[time-nuts] Bob warned me......

[Erik Kaashoek]

Yes, Bob did warn me but I decided to try anyway.
I measured the ADEV of a cheap VC-TCXO in the past starting starting 
with a tau of 1 second and the ADEV nicely started around 1e-10, well 
within the required spec. But there was a hint there where some 
stability issues depending on the Vtune so I measured the ADEV starting 
at much smaller tau (starting at 50ms) and the relation with the Vtune.
The results where remarkable, to say the least and, can be seen using 
this link: http://athome.kaashoek.com/time-nuts/TCXO%20Vtune.png
Ignore the rise of the ADEV with larger tau as the very large capacitors 
used to get rid of any ripple or noise where still (dis)charging
As long as the Vtune was below 0.5 Volt the ADEV at tau=50 ms measured 
about 1.5e-10 (yes, I confirmed using a very stable OCXO that the used 
frequency meter can measure the ADEV with this resolution and short tau)
But as soon as the Vtune got above 0.5 Volt the ADEV at 50 ms started to 
rise and at Vtune of 1.5 Volt the ADEV was 3e-9
Looking at the frequency deviation there clearly is some repetitive 
active switching in the TCXO, no clue what.
Now I need to find a cheap TCXO that behaves a bit better, although, the 
ADEV at tau of 1 s was still well within my target spec but one can hope 
to find something better