Ellipsoid Decentering Tests

Dana has done quite a bit of work testing the foreoptics ellipsoid by itself to determine how good an image it can produce.

Here is the best image obtained so far:

Image of laser fiber focus to focus using the foreoptics ellipsoid alone

The detector here has 5.4µm pixels, so the FWHM across the widest part of the image (yellow cut line) is about 190µm...not sure how this would scale to the detector, maybe someone can comment.  This was done with an f/35-ish beam size incident on the ellipsoid.

To get this image we intentionally decentered the ellipsoid by 10mm in the dimension perpendicular to the plane formed by the centers of the foreoptics mirrors and the laser fiber.  Dana calls this the Y-axis.  With this intentional de-center, running the camera through best focus looks like this:

Decentered focus run (click to see animation)

And here is a similar focus run for the nominally-centered beam:

Focus run w/beam nominally centered on ellipsoid (click to see animation)

There are only three frames is the 'centered' .gif because we did not use as fine a resolution for focus checks, but the plate scale is the same for the two animations, and we get to a smaller spot size when we decenter.  Is this an indication of a clocking error?  Inquiring minds wish to know.

We can also look at the image as we translate the mirror, though it's harder to make a movie out of since the image moves around on the camera:

Image variation with 'Y-axis' translation of ellipsoid (click to see animation)

Translating in the perpendicular direction (in the plane of the back surface of the ellipsoid, if that makes any sense, we don't see much sensitivity to decenter:

Image variation with 'X-axis' translation of ellipsoid (click to see animation)

We'll need to cogitate over what this may mean, if anything.  The foreoptics is now at the position which gives the image at the top of this post.  Dana has talked about shifting the clocking and repeating the decentering tests, but it might be well to compare these images to Zemax predictions to guide the next moves.

Alignment Tests of January 14

After a fruitless first engineering night with essentially no astronomical data collected, we re-examined our optical alignment during the day of January 14.  First we went back to the gas cell to verify that we could still find the Q-branch:

Q-Branch collected 1/14 (unsaved screengrab)

Here the spectral lines look reasonably sharp, and we can see evidence of the re-steering we did during the night.  We're shifted about two orders to the right in the display, and he vignetting at the bottom of the image is gone.

Shifting to lo-med, we could still see the Q-branch:

Q-Branch collected 1/14 (unsaved screengrab)

Then, swinging the echelle out of the way, we see an image of the slit (decker wheel open):

Slit image collected 1/14 (unsaved screengrab)

This is through the 14µm filter, and we see diffraction plus a strong glint that is probably coming off the front edge of the echelle mount.

From here we went to the 5µm filter and opened the slit wheel:

5µm camera-mode image of gas cell collected 1/14 (unsaved screengrab)

Here we see that at least one of the wheel apertures was not lined up...turned out to be the slit wheel, but that was being positioned with reference to the data so not a big deal.  After moving the slit wheel by 4 degrees we got a well-aligned s/d/f/cassette:

5µm camera-mode image of gas cell collected 1/14 (unsaved screengrab)

We see the gas cell is a bit off-center, as it was in our prep-room test of January 11.  The instrument is hanging off the telescope here, and might have flexed a little(?)....looks slightly more off-center to me in this image, but the earlier one was at a different wavelength and it's hard to compare.

Next we re-installed the Z-beam and the laser mount, and put a soldering iron behind the laser mount:

5µm camera-mode image of soldering iron collected 1/14 (unsaved screengrab)

The image of the soldering iron is not very round, even though the aperture we're shining it through is pretty close to the IRTF focus.  The hole in the laser holder is cylindrical so maybe reflections coming through the holder would create an angular dependence.

Evaluation of EXES IRTF foreoptics

The EXES IRTF foreoptics were set up in the lab on the z-beam with the ellipsoid and hyperboloid translation stages as they were set at IRTF; however, those two mirror angles were tweaked to center the beam on the hyperboloid and direct the beam through the KBR lens.

The foreoptics were fed with a 635nm laser beam from a ~5µm diameter fiber core, apertured with a 10mm diameter iris set 350mm from the fiber tip.

The focused beam from the KBR lens was imaged onto a detector with 5.4micron pixels.

These are 'through focus' images of a HeNe beam after the foreoptics and the KBR lens. Images were taken every 5mm along the beam. At best focus the beam diameter is approximately 920 microns.

This is a theoretical 'through focus' spot diagram of the HeNe beam.
At focus the beam diameter is about 20 microns. In this model, the beam from the KBR lens is f/16.

January Optics setup @ IRTF

Once we had the instrument down to liquid helium temperature we set up the gas cell to optimize focus on the C2H2 Q-branch @13.7µm using the gas cell, but not the foreoptics.  Here is an 11 January image of the gas cell off the alignment mirror, showing it to be slightly off-center and somewhat out of focus as we'd expect:

SB375 image of gas cell (test.science.10092.fits)

In this image I've inverted the grayscale so that more flux shows as white, less flux shows as black.  The position of the image on the array is affected by the (randomly pointed) alignment mirror and isn't relevant to spectral work, but the relative centering of the gas cell on the s/d/f cassette aperture is.

From here we did a lo-med focus run, and settled on 21.5 T for lo-med best focus:

SB375 lo-med image of C2H2 Q-branch (test.science.10098.fits)

FWHM for the column plot shown is about 6 pixels, close to the diffraction-limited size.  If we zoom in to look at the individual images of the pinhole we see that they are approximately round...about as good as we've seen.

Zoomed image of C2H2 Q-branch (test.science.10098.fits)

At his point we went on to do a focus run on the echelon side.  We picked 25T for the echelon hyperboloid (narrow slit):

SB375 hi-med image of C2H2 Q-branch (test.science.10108.fits)

FWHM looks like about 20 pixels here and the signal looks smaller, though it would take some work to make a useful comparison.  The log book says 'narrow slit', but not 'narrowest', though the fits header indicate the 108µm slit for this file.  This is not as sharp an image as we have seen in the lab, where we got to ~12 pixels FWHM.  This might be equivalent to 13-14 pixels on the smaller pixels of the SB375, but probably not as many as 20.  Maybe pressure broadening??  The lab book isn't clear on the condition of the gas cell.  We settled on a best focus sum of 46.5T, echelon + echelle.

Next day (12 January), we ran along the 46.5T diagonal looking for optimum focus, imaging the Q-branch with the slit-wheel pinhole.  Here is an image at 23.5T echelle, 23T echelon:

SB375 hi-med image of C2H2 Q-branch, 23.5/23T (B_11_01_12/b001.fits)

And here is a somewhat sharper image, taken at 33.5/14T (sum is 47.5 turns, curiously):

SB375 hi-med image of C2H2 Q-branch, 33.5/14T (B_11_01_12/b005.fits)

In this last image, the spots look slightly de-focussed along the spatial dimension, but are pretty sharp spectrally.  If we look at the order marked with the yellow line, we see a spectral FWHM of about seven pixels (see inset plot).  The 23.5/23T image just above it looks more like about 15 pixels FWHM spectrally.  Surprisingly there isn't a definite note in the book where the mirrors were set for our first night on the telescope.

The image above was acquired at about noon on January 12 in the prep room.  Next notes in the log book refer to more detector linear-range testing in the prep room, but there are some files saved starting at 1:47am on January 13 with much sharper slit-gas cell images:

SB375 hi-med image of C2H2 Q-branch, (11JAN13/c2h2science10065.fits)

We get a spectral-dimension FWHM of about 6 pixels here along the yellow cut line (see inset).  This shows that the focus inside the instrument was pretty good at least spectrally when we went on the telescope on the morning of January 13 for our first engineering night.  We also see the '20-row' noise in this first-minus-fourth-scan difference image.

Pre-IRTF Optics Recap

In planning/executing tests to characterize our optical performance at the IRTF, I'm summarizing the steps taken in the runup to the observation using visible light with a bare mux as the imager...hopefully this will be useful for diagnostics in comparison with the IRTF data.

By November 5 we had the bare mux running in the instrument, and we did a focus run in lo-med mode (Vol. III, p. 5).  We selected 25T as the best focus for the echelle paraboloid.  Here is a comparison of image quality in med-lo mode, one using the SB375 bare mux and the other using the SB226:

 SB375 bare mux image obtained on November 5, 2010 with

our laser fiber @ cold stop and the echelle mirror set at 25 turns absolute.

Similar image obtained September 22 using the SB226

bare mux at an 

absolute echelle position of 23.5 turns.

The SB226 image looks sharper to my eye, but there are plenty of reasons these images should not be directly compared.  Probably most significant is the fact that we had done no bias/clock optimization on the S375, and we apparently used a much stronger laser in the SSB375 image, creating an image about four times as bright as that in the SB226 image. I think it's a little odd that the odd/even effect in the SB375 is really only evident in the brightest part of the inset column cut...not sure this is really significant given the guessed-at operating setup.

At some point (in November I think) we received the foreoptics from Corning.  We installed them on the instrument and performed an alignment.  Amazingly, there is no record of this process in our lab book that I can find. According to my e-mail log, we received the mirrors about 15 November 2010, and Matt came down to Ames on the 16th to test them.  The EXES computer has only one file that looks like it was taken with the foreoptics installed, and it's dated 18 November:

SB375 Bare mux image through foreoptics Nov. 18 (25T)

This image does not appear to be in great focus....the signal level is very high so keep blooming in mind as you compare the images.  The odd/even is pronounced in the coulumn-cut, moreso than in the first image above.  Note the diagonal rays that are not apparent in the first SB375 image above.


I'm frankly surprised that we don't have more in the can from the lab test of the foreoptics, but we were trying to get ready to ship and maybe trying to do too many things at once.


Meanwhile, here is an image taken in the IRTF prep room December 9 with the laser phisterous mounted at the cold stop (similar conditions to the first image in this post but the signal is bombing bright, with no apparent diagonals even at this flux:

SB375 bare mux image obtained on December 9, 2010 @IRTF with

our laser fiber @ cold stop and the echelle mirror set at 25 turns absolute.

Andhere is an image at LN2 in the IRTF prep room with laser light coming through the foreoptics:

SB375 bare mux image at 77K at the IRTF, lo-med mode laser fiber imaged through foreoptics

There are again diagonal (plus quasi-vertical) rays coming off the different mode-images of the laser fiber that were not evident in the pre-foreoptics images.  I don't remember ever seeing these rays without the foreoptics, but it's possible they have always been there at a low level and we happened to accentuate them with the settings we used for these particular tests.  More likely these rays are related to the foreoptics and/or their alignment??


We did some foreoptics tweaking of the foreoptics, with an example image shown here (much lower laser power so it's easier to see the spots formed by the laser modes:

SB375 bare mux image at 77K at the IRTF, lo-med mode laser

fiber at 0.05mW imaged through foreoptics

The spots are elongated in this image, though the diagonal rays don't show up (at least at this power level)