Hi Armin and Manuel.
Here's my LDSS3 notes, please let me know if you have any questions.

We use the VPH_ALL grating. 
We use this grating because it has the best combination of
wavelength coverage and throughput.
We use the CENTER SLIT POSITION (for the slit) because it has
the best balance for QE across the CCD.

Afternoon: Decide on subraster and measure slit position.

Early Evening: spectroscopic standard stars

Nighttime: science observations and Flats 


Catalog:
use the example from
http://www.lco.cl/telescopes-information/magellan/instruments-1/observing-catalogs-1/observing-catalogs


FYI... Clay phone number is (from USA): 626-310-1230

001 sn2006jd      08:02:07.43 +00:48:31.5 2000.0 0 0 -62.6 HRZ 0 0 0 0 0 0
002 sn2008bi      08:35:53.39 +00:42:33.1 2000.0 0 0 -62.6 HRZ 0 0 0 0 0 0
003 sn2008ae      09:56:03.20 +10:39:58.8 2000.0 0 0 -62.6 HRZ 0 0 0 0 0 0
004 sn2007ux      10:09:19.98 +14:59:32.8 2000.0 0 0 -62.6 HRZ 0 0 0 0 0 0
005 sn2008bt      10:50:16.88 -12:06:32.0 2000.0 0 0 -62.6 HRZ 0 0 0 0 0 0
006 sn2007sr      12:01:52.80 -18:58:21.7 2000.0 0 0 -62.6 HRZ 0 0 0 0 0 0
007 sn2008ar      12:24:37.92 +10:50:17.4 2000.0 0 0 -62.6 HRZ 0 0 0 0 0 0
008 sn2004gk_host 12:25:33.23 +12:15:40.1 2000.0 0 0  88.3 OFF 0 0 0 0 0 0
009 sn2007rw      12:38:03.64 -02:15:40.1 2000.0 0 0 -62.6 HRZ 0 0 0 0 0 0
010 sn2008aq      12:50:30.42 -10:52:01.4 2000.0 0 0 -62.6 HRZ 0 0 0 0 0 0
011 sn2006ck_host 13:09:40.42 -01:02:57.2 2000.0 0 0  11.0 OFF 0 0 0 0 0 0
012 sn2008cd      13:15:01.75 -15:57:06.8 2000.0 0 0 -62.6 HRZ 0 0 0 0 0 0
013 sn2007gx_host 17:18:51.49 +22:47:16.6 2000.0 0 0 -62.6 HRZ 0 0 0 0 0 0
014 sn2008bk      23:57:50.42 -32:33:21.5 2000.0 0 0 -62.6 HRZ 0 0 0 0 0 0

finders:
http://cfa-www.harvard.edu/~pchallis/armin.html


notes:
most interesting objects:
sn2006jd IIB 3x300 seconds R=~18.5
Currently SN is smashing into H envelope. Lots of emission lines.
(3x300 because we don't want to saturate the H alpha line, but want to
get flux on the faint emission lines...)

sn2007rw IIb, another interacting supernova.
3x600 seconds  R=~19.0

sn2004gk_host,sn2006ck_host,sn2007gx_host 3x900 seconds
take acq image of 30 seconds, then center on galaxy. Catalog
has the correct angle.
exposures 3x900 seconds for each host.

sn2008bk one of our swift SNe. R=????
(might be difficult to figure out exactly which object...)
might be available during morning twilight.
                                               
secondary objects (in order of interest !!!):
sn2007sr Ia-normal    R=~18.0  1x900 sec, lots of data, now ~5 months
sn2008ae Ia-Pec       R=~19.0  1x900 sec, lots of data, now +2 months
sn2008bi Ia-91bg      R=~18.0  1x900 sec, some data, now + 1 month
sn2008ar Ia-normal    R=~18.0  1x900 sec, lots of data, now +2 months
sn2007ux Ia-normal    R=~19.0  1x900 sec some data, now ~4 months
sn2008bt Ia-91bg      R=~17.5  1x600 sec not much data, now +1 month
sn2008aq II/IIb       R=~18.0  1x600 sec, some data, now ~2 months

standard stars:
http://www.eso.org/sci/observing/tools/standards/spectra/
(use catalog from operator, because it has the proper motion !!)

evening twlight standards
LTT 3218 RA 08 DEC -32 V 11.8 exposure time 20 seconds blueish
LTT 3864 RA 10 DEC -35 V=12.2 exposure time 20 seconds RED STAR

morning:
EG274 BLUE STAR R 16 23 D -39, V=11.0 exposure time 10 seconds
LTT 7379 RED STAR  RA 18 DEC -44 V=10.0 exposure time 5 seconds


 USE the hamuy star catalog.. because it has the correct proper motion !
 Parallactic angle !!!!! remember to ask the operator to make sure the
 rotator is set to the  Parallactic angle.
 (I didn't do this once and its really makes a huge difference in the
 flux calibration !!!


Slit: 1.00" ( we don't use the narrowist slit even on the best nights,
the telescope doesn't point perfectly so it takes more time to
make sure the obejct is in the narrowist slit. The 1" list doesn't
add that much extra light (especially during dark time).)
Don't use more than one slit, it takes too much extra time for calibarions.


 Subraster: 1 2064 1500 2800 (This is close to getting the entire
 4 sections of the long slit) no need to ever change this. You should see
 the yellow boxes on the quick tool match slits.
This subraster includes everything in the X direction (wavelength) and part
 of the Y direction (spatial).

 binning: 1x1

 readout: slow for everything.


 Flats: Lamp Quartz high 5 seconds.
 we don't take flats in the afternoon. Because of the fringing we use
 the flats taken about the same time as the observations.
 after each science exposure. take 3 flats (NO HENEAR).
 its pretty quick. A total of 3 per target is fine.

 Henear: the LDSS3 wavelengths don't change over the course of the night,
 so just take one or two at the beginning of the night and maybe one at the 
 last target exposure time: 2 seconds.

 To be done in the afternoon: measure params for ltoslit and lcobject

 make an image of the slit
 in GUI: 1 second exposure. 1.00 center slit no grating. no filter.
 (I think the scattered light in the dome is enough...)

 the IRAF task "lstitch" combined the 2 CCD images into 1 fits image.

 you can use iraf imexam  "j" I think to get the centering of the slit.
         jimexam    Parameters for line 1D gaussian fit plots
         kimexam    Parameters for column 1D gaussian fit plots

For 1 object at a time, We use a Y spatial position of 2100.
This is the best place on the CCD for 1 object.
It in the center of one of the slits.

Y=2100
now enter the x,y slit position into "ltolist"
and "lcobject"


for any object:
slew to catalog object
ltoslit
wait for guiding and SH
take acquisition image
take acq in r' (exptime depends on bightness)

then
 run lstitch
 run lcobject

 "Do you want to SKIP the measurement of the slit ?"
 answer: YES
 (because you already did this in the afternoon, and lcobject has
 the correct parameters)

 "Did you just acquire the object ?"
 answer: NO, because you are using ltoslit. Then the task lcobject will show you
 an image in ds9 centered on your x,y lcobject parameters.
 i.e. the Supernova should be near the center of the ds9 image.
 Adjust gray levels as needed.  select SN with cursor or centroid (m).
 Double check the pick circle is centered around the SN.

 If the slew if > 10" then you'll need another acquisition image and repeat lcobject.
 This doesn't really hurt, its just an extra minute and it is
 important that the target is really in the slit !!!!

 If the slew is < 10", then  (The telescope usually does a good job of correctly
   slewing and putting the object into the slit when the slew is less than 10 arcseconds.
  Also, we're using a wide slit 1.00"... if we were using 0.75" then we would be more
  picky about the size of the final slew into the slit.)

 Then set in the GUI:
 put in grating - VPH_ALL, filter - CLEAR, aperture - 1.00 center slit, readout slow

take science exposure 900 seconds.
then put in flatfield screen
take 3 flats 5 seconds 
and 1 henear arc 2 seconds

summary of observing sequence:

slew to catalog object
ltoslit
wait for guiding and SH
take acquisition image
   r' 1 second (or more for sne)
lstitch
lcobject

if slew > 10", repeat acquisition.
if slew < 10", then proceed with science exposures.
 
xx minute science exposure
xx minute science exposure
xx minute science exposure
3 flats 5 second each