[time-nuts] Baseline measurement with different standard. (same timebase for clock and DUT)

[Taka Kamiya]

I hear a term equipment noise, base line noise, and noise floor of the equipment all the time. I had a few concerning and unexpected result with recent tests so I've conducted tests of testing methods.  I'd like to verify, my understanding of the situation is correct.  I basically did baseline test of the same interval counter with different time standards.

I took an adev measurement of my time-interval counter using standard-A for both clock and input.I took another adev measurement using the same time-interval counter using standard-B for both clock and input.I repeated the same test with standard-C.Run time was 1 hour and interval was 1 second.  

Method of measurement was, time stamping interval measurement (OK, it was a TICC), clock was 10MHz and input A was fed with 1 pps from the same source.
Each represents Cesium, Rubidium, and GPSD.
My expectation was, I will get an over-lapping (right on top of each other) with very minor variations.  Since clock and the DUT was the same, why would there be any different....
Surprising result (to me) was that they all started with different Adev value (origin) and sloped downwards at the slope -a (same value).  They were EXACTLY parallel from 1 second all the way to the end (which was 900 seconds but run time was 1 hour).  

What did I just do?  How can I make use of this graph?  Why are they all the same slope?  Why are they not overlapping?
I can assume one of the answers to be, clock and the TICC form an "equipment" and that line represents the base line FOR THAT standard.  That's as far as I can get.  If I can get a better explanations, I would very much like to hear it.  I also wonder if this would be the same situation for every counter and every method, but I can do that test later.
Thank you.


--------------------------------------- 
(Mr.) Taka Kamiya
KB4EMF / ex JF2DKG