Re: Quick not on Disk 18a

[Tom Droege <tdroege@veriomail.com>]

Michael and all,

Ah!  I was wondering why you were concerned.  The ascii art explains 
it.  Why not us a running average on the data instead of binning it?  This 
preserves the shape while smoothing the peaks.  There are more complicated 
functions you can use.  For example, because we know the nature of the 
beaste, we might try to undo it.  Suppose we take the even binary 16s and 
spread some fraction of their data over the previous 15 bins.  By adjusting 
the fraction one should be able to undo what the bit error did.  But note 
that it can go both ways.  Or you can just "flat field" it.  Take a lot of 
data and calibrate the ADC.  That is (after all) what they do these days in 
building ADCs.  They (I think, it is hard to keep up) just take data and 
adjust the bin with to make it even.  They even sold an ADC that did just 
this.  You would turn it on and it would be about good to 12 bits.  After a 
few thousand conversions, it would learn to take good data.  I think an 
example is the CS016.  I considered using this device for tass - I recall I 
actually used it for the Mark I and Mark II.

I would expect that the binning errors that you see would be relatively 
constant over time.  Thus "flat fielding" should work.

Tom Droege

At 03:50 PM 5/20/01 -0400, you wrote:

>   I mentioned that raw images from the I-band camera on Disk 18a
>contained pixel values with peaks every 8 and 64 counts.
>
>   Tom replied that this is entirely within spec for the ADC chips,
>and showed that it shouldn't lead to big errors in photometry if
>one considers carefully its effect on the pixels on which light from
>a single star falls.  True.