Site engineer guide to setting up etalons and demonstrating that charge shuffling is working.

How do I balance the etalons?

Go out to Cass cage and put etalon into light table with pupil 8 in light table as well. This is the first thing that meets you when you enter the Cass cage. Screw on the Ne lamp after opening top and bottom windows.

At the s/w level, there are 6 etalon ports, with 3 and 6 clear. Put your etalon in 1, 2, 4, or 5, and then drive this etalon to the access port (light table) using the SMS control window.

i) power up s/w control and select etalon then external=off
ii) switch to R balance, balance needles using R control
iii) switch to error signal, time const highest 1.6ms, gain lowest at 1
iv) balance all needles with the coarse/fine control. Since the etalon is not servo'd, the needle will flick through zero. Just get the needles near zero using the controls. Some etalons are not able to be balanced but will still servo at the appropriate positions.
v) change time const 250ms and gain of 32 (loop is open here)
vi) close loop and integrate
vii) if loop overloads, reduce gain setting
viii) optimize by eye
ix) external = on, and away we go.
It would be useful to add a figure here of what one sees. Show Y axis as normal x-axis, and vice versa.

What causes large etalon oscillations?

(i) incorrect gain or time constant
(ii) large unbalanced R offsets
(iii) open or short-circuited etalon connector
What do we do if the etalon will not servo?
(i) can you balance the etalon? more to come.

How do we test that the shuffling is working o.k.?

For TTF, turn on CuAr or Ne lamp depending on filter; for other etalons, turn on D lamp. These are chimney lamps and do not the secondary mirror.
set pupil = 8
set aperture = 3 (thin slit)
put an etalon in the beam
You will need a blocking filter in the focal plane whose width will depend on the etalon gap. If TTF, use one of the TTF filters, say 670 (Ne lamp) or 710 (CuAr lamp). The UNC 19um requires a 45A blocker, say the H-alpha filter. The 125um and 200um etalons require something narrower than 10A, preferably at H-alpha. If not available, just recognize that there will be order confusion in the output spectrum creating more lines than you would expect.
We will now initiate an 80-step shuffle where the TTF is stepped through 80 unique plate spacings, and the slit image is shifted down the CCD 80 times.
xon aatxt?
rvtaurus ccd1

startup-                                (? on any line gives help)
   ccd_shuffle                           (looks up instrument sequencer)
              csr_default     disk$user:[observer.cs_files]*.csx
              filename        eighty
              preexpose_info  3_1000_100      (2 = 0.1ms; 3 = 1ms clock rate)
                                              (1000 = min phase time for bias frames clock cycles)
                                              (100 = delay time for external device. Leave these!)
              cycle_count     <<1>>           (no. of cycles where exposure time in .csx file)
              bias            <<false>>
              setup_array     <<true>>        (option 2: generate gap numbers; see below)
              etalon_start    <<0000>>        (start gap in option 2)
              etalon_incremt  <<0010>>        (increment in gap in option 2; no. of steps in .csx file)
In the CCD window, you can always return to normal mode to do a glance, say, with "method default". Probably best to do this after every shuffle since bombing in shuffle mode is not good. "Method shuffle" gets you back after "method default".


Figure 1: (A) The shuffle image in Fig. 1 comprises 80 separate images of the slit where each image is taken at different TTF plates spacings, i.e. wavelengths or frequencies (f1, f2....fN). (B) The raster image illustrated in A produces a long-slit spectrum where the degree of curvature in the lines is directly proportional to the gap spacing (the TTF has little curvature at the smallest spacings; the 200um has huge curvature).

File translated from TEX by TTH, version 0.9.