If I use Rg=Rf=1.0k I would expect the noise due to the gain and feedback resistors to be reported as 1.414*4.0 = 5.6nV/rHz.
The DiffAmpCalc says 5.2nV/rHz. Can some confirm if this is an error?
The current version of the ADI diff calculator has a bug and it interprets the temperature as the negative equivalent of the value in the "Ambient Temperature" box. A new version will be released within the next three months that will fix this bug. In the meantime, you can enter "-25" in the "Ambient Temperature" box if the intended ambient temperature is 25ºC and it will give you the exact and correct noise density which is in your case 5.7nV/rHz.
Retrieving data ...