Notes taken on Yerevan preprint Exp-05.pdf Richard Jones, July 29, 2006 1. What is the horizontal scale in Fig. 1? Text says 19.9m between C2 and the PS-6 telescope detectors. Former NIM article says 15.6m between C2 and K3. Is it drawn to scale? Where is the radiator? 2. How are the counters N1-N3 arranged? The text says they make 5 independent energy bins but then only 3 bin widths are listed. It seems that Fig.2 could show the hodoscope geometry. 3. Beam and slit dimensions have changed somewhat relative to the values given in the NIM article. What is the significance of these changes for the MC acceptance of the setup? NIM article Experiment article photon spot at slits 1.4 x 1.4 mm2 1.6 x 1.6 mm2 slit spacing 20 mm 26 mm 4. What is the unit "mkm" in the label "dAL = 20 mkm" in Figs 6-7 of the earlier NIM article? It seems to be German for micrometer. 5. The analyzing power with slit (Fig. 3) is different from the former paper Fig. 5. Both used the same CW form factor. Presumably it is the wider slit opening that makes the new curve somewhat lower. 6. In the case of multiple counters one should define what quantity the asymmetry in Fig. 3 refers to. My guess it is something like (N1left*N1right + N2left*N2right + N1left*N1right) 7. Something is wrong with Fig. 5 as shown. The caption says that dx=0 is the symmetry configuration but the graph is clearly not symmetric with respect to the sign of dx. Compare this with Fig. 4 where the curve is symmetric about z=0. Charge symmetry requires a similar symmetry with respect to right and left shifts in Fig. 5, which implies a zero slope at dx=0. Also there is a general rule that the asymmetry is maximum for symmetric pairs which corresponds to the dx=0 position. Could this plot have been generated for a different pair of counters, say N1right * N2left, so that the symmetry point occurs for dx != 0 ? 8. Adding a length scale on Fig. 6 would be useful. 9. The statement on p.3 that count rates are the same in MC between the two configurations shown in Fig. 6 is puzzling to the reader since the programmer has forced them to be equal by the geometric symmetry. This does not seem to teach us anything. 10. The meaning of Fig. 7 is not clear. The PS-6 detects e+e- particles, not photons. Each coincidence channel corresponds to a summed energy E(+) + E(-). Is that what is meant by "photon energy"? There should be a table showing the different coincidence channels and what central value corresponds to their summed energy under nominal field conditions. Then the caption in Fig. 7 could point to one of the coincidence channels in the table and show how the summed energy of that channel varies with the central field in PS-6. 11. I tried computing the CB polarization for your geometry myself. My diamond orientation was -3.371 mr for theta_par and 100 mr for theta_perp. For the central bin around 1000 MeV I got 68.4% for the peak polarization. How can we understand such a low value as 0.53? Hrachya mentioned this discrepancy in his email, but said that it was what emerged from spectral shape analysis. Does this mean that the incoherent part is significantly more enhanced than from a perfect crystal? If so, we should point this out in the paper and suggest reasons why. One reason I can think of might be that the diamond has some disoriented regions or inclusions that contribute only incoherent intensity. Can you think of any others? I seem to recall that the shape analysis method allows the incoherent level to "float" in the fit. 12. What are you using theta_perp? Maybe I don't understand what you mean by the "peak energy at 1000 MeV". I place the primary coherent edge at 1000 MeV, because that is what the kinematics defines. Where the maximum intensity appears depends on details like emittance, mosaic spread of the diamond, etc. 13. In Fig. 9 it is not clear how you measure a spectrum with such resolution in PS-6 with only three left and three right counters. Also for the PS-30, one expects to see only 30 points but there appears to be many more than 30. Perhaps the answer is that several spectra were taken at different magnetic field settings and then they were combined into one plot. If so, the text should explain how this was done, and if there were any systematics involved in trying to fix the relative normalization of each sub-spectrum. These spectra are very smooth and continuous and give the appearance of having been taken all at once. 14. In Fig. 4, the data points make a good case that the alignment is correct. I do not believe that the accuracy of this determination is as good as 0.1 mm, as stated in the text. With error bars of this size, I would guess something like 0.3 - 0.5 mm is the error on the fit. There are not enough measurements to split that peak by a factor of 100, I would say. 15. There are 7 data points in Fig. 4 which shows that more configuations were tested than just the two in Fig. 6. That leaves Fig. 6 only as illustrating an elementary MC exercise (see point (9)). I suggest that it be replaced with a figure showing the hodoscope geometry for PS-6 and indicating the range of motion of the remote controls on each side. 16. No description is made of how the deltaX alignment was carried out. Only energy calibration and z alignment are described. Small shifts in deltaX in PS-6 would not affect the energy calibration because it would shift the left and right energy channels equally in opposite directions and leave the sum unaffected. This seems to be a missing piece of the setup procedure for the beam line. 17. The mention of comparing rates between an amorphous target and the two crystal orientations does not seem to teach us very much. The two rates would normally be different because the two radiators would not have exactly the same thickness. Even if they did, the spectral shapes of the bremsstrahlung are very different, and there is no physical justification for expecting their integrals to be equal. 18. The measured asymmetry is stated without a range in energy over which it was measured. I guess that the central bin was used which is 20 MeV wide. Is this correct? Also it would help if the diamond angles were listed in the text (or a table) so that the user can compute exactly where the theoretical coherent edge should be. 19. This may be a matter of taste, but since the title of the paper contains the word "polarimetry", perhaps the final result should be a plot of the measured polarization P. That way it can directly compare the results of two polarimetry methods, one direct and the other indirect. I suggest that both Fig. 10 and the final result be stated in that way. We should come up with an estimate for the systematic error on the P(indirect) as well as the existing errors on P(direct). 20. In one of his email messages Hrachya mentioned that the final polarimetry data were taken with a slit setting of 13 mm. From the text I understood that it was 10 mm. Can you recall which it was? 20. I suggest that we add a few experimental details such as the properties of the diamond used (origin, thickness, size) and of the electron beam, such as current, energy spread and approximate value of the emittance, if you know those things. We should also state at what beam current the polarimetry measurements were taken because that could be an important constraint on direct polarimetry.