Echelon parameters

I've been looking at the echelon parameters so we can predict which echelon order lines will end up in.  The best data I have are with the filter open and the gas cell (see below) so we know what wavenumbers fall in echelon orders that are divisible by simple integers.  The order in the center has two clear lines in it.  I know from previous observations with appropriate filters in place that the line closest to the center of the detector is 600.103 cm-1 seen in echelle order = 3 and the line in the same order near the bottom is 400.2217 cm-1 seen in echelle order = 2.  There is also a line near the top near column 400 that I know is 400.48451 cm-1. The central order is much stronger than all the other orders and also much cleaner at the ends of the slit (horizontal).  This makes it apparent that the echelon orders overlapping in the central order are multiples.  To the right and left of the central, clean order are various less intense, less contrasting orders.  Eventually, we see a clean order on both sides.  Again, this is an order where different echelon orders overlap exactly.  Mid-way between the clean orders are somewhat intense, somewhat clean orders.  These locations are echelon orders that are multiples of 2, but not of 3.  Since the echelon order separation is inversely proportional to the echelle order number, we can figure out the location of echelon orders divisible by 4 as well.

Starting with the TEXES echelon parameters for the echelon groove spacing (7589 microns), R number (tan of incidence angle: 9.8), and gamma angle (0.025 radians), I varied the parameters to have the 600.103 line near the center of an echelon order (near M1.00, the 400.2217 cm-1 in the lower half of an echelon order (>M2.00,<M2.40), and the 400.48451 cm-1 line in the upper half of an order (>M3.50,<M3.75).  Further more, M1 = 3/2. M2 and M3 = M2 + 1.

Results will be better when we do this at other wavelengths (13.7 microns? 1315 cm-1?), but currently have:

echelon groove spacing: d = 7589.0 microns
echelon R number (tan of incidence angle): R# = tan(theta) = 9.8
echelon gamma angle: gamma = 0.20


Working on IDL routine to simulate gas cell spectra using HITRAN files with these parameters (and echelle parameters).

Changes to the detector box for the tilted detector experiment

We are considering tilting the detector by ~9 degrees in the echelle dispersion direction for the next cool-down, CD014.  This would allow us to move a spectral line along different focus values (+/- 2 mm) and find the optimal detector position for EXES. In addition to a tilt, we also wish to move the center of the detector forward (toward the camera) by 0.125", since we believe that to be equivalent to the 5 turn difference in the echelle collimator mirror position between best focus in med and high res modes that we uncovered in CD013.

This drawing shows the current configuration.
To do the tilt operation, we will add spacers to the posts marked "A", and we will need to make the standoff spacers at "B" and "C"to be 0.00" and 0.73", respectively. The latter results in pulling the detector center back by 0.24"-- so the length of the spacers at "A" needs to be 0.24+0.125=0.365". 

These requirements create a few spots for mechanical interference in the current D-box.

The chimney baffle will have to be widened from 2.46" to > 2.56", and rear plate of the D-box will have to have a alcove toward one side to accommodate the tilted heat-sink block.

Ryan from Emmett's shop needs about a few days on the CNC to remake the plate with the chimney - and he said that they could get to it before May 1.

In addition we will need 4 of the detector standoff spacers at "A" of ~0.365", 2 of the 0.73" head sink block standoff spacers at "C", and a new rear plate with an alcove built into it. I didn't get a chance to ask about the lead time for these parts.

(As for their recent delivery record: last Friday and early this week they have managed to get us the new cassette cover plate and add screw holes on the window holder to hold the window cover).

Tomorrow, I am meeting with Ryan to edit the solid works drawing to double check all the above measurements.


In the mean time, we should discuss this. If this effects when the next cool-down begins, at what point will it better to simply try +0.125 to the detector position with no tilt, rather than waiting for new d-box parts?

System Power Draw

Did a rough measurement this morning of the current drawn by the EXES racks as various items were turned on.

Bottom line:  Lots of headroom on the 2kVA UPS supply.

System draws about 6.5A total (absent the cal source supplies), which would be ~0.78kVA if we just multiply by 120V.

Here's a rough listing:

PC Host computer:  2A (momentary 4A on startup)
eBox supply:  2A
macmini + monitor:  1A
Lakeshore:  0.5A

Pot power supply + agilent + motor boxes + network switch < 0.5A total

The resolution on the current meter was only 0.5A, so these numbers are not precise below that level.  Hard to see where we'll have a problem that would require us to use any of the non-UPS power.

Slit feature resolutions at 14um in low mode

Echelle Voltage: 0.9062 V
files are from B_12_02_27

#filenum slitfeature   1             2              3
b061 1.5km/s          6.50        6.25         7.60
b060 2.0km/s          6.72        6.72         7.42
b059 2.7km/s          7.40        8.80         9.08
b058 3.5km/s          8.98        9.12         11.90

The scale is: R=36400/N_pix or 8.24 km/s per pixel

Resolutions:
Slitwidth     R
1.5     5600
2.0     5500
2.7     4130
3.5     3990

Slit feature resolutions at 7.5um in medium mode

The picture above shows the line IDs for the H20 band near 1315 cm-1 in medium resolution mode.

I calculate a velocity per pixel of:  2.53 km/s/pix

I was forced to only use frames 1 and 2 of the SUTR data to avoid saturation.

And below is a table of the FWHM of the lines, which were calculated by fitting 5-parameter Gaussian functions across the 20 central-most rows of each line. The values below are the median(sigma*2.355) for the 20 fits across each line.
The line pair at 1313.5 and 1313.6 were partly blended, even with 1.5 km/s, so I didn't record their widths. Where other entries are missing, it is either due to blending, or to saturation.

#slit            file    1        2         3      6      7       8
1.5 km/s        b032    7.77     6.81    6.88    6.36    6.10    6.22
2.0 km/s        b031    8.68     7.58    7.80    7.30            7.32
2.7 km/s        b030    9.70     8.85    8.97                    8.78
3.5 km/s        b029   11.75    11.26   11.59                   10.81               

4.5 km/s        b038            13.89                           13.54          
6.0 km/s        b037            18.30                                                 

flux slit       b036       

 For the mean values of FWHM, the resolutions are: 
#slit            mean(FWHM)    v_res(km/s)    R
1.5 km/s        6.69            16.9        17700
2.0 km/s        7.74            19.6        15300
2.7 km/s        9.08            23.0        13100
3.5 km/s        11.35           28.7        10400
4.5 km/s        13.72           34.7         8600
6.0 km/s        18.30           46.3         6500

Also, it seems clear that the right side is in better focus than the left. This was not clear to see in the high res modes. See the plot below.

The loose cassette cover plate has a very small effect on beam position

After CD013, we found that the cassette cover plate was loose - all the screws along the perimeter had a gap of about 1/16" (see photo above). We haven't seen this in any other component of the system, so we suppose these screws were untightened during the button-up before CD013. How big an effect does this have on the beam position at the detector?

Using the red HeNe laser, I acquired the laser beam in medium resolution mode (see image below).
I took a few images with the plate allowed to hang freely on the screws. Then I pushed on the far side of the plate so that there was only a gap on the near side. Then I pushed in the entire plate, so that there was no gap along the entire perimeter.

All files are in B_12_04_02/
image number            comment
b008-009            plate allowed to "hang"
b010-011            pushed in on far side (as seen in blog photo of cover plate)
b012-014            plate pushed in on all sides

The deflections were very small:
with respect to the beam position when the plate was allowed to hang:
when the plate was pushed on far side, beam moved ~4 pix in x
when the plate was pushed on all sides, beam moved ~ -6 pix in y