[Olympus] BPMs, magnetic field, MIE, etc

Tue Sep 6 15:49:10 EDT 2016

Hi Alexander,
	This is a good point, and not one that we previously considered when estimating MIE systematics. Today, I tried to make a rough estimate of the size of the effect. To summarize, I think it’s definitely a non-negligible effect and a more thorough calculation is warranted. This should certainly be added to the table of systematics. 

0.02% for changes in horizontal position (3µm change, compared to 20 µm quoted uncertainty)
0.17% for changes in horizontal slope (10µrad change, compared to ≈10µrad quoted uncertainty)

Including these systematics would bring the total from 0.27% to ≈0.32%, depending on how a more thorough calculation shakes out.

Details of my estimate:

My approach was to integrate the equations of motion between the BPMs using quadratic fits to the magnetic field measurements from 2011, which are right along the beamline. We don’t have any 2013 measurements between between the BPMs, but I have confidence in the 2011 measurements in that region because the agree very well with the 2013 measurements. In general, the field in that region is on the order of a few Gauss, not 50 G as was suggested in today’s collaboration meeting. I assume the beam position is correct at the BPMs, and then I determine the beam trajectory between those fixed points. Then I look at deviations between the calculated trajectory and our simple linear interpolation assuming that the e- and e+ will deviate in opposite directions. The position difference between species changes from 0 to 6µm over the length of the target, with the average of 3µm at the target center. The slope difference is about 10µrad and is much more consistent over the length of the cell. 

If we wanted to be really fancy, we could put these position deviations into the simulation and try to correct for them. We’d still have some residual uncertainty about the true magnetic field in the target chamber region, but we might be able to argue that our systematics are smaller. However, since it probably won’t be an enormous change in uncertainty, it should not be a high priority. There’s other stuff we need to iron out first, IMO.

Best regards,
Axel

On Sep 1, 2016, at 10:59 , Alexander Kiselev <kisselev at mail.desy.de> wrote:

>  Axel,
> 
>  honestly I was expecting a private answer from you first. Since you posted my e-mail to the list, let me rephrase the question:
> 
>  - if (hypothetically) one measures (0,0) offsets in BPM#1 and BPM#2
> for both e+ and e- beams it does *not* really mean, that both beams had zero offsets and slopes in case there was a noticeable B*dl inbetween the BPMs; it is easy to estimate, that given residual fields of an order of ~1mT (see fig.4-11 in your thesis, though it is Bx), ~1.5m distance and
> 2 GeV/c beam energy we are talking about several dozens of either microradians or microns; this does not look negligible to ~10um accuracy in relative e+ vs e- beam line position determination we claim;
> 
>  -> if the above consideration makes sense at all (I do not insist it
> does; just asking), was this effect estimated, is it small (cancells?) and if not: was it accounted for either in calculating e+ & e- beam line parameters (and then in MIE normalization itself) or in 0.27% MIE normalization systematics estimate?;
> 
>  Thank you,
>    Alexander.
> 
> PS: personally I doubt, that SyMB *alone* can ever give better luminosity normalization than 12 degree monitor with full tracking capability (and yes, I carefully read respective chapters in both Axel's and Brian's theses); since two systems provide statistically consistent instantaneous
> measurements they probably should enter the paper this way (i.e. on equal footing for luminosity determination, with all due comments about possible magnitude of TPE effects at 12 degrees);
> 
> PS: I vaguelly remember, there was a discussion once on how to *cross-calibrate* two luminosity systems using any 4 small sets of data
> {e+-,f+-}; was about "native" SyMB usage mode at that time (not MIE)
> though; based on a simple observation, that flipping magnetic field affects charged tracks in two luminosity systems in a known (and different) way;
> 
> 
> On Thu, 1 Sep 2016, Axel Schmidt wrote:
> 
>> 
>> 
>> Begin forwarded message:
>> 
>>> From: Alexander Kiselev <kisselev at mail.desy.de>
>>> Subject: BPMs, magnetic field, MIE, etc
>>> Date: August 30, 2016 at 17:25:56 EDT
>>> To: <schmidta at mit.edu>
>>> 
>>> Hi Axel,
>>> 
>>> first, congratulations with your defence. Second (being myself a 12 degree monitor person), let me challenge your 0.27% MIE method systematic
>>> uncertainty estimate. Starting with a simple question: was BPM data
>>> "corrected" in some sort for the non-vanishing magnetic field along the beam line between BPM#1 and BPM#2?
>>> 
>>> I mean ~10-20um match between e+ and e- data sounds a bit of a stretch to me, far beyond the capability of a rather simplistic detector we had in hands. With ~1mT residual fields (perhaps not precisely accounted), ~1.5m BPM#1->BPM#2 distance and 2 GeV/c energy you can easily get far off beyond
>>> that when comparing electron and positron *beams*. Which IMHO happens *before* the situation you considered in section 6.4 (fig.6-15).
>>> 
>>> By the way, MIE method itself with all the machinery you developed looks fine to me. I just doubt the absolute luminosity ratio normalization
>>> accuracy one can get from it.
>>> 
>>> Unrelated, concerning fig.8-9: does it mean, that you had to waste ~1/4 or so of the raw data (or do I misinterpret the plot)?
>>> 
>>> Cheers,
>>>   Alexander.
>>> 
>> 
>> -------------- next part --------------
>> An HTML attachment was scrubbed...
>> URL: http://mailman.mit.edu/pipermail/olympus/attachments/20160901/3f0e316d/attachment.html
>> _______________________________________________
>> Olympus mailing list
>> Olympus at mit.edu
>> http://mailman.mit.edu/mailman/listinfo/olympus
>> 