[esa-t474] New version of paper available

[Mark Slater]

Dear All

	I have added almost all the corrections indicated by the reviewer.  
Below is a list of the alterations. The only points I have issue with are
(2) where he talks about an artifact for BPM 4x which I didn't realise we
had and point (11) for which I'm going to need someone else to write up,
or at least send me the info so I can write it up.

If people can help with either of these, do let me know ASAP!

However, as I say, the majority were not a problem and I've implemented 
them. You can see find the new version at:

www.hep.phy.cam.ac.uk/~slater/esabpm.pdf
www.hep.phy.cam.ac.uk/~slater/esabpm.ps

the submitted version can be fonud at:

www.hep.phy.cam.ac.uk/~slater/esabpm_old.pdf
www.hep.phy.cam.ac.uk/~slater/esabpm_old.ps

!!NOTE CHANGE OF WEB ADDRESS!!

Thanks,

Mark





Reviewer #1: I believe this is a valuable, relevant to NIM, and very 
detailed manuscript. The community will benefit 

from its publication. I have the
following comments, all of which should be solvable by simple editing:

********************************
1) The discussion starting with line 441 and ending with line 474 is 
confusing. I will attempt to summarize my 

understanding of it: On each shot, eight
pieces of data are collected from the outer two BPM's, including the I and 
Q signals in each direction. The 

correlations between this data and the the I and
Q of the central BPM are analyzed over many shots, such that eighteen 
linear parameters are found that best predict 

the central BPM data (nine parameters for
I, and nine parameters for Q).

If that is the case, then the authors should explicitly state that x_i and 
x'_i in Equation 16 are derived from the I 

and Q signals of one of the outer
BPM's. The phrase "outer BPM coordinates" sounds vague, since coordinates 
could be taken to mean positions only.

Reworded these paragraphs

********************************
2) In several places (line 178 on p. 12, and later, in lines 671 to 674 on 
p. 42, line 587, p. 35), the paper talks 

about unwanted coupling with other modes
in the cavity. BPM 1 is said to have rectangular cavities, and I am 
guessing that the mechanical damage brought the 

detuned orthogonal mode closer to the
filter acceptance window.  Is the artifact in the BPM 4x signal due to the 
same issue? In that case, analysis of the 

residual vs. motion in the y-direction
would give a much clearer indication of the source of the cross-coupling 
(if data is available).


Is he correct in saying that the detuned orthoganal mode was closer to the 
acceptance window of the filter?
There is no artifact in the 4x signal! 

********************************

3) I recommend the description of the SLAC "A-line" (line 82 on p. 6 to 
line 121 on p. 8) to be better "decoupled" 

from the discussion of the ESA beamline.
Initially, I was looking for BPM's 1-9 in the Figure 3 description of the 
A-line, and only realized the general 

layout of the beamline components after
re-reading both sections a couple of times.

Removed most of the ESA references in the A-line section and added another 
sub-heading

********************************

4) Table 7: the fourth entry in the table is "Q cavities". I could not 
find in the text what this refers to. Are 

these the phase reference cavities?

changed to Ref. Cavities

********************************

5) Fig 16 is missing the plot legends.


Fixed

********************************

6) In some of the plots, a label such as "BPM 4x Residual" is easy to 
mistakenly read it as "four times"  the 

residual, thus giving the impression that the
signal was rescaled by a factor of four. A notation such as BPM-4X might 
make it easier to interpret the plots.

All plot labels fixed

********************************

7) The definition of "linked" system resolution starting in line 544, p. 
32 is a bit vague. Assuming a straight line 

trajectory through a set of BPM's, the
residual of each BPM is calculated, which gives up to eight numbers. At 
this point, I am confused about whether a 

single BPM is selected and the RMS value of
its residual is called the "linked" system resolution, or whether a 
further RMS is performed on all the residuals of 

up to eight BPM's in the set. In the
case of a single BPM, how is that BPM selected?

This is incorrect. The residual of a single BPM is calculated from a 
prediction formed from all but the neighbouring 

BPMs. Reworded this to make it clearer.

********************************

8) Table 8 and its discussion would be clearer if: 1) the table column 
headings were called something like "Measured 

resolution, x", and "Predicted
resolution, x", etc. and; 2) The sentence on line 552, p. 33 was changed 
to alert the reader that predicted 

resolution is defined in the next paragraph.

Done

********************************

9) In the discussion of measured system resolution on line 554, p. 33, I 
wonder if the digitizer was operated near 

the peak of its range, or would additional
signal amplification further reduce the digitizer noise?

I have reworded the statement on attenuation optimisation to indicate that 
this is what we do

********************************

10) Figure 18, I can guess that the black points are the mean, and the red 
region is the RMS width, but it should be 

stated explicitly. Also, the meaning of
the green lines is not at all apparent.

Added to captions

********************************

11) In general, the paper would benefit from a statement of the BPM 
sensitivity (signal voltage for a given bunch 

charge and displacement),  typical signal
levels seen at various points in the processing electronics, a more 
detailed description of the electronics 

components used (mixers, amplifier), and a
discussion of the likely sources of electronic noise.

********************************

12) Minor grammatical errors on lines 587, 654, and 656.

Fixed

********************************