[b1_ana] summary of issues

Sun Jun 9 19:07:35 EDT 2013

I think we should be very very careful about switching back to two cups
or what is said about this in the talk.  Its not a surprise this came up
considering the reader but they were not using an electron beam.  We have 
already looked at this in some detail.  We know that its possible but that
the error in the acceptance from one cup to the next can lead to much larger
false asymmetries, among other issues, and with our observable being so 
small its dangerous.  It is completely possible but would require
exceptionally accurate monte carlo to make an error estimate.  This is why
we are using the proposed method.  Definitely worth further investigation
upon approval, but I just wanted to reiterate why we chose the present 
method.

dustin

On Sun, 9 Jun 2013, O. A. Rondon wrote:

> Hi Karl,
>
> Here are some comments to the proposed actions.
>
> Cheers,
>
> Oscar
>
>
> Karl Slifer wrote:
>> Major Issues
>> -----------------
>>
>> I) The significance and interpretation of our measurement.
>>
>>  Status: The iTAC states it is unlikely we can discriminate between the
>> curves in Fig. 7.  They suggest that we need to half the error bars and
>> increase the number of measured points.
>>
>> Proposed Action:  The iTAC missed the point that *any* non-zero measurement
>> of b1 at any of our x points would unambiguously confirm its
>> non-conventional nature.  We can also cite the Theory TAC which states that
>> "*With very little experimental information currently available, any new
>> data on b1 would be clearly welcome*."  If we wanted to directly respond to
>> the iTAC's critique, we can also possibly put together an estimate of what
>> it would take to half the error bar and increase the points (Pzz=40%,50%?,
>> xx additional days of beam?).
>
> Main point is to emphasize that currently there are only three points
> with sizable errors in the x range of interest. Moreover, we aren't
> trying to discriminate among models, just to confirm whether b1 < 0 is
> real, something totally unexpected,
>
>>
>>
>> II) Stability of the target thickness, packing fraction, and dilution factor
>>
>>  Status : iTac is concerned about annealing, and draining helium to
>> depolarize.
>>
>>  Action : We need to make clear that each pol/unpol cyle is an independent
>> measurement, and annealing does not impact each "slug". We need to make
>> explicitly clear the specific sequencing and how each slug is combined to
>> get the final results.  The most efficient way to depolarize is to drain
>> helium, and we have convincing data to defend the use of this method, but I
>> think it is probably best to just agree to depolarize the target with
>> microwaves or AFP (we already say we'll do this in the proposal pg 30). I
>> don't think it is productive to argue about how safe draining Helium is
>> when we have practical alternatives.   Dustin talks about target thickness
>> in the note. Luminosity can be monitored with lumi or with the unpolarized
>> yield but to what precision?
>>
>
> I think we need to explain how the material is contained in the cups and
> how the rasters work in averaging the length. All the needed stuff not
> already in the draft response is in the draft on target thickness that I
> posted, see below.
>
> The point is that only those who have used the UVA target at high
> current with the slow raster are familiar with these details, so we
> should explain them.
>
> Latex here
> http://twist.phys.virginia.edu/~or/b1/tac-itac-pf.tex
> pdf here, to read what I said
> http://twist.phys.virginia.edu/~or/b1/tac-itac-pf.pdf
>
> Ewa mentions measuring pf. We can say that the dilution factor is a
> scale factor, so its error is relative to Azz, and that we determine the
> dilution factor by the ratio of simulated radiated rates on D to total
> rates. The pf is calculated by interpolating the ND3 measured rates
> between simulated rates for different pf's, with the simulation
> calibrated to the measured rates on the C targets of known thickness.
> More details here (RSS tech note)
> https://userweb.jlab.org/~jones/rss/dfpf_writeup.pdf
>
>>
>> III) Systematic errors (particularly from drift)
>>
>> Status: TAC says our systematic analysis is "on point" but lists several
>> systematics it is concerned about.  TAC also explicitly states that it is
>> possible to handle these with ugrades and sufficient committement from our
>> collab.  Most of the questions from our PAC readers are about systematics.
>>
>> Action:  Make list of the upgrades (temp stabilized bpm, new faraday cup,
>> lumi) which Dave suggested (and listed himself as contact).  Ask Dave to
>> provide whatever details we can get before the PAC so we can show a few
>> slides on the new hardware.  In the collab, identify who will take
>> leadership on this,  and point out that having a team work on this is
>> similar to how we handled things in g2p (which also had heavy target
>> commitment, but still had independent groups working on several other
>> projects).   Reference Dustin's detailed note to assess value of the pieces
>> not explicitly discussed in the proposal.  Address the new comments from
>> the PAC readers in the note.
>>
>> Comment: I think in Dustin's note we should summarize the new value
>> (6*10^-4) in the abstract and explain why this is less than Eq 30 of the
>> proposal (further studies and input from the TAC + sqrt(N) from multiple
>> cycles).  An additional column should be added to Table 1 of Dustin's note
>> with the factor sqrt(N) included.  Also a big point that should be
>> emphasized in the abstract is Dustin's conclusion (from Eq 1) that dAzz
>> reduces by the same factor that we can improve Pzz. Finally, the abstract
>> should note that we have already started thinking about how to extract Pzz
>> from the NMR line shape.  If Dustin doesn't object, I would like to make
>> these structural changes to the document before we send to the readers.
>>
> First, the 6E-4 is the drift per cycle, eq. 30 is the Azz error from
> drift per cycle, based on the older 1E-4/f/Pzz. So, now, eq. 30's
> equivalent is 11.1E-3, with f=0.27, Pzz=0.2.
>
> I think it should be enough to emphasize that the number 6E-4 is per 20
> hours cycle, and it propagates to error on A_zz from drift, reduced by
> the sqrt(number of cycles), and that for the low x point the cycle is 10
> h, not 20.
>
> One important detail is that 6E-4 is an overestimate, as explained by JP
> at the last meeting, since the individual components of the yield drift
> are really being double counted. But overestimating is OK, up to a point.
>
>>
>> Secondary Issues
>> -------------------------
>>
>> 1)Value of Pzz without hole burning.  (This could be considered a major
>> issue, but I feel pretty confident we can defend large values of Pzz)
>>
>> Status : We claim Pzz=20%, Tac claims 12%.  Strictly speaking the TAC is
>> right and 12% could be considered the "safe" value.  Our 20% value depends
>> on some incremental improvements to the existing target, but we stated as
>> much in the proposal.  In particular, we were assuming using a larger
>> magnetic field.  We also have some refs in the literature that support the
>> 20% value, but none of them are quite exactly the same as the proposed
>> conditions.  For example, Z. Phys. C -Particles and Fields 49, 175-185
>> (1991) ran a very similar target in an electron beam with <Pzz>=22.5%, but
>> with a 3He/4He fridge.  On the other hand, TAC/iTAC are very confident that
>> hole-burning will produce large Pzz, and we have strong support from JLab
>> target group that this will work also.
>>
>> Proposed Action: We could defend our 20% value with the citations from the
>> literature along with a display of how the vector polarization improves
>> with B, and a list of refs with Pzz ~ 20%, but this seems
>> counter-productive to waste much time arguing for the lower value of Pzz,
>> when the TAC/iTac seem to be pushing us to Pzz=30%.  We do need to prepare
>> some slides with the description of the basic process of hole burning and
>> the evidence indicating it should be successful.  We also have to make
>> clear that there will be no improvement in Pzz without an approved
>> experiment to justify the R&D.
>>
>>
>> 2)Parity violating Asymmetries
>>
>>  Status: This issue was noted by the TAC, but they also suggested that they
>> will be small and manageable.  This conclusion was confirmed by Wally.
>>  Oscar investigated it in detail, and the results are summarized in
>> technote by Ellie/Oscar.
>>
>> Action: The TAC already says it will be manageable.  We should cite the
>> technote as supporting document and have our study results ready as backup
>> slides.
>>
>
> Just mention in the main response that only Pb dependent A_PV is
> important, that Pb = 0, and that if Pb.ne.0, A_PV is the number on the
> E155x paper: 8E-5Q^2, or A_PV < 4E-4 for the proposal. It's OK to cite
> the tech notes, but I don't think they'll have time for them.
>
> There is no target-only A_PV.
>
>
>> 3)Polarized vs Unpolarized beam
>>
>> Status : Unpolarized beam simplifies the analysis, but we are fine with
>> polarized beam as long as we rely on feedback from one of the other halls
>> which are running at larger current. Also have to take some special care,
>> to cut around beam trips for example, but using polarized beam for
>> unpolarized experiments has been done successfully several times.
>>
>> Action: We need a few plots from Transversity to have as back up, but the
>> TAC/iTAC states its not a problem so no action is needed.  JP pointed out
>> that parity feedback will allow control of the (non)polarization of the
>> beam to the 10^-5 level. Do we have a plot to show this?
>>
>>
>> 4)Alignment of vector polarization along beamline vs. q-vector.
>>
>>  Status:  We've investigated this and found it to be negligible (at worst,
>> a small multiplicative factor to Pzz at largest x point ).
>>
>>  Action: This was never even raised by the TAC/iTAC.  We can include as
>> backup slides.
>>
>
> The main point here is this: Azz depends on some types of form factors
> (F_T_20, F_L_20, ... b1 may be a low x F_T_20). Getting to the form
> factors involves the rotation from the beam to the q vector, but Azz
> itself can be measured in any direction. Of course, other contributions
> than b1 may mix in, but, as I said, if we want to take data with SHMS
> and HMS, our only option is longitudinal field.
>
> Also, the rotation is cos(angle) with angle < ~ 20 deg, so cos ~ 1.
>
> The tech note on extra asymmetries discusses much of this.
>>
>>
>
> One final comment: Ewa mentions the two cell method. I still think it's
> the best. I think we should mention that we have considered it and that
> we would go for it if we conclude that the best way to do the experiment
> is to build a new target with a a high field solenoid, but that our
> initial approach is to use existing equipment.
>
> Just as background about Ewa's comment, the COMPASS target is a three
> cell in-line set of targets, 30 cm - 60 cm - 30 cm, which can be
> combined to cancel all linear drifts, and only require reasonable
> corrections for acceptance and packing fraction.
>
> By the way, I also realized recently that the downstream cup receiving a
> different flux than the upstream one is really no problem, because what
> matters is that it's the same beam. We analyze events as being produced
> by primary beam electrons, not by secondary particles. The later are
> related to the radiative corrections, but we need to do external RC's
> anyway, and they can be controlled substantially by the planned switches
> in order of polarization.
>
>>
>> ---
>> Karl J. Slifer
>> Assistant Professor
>> University of New Hampshire
>> Telephone : 603-722-0695
>>
>>
>>
>> ------------------------------------------------------------------------
>>
>> _______________________________________________
>> b1_ana mailing list
>> b1_ana at jlab.org
>> https://mailman.jlab.org/mailman/listinfo/b1_ana
>
>
> _______________________________________________
> b1_ana mailing list
> b1_ana at jlab.org
> https://mailman.jlab.org/mailman/listinfo/b1_ana
>