WIYN / Bench Upgrade: VPH 740 l/mm grating

Status: Component fabrication is complete. Design and drawings are courtesy of Bill Schoening. Configuration for low-density grating (small angle). Current grating turret (#1) holds flat for low-density VPH grating configurations. VPH grating is in second grating turret (#2). Fabrication costs, etc: If you look at the VPH mount in the bottom left panel of 4, there is a stand underneath it. The orange part is a rotating stage that is commercial (from Bill's office). The grey pieces above and below below it adapt to the grating and to the table. Those were fabricated. They are relatively simple pieces, but the costs were (???). The bottom set of 3 drawings show Grating cell detail for 740 l/mm grating.
"As built" Dimensions:   outer dimension of exterior housing Height = 406.4mm Width = 463.55mm Depth = 111.125 - 155.575mm depending on whether you include the clamping knobs All dimensions were measured to the nearest 1/8 inch and converted to mm; so the error is +/-3.175mm (an overestimate). These are "as built," as measured measured on the spectrograph by Gene McDougall (28 mar 2006). Also note: The turret that holds the VPH cell extends beyond the cell anywhere from 2-5 inches depending on where you are measuring. The excessive width of the exterior mount is to accomodate larger-width gratings to be used at larger angles. The negative ramifications of this choice is that larger fold-angles are required to place VPH grating close to fold-flat and not vignette incident beam. And finally: YES this will get annodized!

Preliminary result: comparison of VPH 740 l/mm vs SR 790 l/mm grating throughput. The VPH grating has 2 times (1.7-2.5) the throughput as the SR grating. Note that only the central third of the sampled wavelength range has been used in this analysis because of focus degradation away from the central wavelength. Likewise, this result applies to an average over the central third of the fibers. In June run, more time was available for focus, and significantly higher uniformity was achieved with the VPH grating. To Do: Check the extraction width as function of wavelength. Credit: K. Westfall, D. Harmer, C. Harmer.

Jan'05 T&E SETUPS for HYDRA-RED
order	alpha	VPH 740 l/mm				SR 790 l/mm
		starting wave (A)	ending wave (A)	central wave (A)	dispersion (A/pix)	starting wave (A)	ending wave (A)	central wave (A)	dispersion (A/pix)
1	15.0	6010	8259		1.08			~790
1	17.0	6907	9188		1.00			~790
1	19.0	7860	10044		1.06			~900
1	21.0	8688	10948		0.99			~960
2	19.0	3904	5038		0.54			~460
2	21.0	4332	5435		0.54			~490
2	23.0	4789	5872		0.53			~540
Table Note - we estimate spectral FWHM is ~2 pixels for Red Hydra and ~4.5 pixels for SparsePak in Littrow configuration for the VPH 740 l/mm grating. The FWHM will be somewhat smaller for the SR 790 l/mm grating because of additional anamorphic factors (i.e., non-Littrow).

Shared-risk run, Jun 2005. (Obs: D. Harmer, K. Westfall, C. Harmer) 740 l/mm only, but configured with large flat (215mm x 260mm clear aperture). Did not get on sky due to weather, but engineering data were gathered with both SparsePak and Red Hydra cables w/ and w/o the masked used in the Jan'05 T&E run to mimic the small-flat footprint. Setups are summarized in the table below; data included domeflat and calibration lamp spectra.

Jun'05 Shared-Risk SETUPS for VPH 740 l/mm Grating
order	alpha	cwl	cable
1	20o	~ nm	HYDRA-RED
1	21o	~ nm	HYDRA-RED
2	20o	~490 nm	HYDRA-RED
2	21o	~490 nm	HYDRA-RED
2	23o	~540 nm	HYDRA-RED
2	21o	~490 nm	SPSPK

With these two sets of data (Jan and Jun 2005) we will determine answers to the following questions:

• How much throughput gain is realized from large flat? Jun HYDRA data w/ mask and w/o mask: Dome-flat spectra. [1st priority]
• How different is throughput for SparsePak and Hydra cables? Jun HYDRA w/o mask and SPSPK w/o mask data for alpha = 21 deg (m=2) setup: Dome-flat spectra. After correcting for difference in fiber aperture, the remaining difference at the 21 degree setup (wavelength range bteween 470 and 510 nm) will be due to a combination of (a) fiber transmission (at the 21 degree setup the wavelength range is roughly 470 and 510 nm, which is roughly where low-OH, i.e., red, fiber transmission starts to role over); and (b) toe vignetting. The latter will be wavelength independent. [2nd priority]
• What is the spectral resolution vs wavelength for both gratings and both cables? Jan VPH data: spectral resolution vs wavelength for both gratings, all setups, and both cables. Line-lamp spectra. Initial measurement for one line near cwl for each of six setups and both cables. [3rd priority]

  Later, we will measure FWHM (sigma) for all good lines in the spectrum and for all fibers.

• How does spectral resolution change with beam size? Jun Hydra data: Compare line-lamp spectra for masked and un-masked data. [4th priority]
• Repeatability, lamp-stability, and mask fidelity: Jan HYDRA data vs Jun HYDRA data: Compare masked Jun data with large flat to Jan data with small flat. Dome-flat spectra. [5th priority]

Preliminary results:

• Throughput gain achieved from large fold-flat: A direct comparison was made of dome flats taken with the red Hydra cable in 6 setups with the VPH 740 l/mm grating with and without a mask simulating the aperture of the small fold-flat used in January 2005. All data were taken in close time proximity in June 2005. These are 1st and 2nd order grating setups in the range of 20-23 deg. The mean change in throughput is a factor of 1.64, i.e., a 64% increase in throughput on average across fibers and spectrum. The gain is dependent on field angle, i.e., both slit position and proximity to the central wavelength. The largest dependence is with slit position, varying from a gain of 1.2 for edge fibers and 2.2 for central fibers; the gain for central fibers is significantly larger. For a given fiber, the gain is down by 10-20% at the extreme wavelengths.

• Relative throughput of SparsePak and red Hydra cables: A direct comparison was made of dome flats taken with the red Hydra and SparsePak cables in the same grating setup (order 2, 21 deg). All data were taken in close time proximity in June 2005 using the large fold-flat, unmasked. In the spectral range between 440-540 nm SparsePak has a 47% higher throughput -- accounting for different fiber apertures, i.e., SparsePak passes a factor of 1.47 times more photons per unit fiber area than the Red Hydra cable. There is a small wavelength dependence of about 15% (1.42-1.57), with SparsePak becoming somewhat higher throughput at the red end of the sampled wavelength region.

  This result agrees qualitatively with a comparison of system absolute throughput with SparsePak and the blue Hydra cable, as reported in Bershady et al. 2005 (ApJS). The reasons for throughput differences include differences in fiber transmission, and differences in the fiber-exit apeture formed by the cable toes. Based on a comparison of the fiber transmission, SparsePak fibers are considerably redder in the 440-540 region than the red Hydra cables. In the Sep 03 Concept Design Report we estimated differences in the fiber toes should amount to roughly a 20% effect (see Section 7), compared to 47% here. However, the red Hydra cables appear to have the largest FRD (see Figure 18 in the Sep 03 Concept Design Report), which would lead to larger vignetting from the toes than estimated in the Report. We conclude that SparsePak continues to show evidence for significantly higher throughput than existing cables. Further tests should be made (any grating) where observational control is provided (e.g., repeat and bracketing observations) for changes in the dome-lamp intensity and color, and other cables are included.

The highest priority for performance-verification for this run is on-sky measurement of spectrophotometric standards with SparsePak to determine the absolute throughput of the Bench Spectrograph with the large flat and VPH 740 l/mm grating. Measurement in at least one setup is needed.