[esa-t474] 3rd Draft of Paper

[Woods, Michael B.]

Hi Mark, a lot of good progress!  I still have a number of issues though and think it needs some more work.  Would be good if we could schedule a 2-hr block of time this week for a meeting for several of us to discuss.  Some comments (especially note comment 15 regarding calibration procedure and suggestion to modify):

1.  author list -- R. Arnold (not Arnolds); should include Doug McCormick (see notes below on mover system)

2.  lots  of spelling errors; need to run spellcheck

2.5  should consider having an index

3.  Introduction.  On p.3, centering and stability of the beam orbit to 1 micron in the Linac is not needed to maximize acceleration; presumably it's for wakefields and emittance growth (Chris Adolphsen should comment)

4.  Table 1.  leave out line on polarization, since we aren't using that (not running with polarized gun).  For bunch charge in ESA, should probably give 1.6e10 (or whatever we ran with in july 2006)

5.  Section 2 on ESA beam
- I need to give you some text here (still, promise soon really!) borrowing some from Frank Jackson (Technote 2006-1, http://www-project.slac.stanford.edu/ilc/testfac/ESA/TechNotes/TN-2006-1.pdf)
- Figure 2; would be good to clean up text on figure.  Use A28, A29 for xcor28 and xcor29 and similarly a32 and a33.  Use C1 for first collimator and C2 for second collimator.  Use WS1 and WS2 for wirescanners.  Move "to beam dump east" text lower.  Then in caption, include "C1 and C2 are collimators.  WS1 and WS2 are wire scanners.  A28 (29) and A32 (A33) are horizontal (vertical) correctors.
- in last paragraph of 2.2, I believe the jitter was much smaller than 190 microns in x (probably about 50 microns in ESA at 3bpm1,2 and maybe about 80 microns at 31,32 in aline).  Will want to quote spotsizes and emittances too.  I can work with Frank on text for the beam section.

6.  Section 2.3 Cavity Specifications (suggest Cavity BPM Specifications)
- "as described in Section 2.2" rather than "as mentioned
- Table 2; would think that 31,32 apertures are 51 mm
- 2.3.1:  "BPMs 31, 32, 1 and 2 were built for the A-line."  ...(see Figure 4)
- give part # / description for tuners;  at this point for the slac bpms should reference paper by Yury and Whittum (ref 14)
- 2.3.2, "thus reducing monopole contamination of the dipole mode..."  also, regarding the lower Q value just simply state that the Q factor is ~500 and is chosen to be relatively low so as to achieve short enough decay time to resolve single bunches with 300ns spacing at the ILC (I don't see that signal analysis and processing become more difficult, though the resolution will be a little worse)

7.  2.4  Processing Electronics (suggest BPM Processing Electronics)
- Figure 6;  should give part numbers for devices shown; need description for variable attenuator and phase shifter before first BPF; need description for the test port for calibration signals and limiter; need to show final attenuator before adc?
- spllitting of signals for 31,32,1,2 need a little more description (Yury or I can probably do this)

8.  2.5 Data Acquisition
- should have Zen review 
- note that SIS digitizers are VME modules;  
- other 10hz event data recorded included 3 bad spill monitors (3c1, 3wake1 and 3c2) and the ws1 and ws2 pmts;  these should be used for beam quality cuts

9.  2.6 
- should get parts specification and terse description from Doug McCormick

10.  2.7
- should provide a schematic for the interferometer to indicate light paths, beam splitters, retroreflectors etc and give parts list
- Table 3; should have this as a single 3-column table.  In caption, suggest for last sentence "The errors quoted are the spread of the rms jitter measured in 20 blocks of 1000 events within a 30-minute run (run 1421)."

11.  3.1
I get confused in this section with which voltage and frequencies we are discussing; would be good to add references to figure 6.  Equation 10 is stated to be at the bpm output coupler, but then after that equation is a sentence that "...waveform described by equation 10 is what is recored at the digitizer."  what about the IQ mixer shown in Figure 6?  This paragraph should then make clear that the second downconversion is done in software and call it digital down conversion DDC.  
- figure 9, should label as a), b), c)
- would be good if paper gave some results for capability of bunch tilt measurements, noting its importance at ilc for diagnosing wakefields; perhaps there are some references?

12.  3.2
- is this the first time precision and accuracy are mentioned?  Need to define!  For precision would prefer instead to use resolution; isn't that much more commonly used?  Accuracy seems to refer to the stability of BPM residuals.
- for these 5 systematics, in results section would be good to have a table summarizing their contributions
- in 3.2.2, "both" appears twice in 3rd sentence of 2nd paragraph

13.  4.1
- need to explain the different frequencies, starting with the BPM frequency and how the ~36MHz frequencies in Table 5 arise from sampling the 83MHz downmixed signals at 119MHz
- figure 10, should note this is the stability of the mean frequency listed in Table 5.  horizontal axis for this and other related 20-hour plots should be like Fig. 23.  Since this can be done independent of calibrations, would be better to show more results distributed throughout 20 hours.  Should show correlation with bpm10 temperature or refer to Fig. 30.  
- how did you determine that the large systematics in bpms 1,2 frequency determination were from monopole leakage and x,y coupling?  
- for table 6 and associated text on p.19, need to clarify what precision and accuracy are

14.  4.2.1  
- "...After computing the change in beam position due to the change in corrector current, the measured position was plotted against the predicted position... the slope of which was used to determine the scale factor."
-" ... in the corrector dipoles themselves."

15.  4.2.2
- "...using data from the two neighbouring BPMs, the beam jitter and drift could be removed ..."
- Figure 13, one plot does not have axes labeled
- the SVD gets used to tweak the relative calibration constants for the "prediction" bpms with respect to the bpm under test; here it's described for bpm4 using predictions from bpms3 and 5.  Seems that when used in combination with the mover scan, this should give good accurate calibration constants not just for bpm4 but also for bpms 3 and 5.  Can then apply this to other bpms; ex. Use bpms 1 and 9 instead of 3,5.  why can't we use beam jitter plus the bpm4 mover calibration to get all the calibration constants for all the bpms?  

16.  4.3.1
- it's noted that variation in trigger time was negligible.  Would be good to quantify this.
- Table 9; results for bpm4?  

17.  4.5
- need to quantify results.  Linearity to 1%? 0.1%?  Should show a signed result in Fig. 14 and perhaps show a plot of residuals wrt a linear fit.  Which bpms are plotted?

18.  5.2
- should use standard set of axes in plots for the 20-hour run
- for figure 16, should comment why bpm4 results are so stable;  presumably since just use the mover determined scale;  indicates that similar stability should be achievable at least for 3,5 and likely also for 9-11.  I think by including a modified calibration procedure described above in comment 15 might achieve this.

19.  5.3
- prefer a different title for section
- table 11; why large values for precision for bpm1 and bpm11?
- figure 19, should label figures a and b (this happens for a number of figures); 2nd plot should be 41y?  Should comment about not being able to remove angle effects with only 2 bpms?
- need more explanation and plot of frequency change in x11
- need more discussion/investigation of jumps in residual stability plots in fig. 23, 25, 26.  why are the rms blown up to 20 microns for x10,y10 etc. 

20  5.4
- "one of the primary goals of the commissioning was to investigate the dependence ..."

21.  5.4.2
- give part numbers and references for flux gate monitors; also measurement results -- how stable?
- give equation for prediction of orbit change due to energy change

22.  5.4.3
- figure 33 refers to top and bottom plots, but only 1 plot is shown
- for the 400 micron drift observed, show calculation for expected drift versus temperature and plot correlation with temperature.  Seems too large?  Must be 400nm?

23.  5.4.4
- would like to see plots of linked residual versus x4, y4, energy, Q; need to bin data in figures 34-36

24.  no mention was made of neslab chiller to stabilize bpm cavity temperatures; should note this in hardware section

25.  no mention is made of beam quality cuts; do we really make no cuts on beam data?  Would think we would want cuts on beam loss monitors like the bad spill counters and the wire scanner pmt signals; also cuts on energy bpm and on pulses with unexpected large orbit deviations 

mike

 

-----Original Message-----
From: Mark Slater [mailto:slater at hep.phy.cam.ac.uk] 
Sent: Monday, September 10, 2007 4:02 PM
To: esa-t474 at hep.ucl.ac.uk
Subject: [esa-t474] 3rd Draft of Paper

Dear All,

	I've just uploaded the 3rd draft of the paper to UCL's CVS:

http://cvs.hep.ucl.ac.uk/viewcvs/papers/esaBpmNote/built/?root=ILC+Accelerator+code

(note the change in address)

This includes almost all the comments discussed at SLAC during the July run with a couple of exceptions, most notably that I could not find a way of improving the stability by performing the SVD for every set of events as Yury was suggesting. It turns out though, that I doubt this would help as we seem to be dominated (with a couple of exceptions) by changes in scale based on temperature changes than from IQ Phase variations. However that said, we can now demonstrate stability of +-100nm over 8 hours which is better than a kick in the teeth :)

Additionally, I believe the paper is now self consistent and as much as is possible, explains the results that we've reported. However, PLEASE CHECK THIS!!! If you could spend some time going over the paper and make sure there is nothing missing/contradictory/incorrect/just plain dumb. I've now got to the point where I can't really see the wood for the trees and so will probably have missed several things.

The plan now is to give 10-14 days for everyone to go over the paper, after which, if you could send me comments, that would be very much appreciated. Any UK people who can make it (and anyone else for that
matter!) will then hopefully devote a full day to going through and making sure everyone's happy. If all goes well, I will then submit the paper to NIM by the end of September.

I hope this all makes sense,

Many Thanks,

Mark

_______________________________________________
esa-t474 mailing list
esa-t474 at hep.ucl.ac.uk
https://mail.hep.ucl.ac.uk/mailman/listinfo/esa-t474
-------------- next part --------------
An HTML attachment was scrubbed...
URL: https://mail.hep.ucl.ac.uk/pipermail/esa-t474/attachments/20070923/78958fba/attachment-0001.html