[d2n-analysis-talk] BPM/Harp Scan calibration (fwd)

[Brad Sawatzky]

Hi Diana,

I expect 0.1--0.5mm should be fine -- in the end it will just 'fuzz out'
the edges of our momentum bins (HRS and BB).  It needs to be checked (of
course) but I doubt it will be a large effect.  Seamus might be able to
give a ballpark answer on how uncertainty in the raster correction
propogates to uncertainty in Q2 for BigBite.  Any thesis using the HRS
should have the same information for those packages.

Arne's comment that the BPMs are stable is essentially a statement that
there is no need to recalibrate against the harps.

        +----//--- MCC
        |
BPM ----+----//--- ADC in BB DAQ
        |
        +----//--- ADC in HRS DAQ

"Calibrating our BPMs to the EPICs" variables is really just identifying
any difference between the coefficient MCC uses and what we use.  That
difference can only come from minor differences in cabling and ADC
response.  At some point (in the distant past) MCC cross calibrated
their BPM signal to 'our' A/B harps to get their coeffs.  So, what we're
really doing by calibrating our BPMs to EPICs is calibrating our BPMs to
a set of Harp scans done at some point(*) by MCC.

(*) It would be nice to know when the harp scans MCC used were done...

FWIW, here's a link to Vince's analysis showing the discrepancies in
late 2008/early 2009:
  http://www.jlab.org/~adaq/halog/html/0811_archive/081101190118.html

There is some evidence to suspect that the downstream intrumentation
package (i.e. Harp B and/or BPM B) may have been shifted during the 3He
target installation.  I don't know if that is taken into account in MCCs
coeffs.

In the end, I think we could look at sieve and elastic data with the
raster on/off in the HRS.  If the holes and y-target reconstruct equally
well then our raster correction and (relative) BPM coeffs are in good
shape.  The absolute momentum resolution, which is a (weak) function of
the absolute interaction point can be cross-checked with elastics data.

I think most of this work has been done, but it needs someone (you?) to
flesh out details and turn my ramblings into a semi-coherent picture.

-- Brad

On Fri, 05 Feb 2010, Diana Parno wrote:

> I see that Vince copied this to Brad, as well, but I thought the rest of 
> you folks would be interested in Vince's response regarding the BPM 
> calibrations (below). Below Vince's response, I have also copied the 
> emails from Arne Freyberger, which Vince kindly sent along to me last 
> night.
> 
> So, we've got a few outstanding questions: what kind of precision do we 
> need on the beam position? Is it worthwhile to try to reproduce Vince's 
> results that the harps are not trustworthy during d2n? What do y'all 
> think?
> 
> Diana
> 
> ---------- Forwarded message ----------
> Date: Thu, 4 Feb 2010 20:18:07 -0500 (EST)

> From: vasulk at jlab.org
> To: Diana Parno <dseymour at andrew.cmu.edu>
> Cc: Brad Sawatzky <brads at jlab.org>
> Subject: Re: BPM/Harp Scan calibration
> 
> 
> Hi Diana,
> 
> I only very reluctantly came to the conclusions in my earlier email.  I
> would prefer to use the Harp information, but it is very clear that the
> Harps have become unreliable.  Brad does have a point that the information
> feed to the ADC's and EPICS read back have a common source.  But MCC has a
> separate calibration procedure from how we have typically done the
> calibration.  So we are using MCC calibrated numbers to calibrate our ADC
> readout.  Not ideal, but I do not think we have any other options that
> will give us the absolute beam position.
> 
> I had several email correspondences with Arne Freyberger, who used to be
> an expert on the Harps.  I can forward you those conversations, since all
> the information I have provided you came from him about the BPM and their
> calibrations.  The claim is that the BPM girder, which includes the harps,
> has not been disturbed and the BPM offsets have been fairly stable for
> years.
> 
> Arne indicated to me that the BPM's are routinely calibrated on MCC's
> side.  I can also send you that email as well.
> 
> As an aside, back when Harp B had it's coupler problem.  I
> cross-calibrated the MCC's EPICS readings versus the Harps positions.  The
> cross-calibration was fairly good.  I have various plots that I can send
> to you.  Then I have the data from the beginning of Transversity where the
> calibration began to breakdown and from d2n, which is rather bad.  The
> good agreement from before Transversity is one of the main reasons I am
> willing to use the EPICS readings.
> 
> Finally how well do you need to know the beam position for d2n?  Arne
> claimed we should know the absolute position to 50 microns.  I am not as
> convinced about this fact.  Typically you only care about this level of
> accuracy when you need great momentum resolution (100's keV).  You do not
> have the resolution on the BigBite side.  The HRS on the other hand you
> do, but you are in DIS kinematics.  The resolution is not important here
> for the cross section.  I think by using the EPICS information, you can
> get at least 0.5 mm accuracy in beam position.  On top of it, we were also
> using a fairly large raster size.
> 
> Perhaps you should do a simulation to test how much a 0.5 mm or even 1 mm
> beam position uncertainty will affect your absolute cross-section
> measurement.
> 
> I will find the emails from Arne and send them to you.
> 
> Vince
> 
> -------------------------
> 
> ---------------------------- Original Message ----------------------------
> Subject: Re: BPM Calibration in Hall A

> From:    "Arne Freyberger" <freyberg at jlab.org>
> Date:    Sat, October 24, 2009 11:26 am
> To:      "Vincent Sulkosky" <vasulk at jlab.org>
> --------------------------------------------------------------------------
> 
> Vincent,
>       The BPMs are quite stable, and well within 0.5mm.  Probably stable
> within 0.05mm,  as long as:
> 
>     1. They are not physically bumped or moved
>     2. the software offsets (SOFF, GOFF) are properly restored after an
>        ioc reboot.
> 
> BPMs physically moving is a rare event associated with beamline work.
> Aside from the compton region your beamline has been relatively
> constant.   The SOFFs and GOFFs   history can be tracked via
> archiver/allsaves to insure that they have been stable as well.  In
> general OPS would notice a large problem with SOFFs and GOFFs changing
> since the relative orbit would not restore after a reboot.
> 
> Yes, the harps have not been re-surveyed.   This should be scheduled in
> the Winter2010 down.
> 
> If you want detailed information on positions, SOFFs and GOFFs, give me
> a list of BPMs and date ranges, I'll ask an operator to dig the data out
> for you.
> Arne
> 
> ---------------------------
> ---------------------------- Original Message ----------------------------
> Subject: Re: BPM Calibration in Hall A

> From:    "Arne Freyberger" <freyberg at jlab.org>
> Date:    Mon, October 26, 2009 10:44 am
> To:      "Vincent Sulkosky" <vasulk at jlab.org>
> --------------------------------------------------------------------------
> 
> Vincent,
> I've attached a spreadsheet that went through our twice-daily snap files
> of the machine.  The entries make it into the spreadsheet only if the
> values are unique, so you can see that:
> 
>     1. The GOFFs are consistently zero, which is what I expected
>     2. the X soffs do not change
>     3. the Y soffs were not in the "snap" script until jan 7th 2009, but
>        since the X soffs did not change I would be surprised if the Y
>        soffs did change
>     4. The snap files between March 25th 2009 and March 30th  read back
>        zeros, but I don't think A was taking data at that time.
> 
> So the values appear to be stable during this time period.
> 
> 
-- 
Brad Sawatzky, PhD <brads at jlab.org>  -<>-  Jefferson Lab / Hall C / C111
  Ph: 757-269-5947 -<>- Pager: 757-584-5947 -<>- Fax: 757-269-7848
The most exciting phrase to hear in science, the one that heralds new
  discoveries, is not "Eureka!" but "That's funny..."   -- Isaac Asimov