View source for Focusing the telescope

== General focusing principles ==

Because of the Ritchey-Chretien design of the telescope, there is not a lot of room to change the position of the focal plane.  Any instrument needs to have its sensor placed very close to the nominal focal position of the telescope.  The RCOS focuser has readouts of 0 to 32000, in units of 1/16,000 of an inch ('''check this'''); in other words, the nominal focal position only moves by about two inches when running through the full range of possible focus.  

The most important and useful thing to remember when finding the focus for a new instrument is that the focal plane is 11.78 inches from the back plate of the telescope.  As a first step, eyeball where the focus of the instrument or eyepiece will be (e.g. the CCD sensor for an imaging camera, or the slit for a spectrograph) and insert or remove the necessary number of spacers to bring the instrument close to the telescope focal plane position.  The remaining adjustments can then be made with the RCOS focuser.


== Instrument-specific focus positions ==

Approximate spacer configurations and focal positions for current instruments ('''still to be confirmed'''):

* LHIRES III spectrograph:
** Use all spacers except the thinnest one (approx. 1 inch thick), and set focuser to about 17,600.
* SBIG ST-10 camera with CFW-8 filter wheel:
** Use all spacers, and set focuser to about 15,600.
* 2-inch eyepiece with diagonal:
** Remove three spacers (we need to mark these) and set focuser to about 18,000.

== To-do items ==

I've put a few possibilities here of ways to improve/simplify our focusing.  They're roughly in order of which I think we should try first.

*  OK, so this looks useful: [http://users.bsdwebsolutions.com/~larryweber/ Focus Max software] (and an explanation of [http://users.bsdwebsolutions.com/%7Elarryweber/ITSPaper.htm how it works]).  It apparently uses MaximDL to take images (like PEMpro does), though it is stand-alone software.  I can't tell for sure if it uses MaximDL to connect to the focuser, or does that itself.  There's a [http://www.tpo.ca/Using_FocusMax.htm tutorial] on how to get it calibrated, a one-time thing for each hardware setup.   So we would need to set up hardware profiles for both imaging and spectroscopy, as we would need to start the focuser in a different place for each.

* We could try out the interface in the Nebulosity software for focusing; Josh and Eric tried this a bit on 7/16/09.   This is a nice graphical view of a given star, showing a historical plot of peak values and star half-widths as you change focus, so you can tell where the best focus is.   One thing to watch out for is that if both it and MaximDL try to connect to the camera, they can lock up.  So start this first, before connecting to MaximDL, then finish focus and disconnect.  After quitting Nebulosity, you can connect to the camera from MaximDL.  (This doesn't waste cooling time, since Nebulosity also turns the camera's coolers on.)  There's a [http://www.stark-labs.com/blog/files/NebBahtinov.php video] showing how to use the Nebulosity interface to focus.

* It would be better, though, if we can get the focusing interface in MaximDL to work better.   It doesn't have to be auto-focus; it just would be nice to see a simple plot of the star and its FWHM, for example. 

* If this doesn't work well, we might look into creating/buying a Bahtinov focusing mask.   We could create a template [http://astrojargon.net/MaskGenerator.aspx?AspxAutoDetectCookieSupport=1 here] and then find someone to laser-cut it for us.  Maybe Steve knows a place we could take something to be cut.  Or we could go simpler and just have Steve make us a Hartmann mask.


== Troubleshooting ==
The focuser readouts have some repeatability, but not 100%, and from time to time the focuser seems to lose track of where it is.  If this happens, send it to "HOME" from the RCOS TCC software, and that will re-zero the positional readout.