From: Greg Aldering (aldering@panisse.lbl.gov)
Date: Fri Nov 01 2002 - 16:52:31 PST
Hi Isobel,
>We should think again about what seeing we can tolerate, since that
>will increase the chance of us getting data. We had originally
>specified median seeing or better which is approx half arcsec. I
>suggest we relax that to 70%-ile which corresponds to about 0.8". I
>don't think there's much point allowing worse than that because Keck
>and Subaru will also be observing during much of the same period, so
>they will have to use those condistions anyway. However the exception
>to that is the 3rd Nov when no-one else is getting spectroscopy as far
>as I know. If the seeing is 'bad' that night, i.e. >0.8" we could get
>to the top of the queue earlier since the programs at the top require
>seeing < 0.8".
>So we could require <0.8 for all except 3rd (if we're allowed to do that).
>Any comments?
I think the main point here is that we should say that we will accept
seeing < 0.8 on Friday the 8th, when we are not on ESI or FOCAS. On
other nights, is it possible to taper our request to be most demanding
on the early nights and more accepting on the latter nights?
Alternatively, can we request the best seeing only for the faintest
targets and allow worse seeing for brighter targets?
>Second question is whether we want to use Nod & Shuffle (for improved sky
>subtraction) for the GMOS observations. It certainly improves the sky
>subtraction, but the downside is a lower observing efficiency, by around
>25%. Personally I would like to use that mode just to see what it can do
>on these faint red objects, but please let me know if anyone objects. We
>will need to put in a request to use it because it was not in our original
>proposal (and in fact wasn't formally offered for 2002B).
Nod and shuffle sounds like it is what we would like to use due to
GMOS's strong fringing, but I'm surprised that the efficiency loss is
only 25%. Since one is performing a pixel by pixel sky subtraction, how
can the loss in S/N be any less than sqrt(2), corresponding to any
efficiency loss of a factor of 2? I must be missing something.
>Does anyone have a strategy for which telescope should be observing which
>targets (bright vs faint?). I would quite like to try at least one really
>faint one with Nod & shuffle but am open to suggestions.
Right now we are quite uncertain about the current efficiency of ESI
(degraded due to oxidization spots on the spectrograph mirrors); we
also lack an efficiency curve for FOCAS. So it is hard to compare ESI
and FOCAS to GMOS and FORS2, and those are the comparisons we really
would like in order to decide priorities for each spectrograph. I would
roughly guess that FOCAS and GMOS will be comparable to each other,
that FORS2 will be better then either for higher redshift objects
because it has a MIT-LL CCD, and that ESI will be no better than FORS2
in terms of efficiency but will have a gain for the highest redshift
objects because OH lines can be suppressed. Even with nod and shuffle,
GMOS still faces the low red QE of its detector. Since our primary
emphasis is on three z > 1 SNe for HST, it may well be that all of the
spectrographs will have to try a very faint target, but that a mixed
portfolio including brighter SNe will make sense for the 8m's.
- Greg
This archive was generated by hypermail 2.1.4 : Fri Nov 01 2002 - 16:52:50 PST