - - - - - - - - - | - - - - - - - - -|
>> | Email: kohlm@mit.edu | K.Michael.Kohl@gmx.de |
>> | Work: +1-617-253-9207 | Home: +1-617-629-3147 |
>> | Fax: +1-617-253-9599 | Mobile: +1-978-580-4190 |
>> | http://blast.lns.mit.edu | |
>> +-------------------------------------+--------------------------+
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>
>
>
> --------------------------------------------------------------------------
> Eugene Geis
> PhD Student, Physics Department, ASU
> Research Affiliate, MIT-Bates Laboratory of Nuclear Science
> eugene.geis@asu.edu
> --------------------------------------------------------------------------
> http://quickreaction.blogspot.com
>
+-------------------------------------+--------------------------+
| Office: | Home: |
|-------------------------------------|--------------------------|
| Dr. Michael Kohl | Michael Kohl |
| Laboratory for Nuclear Science | 5 Ibbetson Street |
| MIT-Bates Linear Accelerator Center | Somerville, MA 02143 |
| Middleton, MA 01949 | U.S.A. |
| U.S.A. | |
| - - - - - - - - - - - - | - - - - - - - - -|
| Email: kohlm@mit.edu | K.Michael.Kohl@gmx.de |
| Work: +1-617-253-9207 | Home: +1-617-629-3147 |
| Fax: +1-617-253-9599 | Mobile: +1-978-580-4190 |
| http://blast.lns.mit.edu | |
+-------------------------------------+--------------------------+
This archive was generated by hypermail 2.1.2 : Mon Feb 24 2014 - 14:07:33 EST