[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: experiments



As far as I can see, there is no meaningful way to read the Mark IIIs 
out faster.  I say "meaningful" because you can read them out faster,
but I don't think any calibrations would hold.

Note that this is exactly what is done during the clearing process.  I 
don't know how fast this is, but I would guess an order of magnitude
faster than normal read.  In principal, one could read out one line
from one camera in about 20 ms, or the whole camera in 10 seconds.  

You may remember the original cameras came with BASIC code.  One could
modify this and probably read out a single camera in 5-10 seconds.  I
am not sure that I would trust such a scan as a flat, though, without
a lot of comparative runs.  Too much is different, the dark current for
one.

Tom Droege

On Fri, 22 Aug 1997, Glenn Gombert <gleng@infinet.com> wrote:
>
>        I have a week of vacation scheduled for Labor Day week (also new
>moon as well) , I will try and get as much done during that time as
>possible. I would prefer not to duplicate what others may be doing *there is
>much to do* so will try and cooridnate with what Mike G. and others are
>working on during that time frame.
>
>        One thing I have not figured out is how to read-out the Mark III
>camera for a full frame in say 3-10 seconds. Maybe Tom has tried this? I
>used that exposure range when using my ST-6 with John Chumack's 16 inch
>F/4.5 Newtonian and it worked out well in that exposure range when taking
>twilight flats. I can't read-out "dome flats" in the basement with eight
>minutes per readout. The mark III would saturate *way* before the readout
>was done, we some some other mechanism for reading out the Mark III rapidly
>in less than 10 seconds.
>
>
>Glenn G. 
>
>
>At 02:58 PM 8/22/97 -0700, you wrote:
>>Glenn G. offered to run some experiments to track down some of these
>>errors.  Here is a preliminary list of things I'd like to see tested:
>>
>>FOCUS
>>  1.  Chris suggested two experiments here.  Nick and the APL folks
>>      devised a focus scale for the lenses, and I think such scales could
>>      be put to good use.  Use one camera only (each lens will have a
>>      slightly different focal length, especially in switching from V
>>      to I filters) and adjust the VCO so that the central column has
>>      the least amount of RA drift.  Adjust the focus so that camera
>>      has its best focus.  Then move the focus ring a little each way
>>      and see if the best psf moves in declination.  This will check
>>      the lens alignment.  Then, rotate the lens in 90 degrees steps
>>      as allowed by the 4 screw mount, as suggested by Tom, to check
>>      the lens axis with respect to the glue joints.  At least at the
>>      180 degree points, one should see a change in the gradient of
>>      the psf.
>>  2.  The question was raised early on whether these lenses work well
>>      at all in the NIR, since the psf's for the I filter were so much
>>      worse than for V.  Tom now has a mark III with V,R,I filters.
>>      It would be useful to see the psf's for each filter using one
>>      of the Augat(?) lenses, to see if the best images go in the
>>      direction V better than R better than I.
>>
>>FLATFIELD
>>  1.  Use reflected light off a white card held in front of the triplet.
>>      The card should not be shiney, and should be at some distance from
>>      the lens (a foot?).  I'd consider either diffuse light near
>>      twilight or several flashlights spaced around the triplet for
>>      illumination.
>>  2.  Use a diffuser in front of the lens.  Two pieces of ground glass
>>      spaced by a half-inch or so should be adequate.  Again, some sort
>>      of diffuse illumination if possible.
>>  3.  Look at the flatfield vectors derived from a moonless sky, and one
>>      derived near the moon (good to do this time of month).  That will
>>      tell you the kind of errors to expect in sky-based flats.
>>
>>These are very obvious tests, and I'm sorry if I repeat something someone
>>has already performed.
>>Arne
>>
>>
>Glenn Gombert <gleng@infinet.com>
>
>
>
>