Slit guiding and related issues

CCD cameras (CMOS/CCD), DSLRs, telescope mounts, focusers, autoguiders, domes, Roll off roofs, etc..
Post Reply
keving
Posts: 53
Joined: Thu Dec 14, 2017 4:49 am

Tue Jul 23, 2019 8:08 am

Does anyone know how the slit-guiding algorithm works?

I am trying to find the best method for guiding in spectroscopy with a
slit (not fibre). I have used slit-guiding reasonably successfully
with bright stars (mag 3-6) and a high resolution instrument (Lhires
III), but have now returned to using my low res instrument (Alpy
600). Here the challenge is to guide on fainter stars (up to mag 13 or
so). While I can see a star reasonably well in the guide camera at
these magnitudes(by messing with the display histogram), the
slit-guiding seems to get lost. It thinks its on target, but with the
star completely outside the target point. Just to be sure I understand
the display - the calculated star position is given at the centre of
the 'arrowheads' style cross, in black and red, right? This is at the
right place, but the star is displaced.

After several experiments, my best guess as to how the slit-guiding
algorithm works is that it uses the luminance across the whole (albeit
small) window selected for guiding. It then partitions this into two
'lobes' either side of the slit and tries to make this symmetrical
across the slit. This will work with bright stars where the star ADU
count is significantly above background, but fainter stars will fail,
especially if the skyglow is significant (as it is in the UK in
summer...). Then, small luminance gradients take over and drown any
stellar signal. Increasing guiding exposure can help, but there is a
limit to this, and it can be difficult to get an idea whether the
algorithm will work if you are used to stretching the guide image for
clarity....

If this is *not* how it works, I would be very keen to find the secret
behind it!

As a workaround, I tried two alternatives. This may be useful to
others, but I would also appreciate input here, as I have a few questions.

(i) Use 'fixed point' guiding. This is (I guess) designed for guiding
with a fibre. If used 'straight out the box' with a slit, it suffers
from the problem that it chases an asymmetric (dual-lobed) star image
back and forth across the slit: this is why slit guiding is
potentially useful. However, I found that, by winding down the RA
aggressiveness to 0.2, this damps these oscillations considerably
(assuming you have a slit oriented along the DEC axis in the guide
cam.) I was able to guide with an error of around 1 arc sec (and I
suspect the errors are over reported because of the asymmetry?)

(ii) Use 'all field' guiding. This worked well. Is there any reason to
use it as a default in all circumstances? One thing I didn't understand
was, what did the cross/cursor indicate on the brightest star? How
does this algorithm work?

Note that the disadvantage of using 'regular' guiding is that the only
image displayed is the region of interest around the guide star. If
this isn't the target star (on the slit) then you are trusting that
the guiding is maintaining the slit-star relationship (yes, I know it
should be, but its nice to have visual confirmation :). One request
for future versions therefore is to have the option to see teh entire
guide camera image (like PhD).

Kevin
Post Reply