Changes to Chaplin et al. manuscript on the VUV and SXR measurements at SSX vernon_v8.pdf D. Cohen 25 August 2008 - The abstract has changed somewhat, including a re-write of the SXR results near the end. - Tim, MB, on the rhs of p. 1 we'd previously described the radius of the SSX barrel as 0.2 to 0.25 m, but I've changed that to 0.2 m. - I've swapped in a higher fidelity image of the SSX schematic (Fig. 1), and I've also added two white squares at the rough position of the spectral diagnostics, and connect these to the photo of the ports in Fig. 9. How's this? In the caption we talk about the position of the SXR, VUV, and IDS, but the photo in Fig. 9 shows/describes two ports which are assigned to the SXR and IDS, respectively. Is this all OK? - There are several places where I've changed the description of the reconnection timing and timescales (mostly to emphasize that reconnection is not instantaneous); please keep an eye out for them and let me know if you see any inconsistencies or mistakes. - I've removed the old Fig. 2, with the three model spectra showing the validity of the steady-state modeling approach. I've kept the description in the text, but I no longer think the figures add enough to be worth the space. - sec. IIB - Vernon, do we want to add anything to the description of the time-dependent tests? I was thinking maybe something about the assumed initial and final densities. - The new Fig. 2 (old Fig. 3), showing two model spectra in the VUV range (T=20, 40 eV) is the same as it was before (1% C and 99% H), but we could recalculate it, including N and O (though I don't think this is necessary; the point of the figure is to show the two strong carbon lines and their dependence on temperature). I've removed the label for the 229 line. (And by the way, we now do not discuss the 229 "anomaly" at all.) - sec. III - I removed a paragraph that mostly described the wavelength calibration. There are several other places in the paper where I removed between a sentence and a paragraph, mostly in cases where I felt extensive description of our procedures was not necessary. - Fig. 5 - Two VUV traces: I believe these two panels were reversed. I've swapped them now, but Vernon, please have a look and confirm that I was right to switch them. The 155 line in a merging shot must be stronger and higher signal-to-noise than the 97.7 line in a single spheromak shot. - I wonder if we should mention Vernon's recent modeling of the Ly-alpha line strength (ratio to C) in sec. III. Here's Vernon's summary: http://astro.swarthmore.edu/~vchapli1/Results.html (scroll down a bit to the 2 July 2008 entry); He finds that the measured line ratio is consistent with 100:1 H:C ratio. This is right in line with what we expect, but I believe one of you told me that the Ly-alpha signal is dominated by plasma from/near the wall, and so the line ratio may be comparing photons that don't originate in the same plasma. - I have updated the discussion of N impurity levels in sec III (pp. 4 and 5) to include Vernon's new results on the N 124 nm line, thanks to Tim's prodding. A new panel has been swapped into Fig. 6. - Should/can we comment on the source of the carbon (and why it's so much higher-concentration than the O and N; shouldn't there be O in hydrocarbon gunk?)? - P. 6, sec. IVB - I've characterized the IDS resolution as "3 eV for carbon" - is that right? I pulled the number out of my ass. I think it's useful to quote resolution in more practical/physical units than simply a wavelength interval per pixel. - My description of the IDS-derived T_i is more detailed than it used to be. I state that we remove data that shows obvious jet-like, non-Gaussian lineshapes, but I also state that some non-thermal contribution is likely still present in the data, and thus that we should consider T_i to be a "mean ion energy." I also describe the decrease in T_i at t>40 us. Is everyone OK with all this? - I have significantly reworked section V, on the SXR measurements, primarily to make it more positive (I've done similar things, on a smaller scale, in several other parts of the paper). I emphasize now that (a) we measure a burst of x-ray emission starting at the beginning of reconnection, (b) there's evidence for spectral evolution of the x-rays as reconnection progresses (i.e. the channel intensity ratios change), and (c) that we can try to back out gross information about the changes in the spectral energy distribution by looking at the specifics of the response curves of each of the three filters. Note also that my dismissal of the Sn data is much terser than it was in the old version of the manuscript (bottom of p. 7); please read it carefully and let me know if it's OK. - I have made various changes to the figures in this section. We begin by discussing the instrument, so I put the instrument figures first. We then discuss the data, so the data figures go next (I've split them up, making a version with the Sn eliminated, and also fixed the color scheme to match that in the response curve figure). Then we finish up with the Prismspect model figure (which is the new one, including N, that Vernon made last month). Oh, and I've also updated Fig. 9; now we show the version with the pink H recombination emission in the windows (and comment on it in the caption). - Vernon, the data we show in Figs. 11 and 12 is claimed to be from one shot. I wanted to make sure that this is right; and in other places, we describe various figures as representing averages over specific numbers of shots. Could you check these and think about whether I've characterized each one correctly. (One thing - what does it mean to make a monochromator line ratio measurement from 25 shots? Is that 12 for the numerator, let's say, and 13 for the denominator? This raises some possible ambiguity, for example, regarding how we calculate the standard deviations that the 1 sigma error range in T_e vs. time is based on.) - MB, I have not used your exact words in the summary (see your email of July 1), but have written something relatively similar in spirit (I don't think that the T_e peak delay can be due to euqilibration timescales, as Vernon's time-dependent simulations show that these timescales are of order a few microseconds). But read what's there now, and let me know if you'd like to see changes. - MB (anyone else?) we need a grant number to acknowledge. - Vernon, I hate to be a stickler, but the typefaces used in figures should be uniform. The final - SXR modeling - figure uses a sans serif font right now, and I've made the others in Time or Times New Roman. If you don't feel like changing this yourself, perhaps you can remind me where your Origin workspaces are on the lab PC and I can do it.