That sounds OK. Slightly nervous about running a modified weewx-vantage driver on my production machine, but I guess I will get over my fears:) BTW both machines run ntp to the same server list , so times are always close to identical. Also please be aware that on both machines WeeWx logs to /var/log/weewx.log not syslog. Also the logs get rotated at midnight daily. Values came from weewx.sdb I checked to be sure and there are no corrections being applied by WeeWx or the vantage console. Humidity Values Seen: Last Week 19% - 96% Last Month 13% - 98% Last Year 13% - 99%
Thanks Paul On Wednesday, April 3, 2019 at 8:16:39 AM UTC-4, [email protected] wrote: > > Paul, > > For the parsing of your analog humidity sensor we need probably a look up > table, as we have used similar for the soul moisture and leaf-wetness > sensor data. > > There is a positive relation between the raw humidity value and the > temperature as you can see in the attached table. > Each set with the same raw humidity codes has different humidity > percentages as result of different outTemps. > > For the look up table I will need the raw data over a as big humidity > range as possible. 0-100 % would be ideal, but let's see how far we come. > > Paul, what are the lowest and highest humidity values you have recorded > the last months? With this range we got to work. > And which driver you use on your other computer? Is this the weewx-vantage > driver? > > I suggest we do the following: > 1. I modify both weewx-vantage and weewx-rtldriver so, that they will > output a debug line showing the raw values (rtldavis) and decoded values > (vantage). > 2. You run both programs during a week. The RPI saves 8 days of syslog. > 3. Assuming both computers running these drivers have about the same > system times, we can combine these two sets of debug data in one table and > come with a table similar as attached below, but with a greater range. > 4. With this set we can try to find a correlation formula. > 5. When a correlation formula could not be found, I will compose a lookup > table with segments where the values within can be interpolated lineair. > > What do you think? > > Luc >
