To: Garry Nitta
Subject: Aligning the MCCD slit and camera

Here is the way the alignment of the MCCD should be done.
I will also describe tricks that have been used in the past
to make the data simpler.

The goals are:
	1) Make the wavelength direction in the spectrum
		parallel to the CCD rows.
	2) Make the slit parallel to the grating rulings.
	This give the highest spectral resolution.

	3) Make the slit perpendicular to the scan direction.
	This gives geometrically rectilinear spectroheliograms.

The optical properties of gratings cause the spectral
lines to be tilted on the CCD even when (1) and (2) are
achieved.  Since we never move the scan mirror, we cannot
achieve both (2) and (3).

In the past we have often adopted another goal:
	4) Make the spectral lines parallel to the CCD columns.
	This minimizes the number of columns needed to cover
	the desired spectral range and makes it easy to 
	take spectroheliograms and to establish the wavelength
	scale in the data.

This goal is in conflict with the others.  
If we try to achieve (1) and (4) by rotating the slit, we violate
(2) and (3).  If we try to achieve (3) and (4) by rotating the
CCD, we violate (1) and (2).

The reason we have gotten away with it is that the normal H-alpha
program used the #1 (low dispersion) grating and bins the CCD
pixels in space and wavelength.  The spectrum line tilt is proportional
to the dispersion, so it is small for the #1 grating.  The pixel
binning means that even if we make the columns parallel to the 
spectral lines, the amount of offset in spatial location from 
short to long wavelengths is small, less than a pixel.

Reni also uses (4), to minimize the number of columns 
needed to cover the spectral range and the time needed to 
take the scan.  Her science is not limited by the spectral
smearing (2) or the warped images (3).  So she can use (1) and
(4) in a consistent way.



Method:

I have methods for all 4 goals.  You just have to pick the
ones you need.

1) Wavelength parallel to CCD rows.

	1) Place a fine hair across the slit to block the light
	from a small place on the Sun.  

	2) Put the CCD in 1 x 1 binning, full frame image.

	3) Take a single image.

	4) Use PCOL (PROW?) to plot the intensity parallel
	to the slit at the longest and shortest wavelengths.

	5) Rotate the CCD until the location of the dark
	caused by the hair is the same at both ends of the spectrum.


2) Slit parallel to grating rulings.

	1) Use high dispersion grating.

	2) Make the slit narrow.

	3) Put the CCD in 1x1 binning, full frame.

	4) Rotate the grating to 0 degrees to see the zero order
	image of the slit.  This is very bright, so a neutral
	density filter may be needed.

	5) Rotate the slit

3) Slit perpendicular to scan.

	1) Put a crosshair at a focus upstream of the scanner.

	2) 


4) Spectral lines parallel to CCD columns.

	1) Narrow the slit.

	2) Use a narrow spectrum line.

	3) Put the CCD in 1 x 1 binning, full frame

	4) Take a single image.	

	5) Use PROW (PCOL?) to plot the spectrum at the top
	and bottom of the field.

	6) Rotate the slit until the location of the
	spectrum line is the same at both ends of the slit.
	{Compatible with (1), not with (2) or (3).}

		OR

	7) Rotate the CCD until the location of the
        spectrum line is the same at both ends of the slit.
	{Compatible with (3), not with (1) or (2).}