But then, you tell that "no matter about specs , there is no noise." Do you mean that output of V ref have not noise becouse power V have very low noise? This is absolutely not truth afaik. (yes noise is "low" and not meaningful in your use but there is LOT of noise and many kind of noise what need take count if use these kind of reference chips for other purpose where need high accuracy and low noise level)
Thanks for your message , I am not specialist so to get in-depth about the called noise effect,
but I have do my homework over the years and the " Noise" it could be an friend on enemy ,
and this depends from the application.
At the current design " reference stabilization circuitry " , the hostile " Noise " are the one who cames in from the converted AC to DC .
In theory the " stabilization circuitry " , it should be able to suppress " filter " the input source,
so to help it even more, at becoming " Clear and stable " .
I know that I am not saying anything new up to here .
I agree that in any conversion , we have unwanted factors , like a bit of extra noise or heat.
But the major question are :
Does this amount of side-effects haves an negative influence in our circuitry ?
And the answer comes by it self , by looking at what this design it will be connected with .
The " Noise " as cause , its the major enemy of any circuity used at modulating acoustic frequency.
Like pre-amplifiers and broadcast systems ( example the FM band transmitter ).
If you are happy , with my new description like .... " this circuitry its an (Very Very) Low-noise one "
then I have no reason to defend it , as we all do agree on that.
Personally, I have so much faith to it , that I did not add any capacitor on it , as extra filtering
Additionally .. In praxis the Reference LT1021 stabilizes to 10V with just 10.48 volts as input voltage.
The white paper suggests 12V stabilized input voltage , and I have add 15 V ..
The schematic says that any voltage above the 12V it will cause an small drift .
So I will remove as test, one battery from the chain , and test to see if the Fluke 87-5 it will detect the change, as an perfect 10.0.0.0. volt indication.