I currently have three Davis stations and just went through calibrating the outside and inside temperatures for each. I'm pretty surprised at the delta's I had to put in each to make them match my pair of NIST Traceable mercury full immersion thermometers. My two mercury thermometers are consistently in sync and were last calibrated Oct 2016 when they were manufactured. Here are the offsets I had to enter in each of my stations for outside and inside temps to match the NIST Traceable thermometers.
Inside NIST Outside NIST
Davis Vantage Vue -5.5c 70f 3.4c 68f
Davis VP2 DFARS -7.7c 70f 2.7c 68f
Davis VP2 SHT31 -4.4c 70f 0.7c 68f
No surprise the SHT31 is the most accurate of the bunch needing only 0.7c adjustment. But, 3.4 and 2.7c seems extreme. The Vantage Vue is brand new only being installed yesterday. The inside sensors were way off needing up to -7.7c, that's 18.14f off or over 12%.
When calibrating the outside sensors, the NIST thermometers were in a radiation shield positioned within 18" horizontally and 3" vertically with the ISS sensors .
Another interesting observation. I wondered how much impact a radiation shield had on a day like today. It's breezy with winds gusting to about 5mph and it's overcast so there's no real hard sun. I placed one of the mercury thermometers outside of the radiation shield and left one inside. I measured a 9f delta between the two. So even when you don't think it would make a big difference, radiation shields are a big deal. When we have a hot summer day, I'm going to repeat that test and see what the delta shows.
Not bitching or asking for advice, just sharing some observations.