I have a Sonoff SNZB-02 in my bathroom to control the vent fan. When I take a shower, the humidity turns the fan on, then off when it gets lower.
I had simply had a rule the if the humidity got over 75%, turn the fan on, under 75, turn it off.
Last year, when the weather started getting colder, it kept showing the humidity as being over 75%, and would just turn the fan on. I chalked it up to not having the AC on in the house, and in the spring, it seemed to return to normal. I'm having the same issue again, so I bought another sensor, this one being Centralite. I put that one in the hallway outside the bathroom. With the bathroom door open, the two are about 8 feet apart, and no corners in between.
With the bathroom door open for over 12 hours, the on in the hall is constantly in the 50s percentage wise, and the one in the bathroom is in the 70's, or high 60's. I know that these sensors are not always the most accurate, but a 20% difference between them?
Does anyone have any suggestions, including maybe another brand that's more accurate but won't break the bank to get?
I recently wrote a post about trying to dump as much z-wave and zigbee as I can as one of my resolutions for the new year, and these infernal little temp/humidity sensors inspired it (one of them was a Sonoff, but others were not better. I've found them to be often inaccurate, especially for humidity. I switched to using Ecowitt devices and the driver from @sburke781 and couldn't be happier. No more z-wave or zigbee temp/humidity if I can avoid it.
Maybe I am lucky, but I have a pair of SNZB-02 that have been quite good for 3-4 years. I doubted the readings when I first got them, so I did the salt humidity test and found them to be spot on.
I have a similar routine for our bath fan, with the SNZB-02 installed just outside the shower. We recently installed an Ecobee Premium thermostat in our hallway which includes humidity sensing. The two have been tracking quite well and converge after showers. Looking now - 90 minutes after shower/shave - they are already back within 4%.
My sample size is only n=2, and the units are early production (allowing for possible changes over time?). However, I received way more performance than the price paid.