Auto-indexing with Denzo

Do NOT close the image window yet!

After having obtained the `peaks.file` by running in XDisplayF you need to edit the input file for DENZO. In the same terminal window from which you started the image display you grab a template with
`$ copy $p:[denzo.example]auto.dat *`
or just click here for one. Edit the file with
`$ edit /tpu auto.dat`
(or whatever your favorite editor is called).
It is VERY IMPORTANT that you change all the relevant parameters to reflect the conditions under which your image was taken. Crucial are the

Distance (crystal-detector) in mm
Wavelength (typically 1.08 A)
Format of the image (`mar 18cm` or `mar 30cm`)
Beam center in mm. Approximate values for centered (2theta=0) plates are
`xbeam 90.0 ybeam 90.0` for mar 18cm
`xbeam 150.0 ybeam 150.0` for mar 30cm

Template for the image file name (`raw data file`)
Sequence number for the image file name (`sector`)
Proper description of the rotation range in degrees
After you have adapted `auto.dat` you MUST start DENZO in the same directory that you started hkl1200/hkl2000 in:
`$ denzo`

followed by
`@auto.dat`
which will produce plenty of output. Towards the end you'll get a list of the 14 Bravais lattices and a value for the goodness of fit (or "metric tensor distortion index" as Z.O. calls it).

Make a note of the lowest reasonable value (less than 1%) in the highest symmetry lattice, in this case P222 (with a MTDI of 0.22%). If you already know your cell/spacegroup, this is the moment of truth (one of them at least, the next one is the R_symm)...
You can now set the spacegroup by
`spacegroup P222`
and continue refinement by a few
`go go go`
If things are well-behaved and the chi2 values have stabilized (preferably at values under 10), you can refine more parameters by
`fit cell go go go`
or a more daring
`fit all go go go`
For a good crystal with high resolution data (`res 1.7 10`) you should get chi2 values close to 1.0:

This is a good time to check the updated predictions in the graphics image, they should match the spots very well. Green spots are fully recorded reflections, yellow spots are partials and red spots are overlaps. If you find many red spots you may want to play with the `mosaicity`.

You should also use the ZOOM button to magnify an area to overlay the integration boxes on top of the spots:

The big square box around each spot is the integration `box`. Pixels between that box and the first "circle" inwards are used for background determination. The actual spot intensity only comes from the pixels inside the inner "circle", pixels between the two "circles" are not used.
You can change the size of the box with
`box 3.0 2.5`
The size and shape of the outer "circle" by
`background elliptical 0.7 0.7 0.0`
and size and shape of the inner "circle" by
`background elliptical 0.5 0.5 0.0`
A few `go go go should tell you whether things are improving.`
When you are happy with the refinement, type
`list`

and paste the "Current parameter values" into a file for subsequent inclusion in the DENZO integration input file.
Finally, to quit DENZO you type `quit` TWICE...