Hello folks,
I have multiple (cheap) temperature and humidity sensors and I trust none of them with humidity measurments. I've got few Sonoff SNZB-02 and few Tuya TS0201. Even though sensors were in the same room, sitting next to each other, each of them showed different humiditiy measurements. So I calibrated them with "salty" test. In short, the Sonoffs were off by 10-13% (showing 85% instead of 75%), the Tuya were 2-5% off (showing 71% instead of 75%). I calibrated them (using kkossev driver - thanks BTW) and then again put them in a same room one next to each other. The measurements were "all over the place" with Sonoff showing ~60%, while Tuya showing ~70%...
What is going on? Anyone else with this kind of behaviour? I know I cannot expect spot on measurements of such a cheap sensor but still....
Any thoughts?
Best regards!
1 Like
It’s likely that the sensors these devices rely on have an error range of +/-5% or even +/-10%. I don’t know that for a fact since I’m not sure what’s inside each device, but that’s typical for devices in this price range. Greater precision would require more expensive components.
So even if you calibrate the sensor, it’s pretty likely that subsequent numbers could drift over a fairly large range because it isn’t designed to do better than that.
1 Like
Mark nails it above.
I would additionally contend that many users (me included) don't need these sensors to give a dead-accurate reading at any one point in time; rather, I just need them to reliably determine useful changes.
For instance - I don't ever really care what the actual humidity % is in my bathroom, but I want that sensor to reliably tell me when that normal value suddenly jumps by 5% so that my fan kicks on.
I have a window fan that exhausts out when the AC can't keep up in that upstairs bedroom on hot summer days... It triggers at 82. Is it actually 82 in that room when it triggers? I don't know and I don't care -- going by feel when I set that rule up, 82 was just what I picked.
In these cases, the reported value is just used as a reference point -- I need that value to be reliably consistent, but I don't actually care if it's reliably accurate,
2 Likes
AFAIK inexpensive humidity sensors are just kind of ... Bad. And the more extreme the conditions, the greater the error.
If you need the best output possible you may need to get a lot of sensors and try to pick out a group of them with the most similar readings. They still won't be showing you the NASA-grade true value but at least the group will be more internally consistent.
FWIW I have found my Ecowitt weather station sensors to be pretty OK for cheap stuff, but that requires an Ecowitt wifi hub and using a 3rd party integration. And at the end of the day it still may not be better than what you have now.
Thanks for response.
I tested already calibrated sensors and Tuya was spot on, Sonoff was off again. I'll calibrated it again and we'll see later.
Yep, the salt calibration method seems to be the easiest one and precise enough for our needs.
