Collimator Optimization "Live" Website

The corners of the ep profile accepted by the cerenkov bar were found for z locations separated by 18cm. Figures a and b show examples of how the values of the corners were chosen; the red outline shows the edges of the profile as defined by the corners at that location.

The links to detector set #1-5 are postscript files of the profile at all of the locations throughout the experiment. The top link, for the "real" profile, is for the locations upstream of the primary collimator. This would be the profile of electrons that make it through the upstream cleanup collimator.

There are also datafiles containing the values of the corners of the profiles at each z location - one for the accepted profile everywhere and an additional one for the "real" profile upstream of the acceptance defining, or primary, collimator.

The data files have the corners of the profiles and the z values in this format:

x y z

for the corners in this order:

lower right
lower left
middle right
middle left
upper right
upper left

Figure a - An ep profile close to the end of QTOR for events accepted on the quartz bar.	Figure b - An ep profile close to the end of QTOR for all events.

Detector Set #1, "real" ep peak


Detector Set #1


Detector Set #2


Detector Set #3


Detector Set #4


Detector Set #5




Data file with the corners of the profiles for each z location.

Data file with the corners of the profiles for the "real" ep peak in Region I.

Some text describing our optimization procedure.

Optimization Procedure

What I've done so far:

Use a "stripped" down version of the experiment.
Choose z locations for the collimators (one upstream and on downstream).
Put detectors at the z location of possible QTOR or QTOR support structure interferences.
In Paw++, draw the shape of the interferences for the various z locations.
Trim the collimators to clear those interferences.
Check with SolidWorks by putting the ep peak profile through QTOR.

So we've defined the LARGEST it could POSSIBLY be based on the QTOR support structure (see rows 10 and 11 below)


What I need to do:

Refine collimators
- Get a more realistic target attachment structure and/or design one to not interfere.

Consider the detectors
- Refine main detector for 3 shapes/different z locations (study inelastic % vs. FOM)
- Check moller rates
- Consider size of minitorus and realistic supports
- Consider size of GEMs
- Consider size of Region III chambers

Refine collimators
- Check the dose on the QTOR support and other "close" structures, possibly trim ep peak even more.
- possibly change z locations, redesign based on above considerations and summarize pros and cons

Here are the ep profiles for the slightly modified row 14. There are profiles of the accepted ep peak at 15 z locations in Region I starting 10 cm from the end of the target in 10 cm intervals. There is also a data file that contains the points on the corners of the keystone shape for uploading into a 3D CAD program.

There are also ep profiles for Region III at 15 z locations starting with z = 370cm, in increments of 20cm.

Region I/ Target	Region III/Cerenkov Bars
This figure is a link to the .ps file for all of the z locations. The data file. The ntuple.	This figure is a link to the .ps file for all of the z locations. The ntuple.

Investigations of the possibility of using a straight bar rather than a handle or v-shape.

Cerenkov Bars	Cerenkov Bars
	Width (cm) Lower x (cm) ep Rate (MHz) Inelastic Rate (MHz) % Inelastic Ave. Q2 FOM Error on Qweak 16. 310. 359. 0.005 0.001 0.0254 0.2318 5.778 16. 312. 578. 0.070 0.012 0.0258 0.3838 4.785 16. 314. 786. 0.149 0.019 0.0261 0.5373 4.289 16. 316. 894. 0.345 0.039 0.0263 0.6200 4.117 16. 318. 934. 0.691 0.074 0.0264 0.6514 4.063 18. 310. 578. 0.072 0.012 0.0258 0.3838 4.785 18. 312. 786. 0.149 0.019 0.0261 0.5376 4.288 18. 314. 895. 0.345 0.039 0.0264 0.6223 4.114 18. 316. 940. 0.691 0.073 0.0264 0.6573 4.056 18. 318. 964. 1.322 0.137 0.0264 0.6732 4.027 20. 310. 786. 0.152 0.019 0.0261 0.5376 4.288 20. 312. 895. 0.345 0.039 0.0264 0.6226 4.114 20. 314. 941. 0.691 0.073 0.0265 0.6596 4.055 20. 316. 969. 1.322 0.136 0.0265 0.6791 4.021 20. 318. 987. 2.337 0.237 0.0264 0.6897 4.000
This figure has a link to the pictures corresponding to the table on the right.	Table - Inelastic percentage and error on Qweak as a function of bar width and lower x for a 2.2m long bar.

Figures 1a,c are pictures of the ep peak for a downstream and upstream collimators at the z location of one of the major QTOR interferences. The shape of the ep peak has been chosen so that there is ~2cm of clearance from the QTOR support structures at the upstream end. The horizontal bar at the top is actually further upstream, and the corners at low theta/high phi are further downstream and have been "projected" to this z location in order to show the reason for the choice of cuts.

Possible interefences further downstream with QTOR coils and support structure will be checked by putting an ep peak profile into SolidWorks. Also, there may be interferences from other elements in the experiment, such as a reasonably sized beampipe near the target at low theta.

Figure 1a - ep peak at QTOR support location for downstream collimator.	Figure 1b - GEANT picture of downstream collimator.
Figure 1c - ep peak at QTOR support location for upstream collimator.	Figure 1d - GEANT picture of upstream collimator.

These pictures show the SolidWorks version of the QTOR support structure with the ep peak profile through it.

Thanks Norm!

Figure 2a - SolidWorks picture of the upstream interferences.	Figure 2b - SolidWorks picture of the downstream interferences.

The interference at the upstream QTOR support was removed by lowering the upper theta of the center trapezoid (the z location of that interference is incorrect in 1a and c above). The interference at the downstream QTOR support was removed by clipping the upper theta of the "wings" of the ep peak shape (See rows 10 and 11 below).

Figure 3a - ep peak plots at the z location of the downstream interference, with the interference removed for the downstream collimator.	Figure 3b - ep peak plots at the z location of the downstream interference, with the interference removed for the upstream collimator.

WARNING!

The values in this table (to date) are only for rough comparison. The detectors have not been
optimized for each configuration, and only the interferences from QTOR have been considered.

Row #	Date	Description	Collimator	Detector Shape	Elastic Rate (MHz)	Percent Inelastic (%)	< Q2>	Error on Qw	4-panel plot, ntuple	bash and euclid files
1	4-12-05	stripped down version of the experiment; only fields of the magnets, target and primary collimator are present	downstream	handle	822	.013	.0299	4.16	4-panel plot ntuple	.bash .euclid
2	4-12-05	same as above	downstream	v_shaped	815	.025	.0297	4.17	4-panel plot	same as above
3	4-12-05	same as above	upstream	handle	826	.019	.0283	4.17	4-panel plot ntuple	.bash .euclid
4	4-12-05	same as above	upstream	v_shaped	851	.051	.0291	4.17	4-panel plot	same as above
5	4-13-05	Now upstream version has same clearance of QTOR structure as downstream version	upstream	handle	693		.032		4-panel plot ntuple	.bash .euclid
6	4-18-05	Now the lower part of the trapezoid clears the QTOR support structure by 2cm at low theta	upstream	handle	571		.035		4-panel plot ntuple	.bash .euclid
7	4-18-05	same as above	downstream	handle	699		.032		4-panel plot ntuple	.bash .euclid
8	4-20-05	2cm of clearance for the upstream QTOR support structure at low radius, and increased the upper theta.	downstream	handle	938	.009	.0299	4.03	4-panel plot ntuple	.bash .euclid
9	4-21-05	same as above	upstream	handle	756	.048	.0318	4.19	4-panel plot ntuple	.bash .euclid
10	4-26-05	Using the previous version of the collimator, but trim the upper theta of the "wings" so that the interference in the downstream of QTOR is removed and lower the upper theta so the upstream interference is removed.	downstream	handle	931	.016	.0290	4.05	4-panel plot ntuple	.bash .euclid
11	4-26-05	same as above	upstream	handle	722	.041	.0318	4.24	4-panel plot ntuple	.bash .euclid
12	5-10-05	Added detectors near target and put in preliminary target chamber attachments and beamline near the target.	downstream	handle	935		.0290		4-panel plot ntuple	.bash .euclid
13	5-10-05	same as above	upstream	handle	731		.0316		4-panel plot ntuple	.bash .euclid
14	5-20-05	Old favorite! Lowered upper theta based on FOM study - makes collimators smaller without changing error on Qweak. Also have a minitorus that works well.	downstream	handle	895	.011	.0272	4.07	4-panel plot ntuple	.bash .euclid
15	5-20-05	same as above	upstream	handle	703	.038	.0314	4.25	4-panel plot ntuple	.bash .euclid
16	5-25-05	same as row 14, but with a different detector.	downstream	straight bar	957	.092	.0264	4.03	4-panel plot ntuple	.bash .euclid
17	5-25-05	same as row 14, but with minitorus off	downstream	handle	1015		.0249	4.03	4-panel plot ntuple	.bash .euclid
18	5-25-05	start with row 14, move minitorus, change collimator thicknesses, etc.	downstream	straight	935		.0263	~4.1	4-panel plot ntuple	.bash .euclid
19	7-13-05	Version 5, untrimmed collimator	downstream	straight					4-panel plot ntuple	.bash .euclid
20	8-2-2005	collimator trimmed so unradiated profile fits on 18 cm bar, minitorus off get QTOR field map	downstream	straight	875	.042	.0258	4.18*	4-panel plot ntuple	.bash .euclid "fixed" .euclid
21	8-2-2005	same as row 20, minitorus on, get minitorus field map	downstream	straight	765	.047	.0273	4.28*	4-panel plot ntuple	.bash .euclid

* These errors include Dave Mack's factor of 1.03 on the statistical error, though the other entries in the table do not.

Note: The assymetry for row 20 -.237 ppm, while the hadronic fraction 30.2% is and the axial fraction is 3.4%.