For years I have been very critical of Davis' use of the Sensirion SHT31 in this thread. I have been frustrated that I have been able to get considerably more accurate humidity and dew point results with a home-brew, older style SHT75. I've been convinced that there was something Davis was doing in their mounting of the SHT31 on it's board that somehow was causing a high humidity bias in addition to the high humidity bias of the Sensirion sensor itself. I want to pass along that after testing a new Davis SHT31, I observed a notable improvement in the performance of the Davis SHT31 that matches the results I get from the SHT75. This is the first time I have ever seen this kind of performance and I've similarly tested 6+ Davis SHT31s over the years.
After hearing from a few people on this forum, I purchased a new Davis SHT31 with a manufacture date in 2021 because they noticed what they thought was an improvement in the humidity/dew point performance of the sensor. The Davis SHT31 I bought has 25ft of cable. Out of the box, it actually read slightly LOWER in humidity and dew point than an SHT75 and SHT35 I have. Never have I observed this. Every other new Davis SHT31 I have ever bought typically ready 3-5% higher than these sensors at room temperature.
I started to use the new Davis sensor outside for a few weeks to let it acclimate and experience prolonged high humidity. Then I co-located my home-brew SHT75, inside the same radiation shield, and compared the readings for a series of days. Humidity during these days ranged from saturation to around 30%. The Davis SHT31 followed the performance of the SHT75 in a way I have not observed before. Previously, as humidity dropped from saturation, the SHT75 would read consistently lower, but with this new SHT31, it stayed right with the 75.
The sensor still suffers from the limitations inherent to the Sensirion sensor. It still is reluctant to reach 100% at saturation (although it has jumped to 100%) and it still has a wet bias in the mid and lower ranges in my climate. I use a slope and offset to help correct this bias and to match it to my Belfort Psychron.
I do not know if I am just lucky with this sensor or if Davis has changed something in their processes that improve performance. There is no obvious or even less than obvious modification to the sensor board that I can tell. We'll never know, but I'm going to be an optimist here and give Davis credit for improving their product. I noticed Sensirion published a document for manufacturers specifically addressing the dos and don'ts of mounting their sensors with conformal coatings as Davis does:
https://www.sensirion.com/fileadmin/user_upload/customers/sensirion/Dokumente/2_Humidity_Sensors/Application_Note/Sensirion_AppNotes_Humidity_Sensors_Handling_Instructions.pdf.
I like to think that combined with feedback, maybe even from this forum and the people on it, Davis took action to improve a deficiency in their product.