humidity readings

[manjh]

I have a setup with twi separate ESP units with several sensors attached.
On one of them I use the good-old DHT22. on the other a newer BME280.
What worries me is the difference in readings.
On the BME I can see a humidity value of 42%. The DHT tells me 55%. And just for checking I added a standalone digital unit, which reads 61%.

Is this "normal"? I would expect the BME to be correct, but it worries me that two other devices are so far apart.
Should I calibrate somehow?

[Shardan]

Hello,

the BME280 has a tolerance of +/- 3% for humidity following the datasheet of "Bosch Sensotec".
I'd calculate more if you're using a chinese copy.
Says: If real pressure is 70% you may read a value from 67,9 to 72.1 %rH

The DHT22 is similiar. Most DHT i tested had somewhat big tolerances so i dumped them.

As i don't know your standalone device i don't know how accurate it is.

One point with humidity metering: Devices must have the same temperature.
It is called "relative" himidity for a reason as the amount of water the air can hold depends heavily on the temperature.
Give all devices a lot of time to settle down before comparing.

For a real calibration you need a highly accurate reference meter.
These are expensive and rare, don''t take any 30€ Hygrometer from the market for accurate.
To give an example: If you want to calibrate to a tolerance of +/1 1%rH you need a reference meter that is reliably
calibrated to 0,1% rH ! i doubt that any normal hygrometer is that exact.
As an example for well calibated home sensors take a genuine SHT15 made by Sensirion (not a rebuild from AliExpress..).
This sensor has an accuracy of +/- 2% rH, it is laser calibrated at manufacturing. And it comes around 30€.........

Some more info about accuracy and calibration can be found in the Wiki:
https://www.letscontrolit.com/wiki/inde ... d_Accuracy

Regards
Shardan

[manjh]

In my case, the absolute accuracy is not a priority. If the sensor reads 55 and the "real" value is 52 or 58, I can live with that.
But what bothers me is the fact that I have several sensors, with readings that are 10% apart.
I am considering putting all of them together in an (almost) airtight container, so I know they have the same environment.
Then I can play with it: putting a wet sponge in there should raise the readouts.
The standalone device can't be tinkered with, but I can certainly adjust the ESP attached sensors...

[Shardan]

Right.
Just be patient and wait until the devices settle down inside the contaier - it definitely takes a while.
I'd recommend half an hour as a minimum.

Regards
Shardan

[manjh]

After several days of confusion I decided something must be wrong. The BME280 was consistantly giving readings about 10 percent points lower than the other sensors, including the standalone cheap Chinese unit.
The peculiar of all this is that most readings were quite close together, except for the BME.
Hardware: I had two individual ESP units. One contained a DHT22, a BMP085, a BH1750 (Lux sensor), a Senseair S8 and an 4-line LCD panel. The second unit had the BME280 and an OLED panel.
The OLED was very sharp, but for my old eyes the good-old LCD was much better. So I decided to pull the BME280 from the 2nd ESP unit and plug it into the I2C bus of the 1st.
Re-arranged the LCD screen lines somewhat, so I could display all values on a single display.
Now the readings are much better! Temperatures are all within about 1 degree apart, and air pressure within about 4 hP!
Not sure what caused the earlier weird readings, but I am quite happy with what they are now.


What's not in the picture: the standalone unit shows 24,4C and 61%.