[Halld-tagger] Photon flux calibration

Wed Jul 11 14:03:56 EDT 2012

Hi,
What we can measure is the separate probabilities of having a
TAC hit or PS hit when a trigger occurs in the tagger microscope channel.
The P(PS*tagger) thus can be measured with a thin radiator at higher
beam currents than the P(TAC*tagger) measurement since
the photon producing a pair will not produce a hit in the TAC anyway.
In other words, the TAC and PS can be relatively normalized using different
runs at different currents. This means we can have two types of 
normalization runs:

o low flux normalization runs with "thin" radiator to get P(TAC*tagger)
o high flux normalization runs with "thick" radiator to get P(PS*tagger)

As Eugene pointed out, this will fail if the tagger counter involved 
develops
non-linearity with respect to the beam-current
(PS counter linearity has to be assumed in case of simultaneous
measurement as well). Therefore, one would need to study this
nonlinearity and take it out using measurements at different beam currents.
Basically, one can gain statistical precision at the expense of adding 
some systematic
uncertainty.

Hovanes.

On 07/11/2012 02:04 AM, Eugene Chudakov wrote:
> Hi Mark,
>
> Yes, as far as I understand it, all the measurements discussed are
> done in coincidence with a signal in the tagger, separately in each
> small energy bin.
>
> a) The photon flux after the radiator and the rate in the tagger are
> irrelevant, apart from some second-order effects with the accidentals.
>
> b) Only interactions of the beam photons in coincidence with the
> tagger are used.
>
> c) For example, in order to measure the eta cross section, we need to
> find the eta production rate per incident photon:
>
> N(eta*tagger)/(N(PS*tagger)*K(TAC*tagger)/K(PS*tagger))
>
> Here K are the rates in a special run with a low beam intensity.  The
> collimator acceptance is irrelevant, the tagger efficiency (the
> probability of the radiated electron to be detected in the tagger for
> a beam photon in the hall) cancels out separately for the low
> intensity run and for the main run (they may be different).
>
> This scheme does not seem sensitive to the way the low intensity is
> obtained.  What matters is the stability of the PS and its sensitivity
> to the photon beam spot profile. The latter can be simulated and later
> measured with a harp. The former needs monitoring - are there good
> methods? One should also optimize the thresholds for the TAC and PS.
>
> The rate K(PS) is limited by the capability of the TAC. Assuming a 1nA
> on a 1.e-4 radiator and a collimator acceptance of 0.3 one gets about
> 0.5MHz of photons E>1GeV. Let us assume it is OK. Let us take 50MeV energy
> bins, the PS converter of 1.e-3 conversion probability, a 20% PS
> acceptance.  The PS rate per bin would be about 0.2Hz. In 1h one
> gets 700 events/bin.  The K(TAC)/K(PS) should be a smooth curve, the
> stat. accuracy seems to be OK.
>
> As I have said, I do not see why not to use a thin radiator at 50nA
> instead of 1e-4 one at 1nA. Can one use a 10-20um carbon thread?
>
> Eugene
>
>
> ------------------------------------------------------
> Eugene Chudakov
> http://www.jlab.org/~gen
> phone (757) 269 6959  fax (757) 269 6331
> Thomas Jefferson National Accelerator Facility
> 12000 Jefferson Ave,
> Newport News, VA 23606 USA
>
> On Tue, 10 Jul 2012, Mark M. Ito wrote:
>
>> Folks,
>>
>> The following may be known to all, but I thought that the discussion
>> might benefit from a restatement of "tradition."
>>
>> As Richard mentioned, having the tagger in coincidence is key to the
>> traditional normalization scheme for tagged photon beams.
>>
>> The main advantage of the traditional approach is  that it has no
>> dependence on the efficiency of a particular tagger counter. Those
>> efficiencies are unknowns. One does not even bother to measure them.
>>
>> The basic assumption is that there are bins in photon energy
>> corresponding to the acceptance of particular tagger counters. For each
>> of these bins in photon energy, the corresponding tagger counter is
>> _required_ for any valid event. The attitude then for each bin is that
>> if the corresponding tagger counter does not fire, it is not a photon.
>> With that attitude,  the beam rate in that energy bin is _defined_ as
>> the rate in the corresponding tagger counter. And nicely, with this
>> attitude, the efficiency of the tagger counter does not matter for
>> normalization; the beam rate determination reduces to measuring the rate
>> in the relevant tagger counters.
>>
>> The only thing that defeats this scheme are the cases where the tagger
>> counter fires, and thus we have a valid traditional photon, but that
>> photon does not get to the target. One standard scenario is that the
>> photon hits a collimator. The probability that things work out, and the
>> photon makes it to the target, is traditionally called the tagging
>> efficiency. Note that this is an unfortunate terminology; the tagging
>> efficiency has nothing to do with the efficiency of the tagger! Some
>> have opted to use the term "tagging ratio" instead to avoid this confusion.
>>
>> The total absorption counter method purports to measure all photons that
>> make it to the target, and thus to measure the tagging ratio directly.
>> But it suffers from the requirement that the beam rate needs to be low
>> enough not the blow the TAC out of the water, as we all realize.
>> Traditionally (again, yikes!) the pair spectrometer is used as a
>> relative rate monitor; its rate relative to the TAC is measured in the
>> TAC runs, and thus, properly scaled, it can act as a proxy for the TAC
>> at standard running rates.
>>
>>    -- Mark
>>
>>
>> On 07/10/2012 07:55 PM, Eugene Chudakov wrote:
>>> Richard,
>>>
>>> Basically, the PS is doing the same thing as the TAC (they see the
>>> same beam), but the TAC's efficiency is about 100% while the PS-es
>>> efficiency is unknown to a 1% accuracy, but can be calibrated, say
>>> with the 1nA beam.
>>>
>>> The acceptance of the collimator and the efficiency of the tagger may
>>> be different for different radiators (I suppose it will not be a big
>>> factor, perhaps a few percents). However, I do not see why the ratio
>>> of the tagged rates in the TAC and in the PS (for a given tagger
>>> energy bin) should depend on the radiator, apart from small
>>> geometrical effects associated with the beam spot in the PS. I do not
>>> assume that using a thinner radiator would distort the tagger energy
>>> measurement with respect to the normal radiator. The energy dependence
>>> of the correction factor must be a smooth function anyway. We just
>>> need to calibrate the PS with the TAC using some radiator, and the
>>> calibration should work for another radiator. I may be missing
>>> something - please explain.
>>>
>>> Eugene
>>>
>>> ------------------------------------------------------
>>> Eugene Chudakov
>>> phone (757) 269 6959  fax (757) 269 6331
>>> Thomas Jefferson National Accelerator Facility
>>> 12000 Jefferson Ave,
>>> Newport News, VA 23606 USA
>>>
>>> On Tue, 10 Jul 2012, Richard Jones wrote:
>>>
>>>> Eugene,
>>>>
>>>> One needs to measure the ratio of the pair spectrometer rate to the TAC
>>>> counter *for a particular set of beam photon populations.*  The populations
>>>> are defined by those beam photons that are in coincidence with each of the
>>>> tagger detector channels.  None of this is meaningful without the tagger in
>>>> coincidence.  As soon as you change the radiator, the population being
>>>> selected by the tagger coincidences changes.
>>>>
>>>> -Richard Jones
>>>>
>>>>
>>>>
>>>>
>>>> On 7/10/2012 10:52 AM, Eugene Chudakov wrote:
>>>>> Hi,
>>>>>
>>>>> The yesterday's discussion on the photon flux calibration did not
>>>>> convince me that one desperately needs a 1nA current.
>>>>>
>>>>> One needs to measure the ratio of the pair spectrometer rate to the
>>>>> total absorption counter rate (for a given energy bin in the tagger).
>>>>> This ratio should not be very sensitive to the type of the
>>>>> radiator. Both detectors see the same photon beam. So, instead of
>>>>> using a 1nA beam current run one may use a thin radiator or a scanning
>>>>> wire with a 50nA run. I suppose it is easy to simulate the acceptance
>>>>> of the pair spectrometer to find out what would be the dependence on
>>>>> reasonable shifts in the beam spot profile (say, a 20% variation of
>>>>> the radiator thickness across the beam).  One should also keep in mind
>>>>> that a low current beam might have a different profile with respect to
>>>>> the full current beam, so this kind of uncertainty always exists.
>>>>>
>>>>> Eugene
>>>>>
>>>>>
>>>>> On Mon, 9 Jul 2012, Richard Jones wrote:
>>>>>
>>>>>> Hello,
>>>>>>
>>>>>> Please remember our biweekly working group meeting this morning at
>>>>>> 11:30EST.
>>>>>> The draft agenda is posted in the usual place.  Please install links in
>>>>>> the
>>>>>> agenda page to any materials that you will be presenting.
>>>>>>
>>>>>> -Richard J.
>>>>>>
>>>>>>
>>>>> _______________________________________________
>>>>> Halld-tagger mailing list
>>>>> Halld-tagger at jlab.org
>>>>> https://mailman.jlab.org/mailman/listinfo/halld-tagger
>>> _______________________________________________
>>> Halld-tagger mailing list
>>> Halld-tagger at jlab.org
>>> https://mailman.jlab.org/mailman/listinfo/halld-tagger
>> _______________________________________________
>> Halld-tagger mailing list
>> Halld-tagger at jlab.org
>> https://mailman.jlab.org/mailman/listinfo/halld-tagger
>>
> _______________________________________________
> Halld-tagger mailing list
> Halld-tagger at jlab.org
> https://mailman.jlab.org/mailman/listinfo/halld-tagger