From rayner@galileo.ifa.hawaii.edu Thu Jun 15 10:32:01 2000
Date: Fri, 9 Jun 2000 11:41:45 -1000 (HST)
From: John Rayner <rayner@galileo.ifa.hawaii.edu>
To: bergknut@jeans.ifa.hawaii.edu, cushing@galileo.ifa.hawaii.edu,
    denault@jeans.ifa.hawaii.edu, gching@galileo.ifa.hawaii.edu,
    koenig@jeans.ifa.hawaii.edu, neill@galileo.ifa.hawaii.edu,
    onaka@jeans.ifa.hawaii.edu, rayner@galileo.ifa.hawaii.edu,
    rayner@galileo.ifa.hawaii.edu, toomey@jeans.ifa.hawaii.edu,
    vacca@galileo.ifa.hawaii.edu, vern@hokuhele.ifa.hawaii.edu,
    watanabe@jeans.ifa.hawaii.edu

Tony and Peter,

The XUI survived its first exposure to observers
extremely well. Both George Bendo, and Heidi Hammel
and Mark Marley, were running SpeX relatively easily
after one night each of support.

Below is a list of fixes and improvements we need 
to make. Some are functional, some are cosmetic
and some are somewhere in between.

The changes should be done before the next observing
run, June 23-26. The next obsever can be a miserable 
bastard not averse to phoning you at home at 4am and 
sending you sarcastic e-mail if necessary.

Thanks,

John

XUI changes (* higher priority):

*Include beamswitch deadtime so that guider can 
pull object into slit before spectrograph starts
to integrate. If the spectrograph starts integrating 
before the object is in the slit blurring will result,
which can cause problems for spatially resolved objects
such as Uranus and Neptune. This needs to be very
visible so put it in the itime/coadd/cycle window.

Slowcnts should also be taken out of the engineering
window and put into the itime/coadd/cycle window. Some
objects (bright stars and prism flats) require the 
shortest possible itimes (slowcnts 4 rather than the 
low noise default of 12) to avoid saturation.

*Both bigdog and guidedog need to have function buttons
setup to flush the arrays after observing bright objects
to minimise persistence in subsequent exposures. Peter
and I need to discuss what purges to do.

*Make startup macros (powerup1 and init_0.4V) one 
function button (INIT?).

*Remove unnecessary macros from list. There is an
a lot of clutter currently.

*When guiding only in beam A (guide AB box not ticked)
slowguide should automatically stop guiding in the
B beam. At present guiding continues in the B beam
and can drag the object out of the guidebox when it
returns to beam A. (It sometimes guides on hot pixels).

*The telescope offset and send AB beam offset buttons
need to be in the same panel as the guide A (Ga) and
guide B (Gb). It requies some dexterity to set up the
guide boxes at the moment and this change will make it
much easier by keeping everything needed in the same 
panel.

Remove position angle (pa) from slow guide box. Use the
telescope offset panel pa.

*Incorporate bad pixel mask in slowguide so that guider
avoids using data from bad pixels. The mask can
be embedded.

Using the STOP button to abort integrations seems to
hang the DSPs occassionally. (The symptom is that
subsequent GO commands are ignored.) I was able to recover
from this by running the powerup and init macros. 
Why does this happen and is there a better way to
recover?

*Get xui macros working, including SAVE and RESTORE
commands.

*Boxes drawn in dv are sometimes very difficult to see.
This is causing difficulties in setting up guide
boxes for the guider. Find some way to improve 
visibility of drawn boxes and lines.

*Headers: Check that DIVISOR is being recorded properly.
IDL issues illegal header content message according to
Hammel. 

Both sets of observers complained about the difficulty
seeign green lettering on a grey background (eg. cal
lamps ON indicator). Change colour?

Make autosave button same size as GO and STOP buttons
and bright yellow.

Itime bar should indicate minutes remaining rather
than percent.

Is it possible to issue audible beeps at the completion
of integrations, cycles etc? (I'm sure we've been
here before.)

Setup guest accounts