Data Quality

[Tom Droege <tdroege2@earthlink.net>]

I have been doing little else but looking at mass quantities of data.  I am 
beginning to have some general impressions.

1) Darks Don't Matter
2) Flats Don't Matter
3) The Sky Matters

Of course darks and flats matter.  But one dark or flat gets about the same 
final results as another for me.  This is because the sky noise dominates 
everything.  I sure would like for someone else to work on some data and 
come to some conclusion.  Agree with me or not.

OK, I know this not the conventional wisdom.  Lots of others have been 
doing this for a long time and find that darks and flats do matter.  So why 
am I so confused?  Well, I think there is not so much experience by people 
using wide field instruments in lousy locations.

Darks.  I have built a temperature control system into the Mark IVs.  It is 
not as good as I had hoped, but the darks do not change very much.  Michael 
also corrects them to a covered pixel.  Also there is a dual slope ADC 
system which helps to keep the bias level constant.

Flats.  Some people may be able to see dust donuts, but not me.  Well, I 
may have spotted a few, but they are far down into the noise.

The sky dominates everything.  Well, which lights my neighbors decide to 
shine on my dome at any one moment may also contribute.  I can do nothing 
about this.  So I am content to take the data I can take.  This will not be 
0.001 or even 0.01 mag data.  It will be at best 0.02 or 0.03 mag 
data.   Sigh!  It is too bad that I am not at a better location.   If I had 
to haul a Mark IV to a good location, I might be able to take 5% of the 
data I can take here.  So I plan to content myself with taking long period 
large amplitude data.  I can get that, and I think it will be useful.  To 
do that I seem to be able to get by with monthly darks and flats.  They 
just don't matter with the other problems.

So why not just keep the processed .fits files??  It seems to me that they 
differ from the raw data in a way that will not change significantly by 
anything that I can do later.  They are much more useful.  Rob at a dark 
location may have a completely different set of problems.  I am very 
interested to hear what mag he can detect with a 90 second exposure.  I bet 
he beats me by a mag.  Rob, possibly you can send me a CD of images and I 
can process them through Michael's pipeline for comparison?

When Michael S. gets the new data up on a data base, I can always grab the 
area around a problem star and make a finder chart to match with the 
processed fits image to get the final position.  Possibly Michael S. can 
give me a tool to help with this, but it is not vital.  So further 
improvement of the image WCS values is not needed.  It is in the nice to 
have category.  My preference would be for Michael R. to spend what time he 
has available studying how well my data matches the standard system.  Also 
now that I have a lot of stars measured over 500 days, how stable the 
transformations are.  So my largest interest is in what we can say about 
the quality of the data.

I write this at the risk of biasing possible users  against the final data 
set.  Sigh!  I could just keep quiet.  But I like to talk about 
problems.  It does not mean that I am not doing the best job I can.  It is 
just that I do not plan on wasting time doing things of no value.  So I am 
not taking flats every five minutes or even every 5 hours.  I will take 
them as often as is needed to get the best data I can at this location.  A 
Crow rifle, some express amo, and a finder scope would do the most good, 
but there are a lot of lights in Chicago.  Possibly more lights than I can 
measure stars.  I figure I will get 2-4 million.  There must be more lights 
than that in Chicago.

Tom Droege