This time we used 3 widely spaced FCLS positions to obtain better leverage on the plate scale.
The three FCLS positions that were used with slit 6 (slitwidth = 3.23").
(X1,Z1)= 19.6 mm, 16.7 mm
(X2,Z2)= 25.6 mm, 21.35 mm
(X3,Z3)= 31.6 mm, 26.05 mm
separations:
P1-P2: 7.59 mm --> 31.4"
P1-P3: 15.21 mm --> 63.0"
P2:P3: 7.62 mm --> 31.5"
Here are the three superimposed FCLS positions in the different modes (true camera, echelon camera, medium res and low res).
 |
| True camera mode: slit6, N04679 |
 |
| Echelon Camera mode: slit6, N04679 |
 |
| Medium, 2142cm-1, 12th order (57.91 deg) |
 |
| Low, 2142, 6th order (24.097 deg) |
The separation between FCLS points was:
Interval LOW MEDIUM HIGHCAM TRUECAM (EXES pixels)
P1-P2 (7.59 mm): 160.586 158.717 158.832 161.491
P1-P3 (15.21 mm): 320.243 320.330 318.124 321.236
P2-P3 (7.62 mm): 159.657 161.613 159.292 159.745
If we use the 4.14"/mm value from Ted Dunham, the plate scales are as follows
LOW MEDIUM HIGHCAM TRUECAM (in "/pix)
0.195534 0.197836 0.197693 0.194438
0.196726 0.196672 0.198036 0.196118
0.197298 0.194910 0.197750 0.197189
If we scale it up so that medium comes out as 0.201"/pix, as measured in lineops 2014, then the TAAS plate scales become:
LOW MEDIUM HIGHCAM TRUECAM (" /pix)
0.200031 0.202387 0.202240 0.198910
0.201250 0.201196 0.202591 0.200628
0.201836 0.199393 0.202298 0.201725
 |
| Platescale versus Input Angle for high-medium mode. Dashed line is with Gamma angle =5.43 deg, as per MJR's comment to this post. |
5 comments:
This fits with idea that platescale change caused by grating anamorphism. We will only see the effect when using the echelle for cross-dispersion so the slit is aligned perpendicular to the grooves. With the grating normal to camera (Ebert configuration), we get a larger slit image on the detector and so a smaller platescale.
With alpha = input angle, beta = output angle = (alpha + delta), and delta = angle between them = 2.715 degrees, the platescale should be reduced to 0.201 * cos( alpha+delta ) / cos (alpha). I've put a table in below. Without checking the numbers, my memory is that the effect was a little larger. I wonder if we should revisit the delta = 2.715 degrees value.
Input Angle PS factor Platescale
50.0000 0.942427 0.189428
53.0000 0.936018 0.188140
56.0000 0.928652 0.186659
59.0000 0.920044 0.184929
62.0000 0.909791 0.182868
65.0000 0.897297 0.180357
68.0000 0.881638 0.177209
71.0000 0.861311 0.173124
I went back some available NOD_ON_SLIT data and more carefully measured the shifts, and adopted the NOD_PQ values from the headers. (I included the point at 67.3 deg, 0.160"/pix which was my recent measurement on TAAS at 2142cm-1).
Typically, the stddev of the separations was 2% or less, measured over all available echelle orders (between 5-13 orders).
angle=[43.44,49.4420,50.5594,51.6000,53.9232,56.0421,58.2540,67.2698]
platescale=[0.1799,0.176,0.1798,0.174341,0.174615,0.171,0.168,0.160]
Measurements are the triangles. The predictions are the lines, with the dashed line being shifted 0.012 down to lie closer to the measured points.
Here are the data sets I used (out of order at this moment)
callisto 396cm-1, 2nd
asco 435cm-1, 2nd
callisto 384cm-1, 2nd
pallas 546cm-1, 3rd
pallas 747cm-1, 4th
gdra 1423cm-1, 8th
fcls 2142cm-1, 13th
vesta 584.8cm-1, 3rd
I looked at Curtis's post and decided to see what angle matches his measurements. I can't post the plot here, but Curtis could....
It is in EE/optics/platescalechanges_aug2016.png, if anyone wants to look.
Note, because the TAAS measurement is likely not at our best focus, it should have a different platescale. It was also hard to hit it along with the other points. So, I put his Ames TAAS measurement (67.3 deg and 0.160"/pix) in purple on the plot.
The prediction using beta-alpha = 2.715, the value from exesechelle.xlsx, is the solid line clearly well above the points. If I double that value to beta-alpha = 5.43 degrees, I get the dashed line that fits most of the points other than the purple one. Makes me think our beta-alpha may be off by a factor of 2. We should see if that helps with our grating positioning somehow. Of course, there is likely to be a difference in high-med vs med mode.
Just noticed Curtis gave me a correction factor for the TAAS points since on SOFIA our platescale in medium and camera was 0.201" / pix and on TAAS it was 0.211" / pixel. Adjusted the TAAS point (purple) accordingly and it is now plotted in black and quite close to the line for beta-alpha = 5.43 degrees.
Adding to the data:
1548cm-1 alpha Boo
6" || nod
0.165861 " / pix
Input angle=61.7450 deg
Post a Comment