Re: Measuring noise.

[Tom Droege <droege@wwa.com>]

Chris and all,

Uhhhhh!  At the moment I am clicking a power supply on and off using a
stopwatch.  I figure about 1% if I am paying attention.  Usually I tend to
go to sleep for a long exposure.

This is an insight on how I work.  I always try to do something crudely
before I build the instrumentation to do it right.   This usually proves
that you don't want to do the thing so one saves a lot of time on
construction.  From time to time you fail because the only way to do
something is to do it accurately.

But nature is not (by nature) very accurate.   So most things that can be
done in nature can be done crudely. 

I am off to implement Arne's Janesick algorithm.  

Tom Droege 

At 09:21 AM 10/26/98 -0800, you wrote:
>Tom Droege wrote:
>> 
>> OK, I think I have everything I need to measure noise.  Can someone outline
>> a procedure?   I should be able to get both the gain and the real noise out
>> of the right experiment.  So far, I have been making up wrong experiments.
>> 
>> I am now somewhere between 5 e- and 20 e- by my (obviously wrong) schemes.
>
>I just checked Christian Buil's book.  Under "Experimental
>Determination of Noise" he Use a method diferent then Arne's.
>He fits a line to a plot of distribution v. ADU.  The Y
>intercept is the readout noise
>
>He does make two interesting points: 1) the stability of the
>light source needs to be *very* good. He gives a figure of
>10E-6/hour 2) You need about 100 measurments.  Is your LED
>pulse width and voltage constant to better then 16 bits?
>
>
>   --Chris Albertson             home: chrisja@jps.net        
>     Redondo Beach, California   work: chris@topdog.logicon.com
>
>
>