Recommendation for humidity sensor for Humidor Monitoring Project

[Crix1990]

Hello together,

sorry for my uncertainty on this topic, I'm using Arduino stuff not that often and learn it on try & error base.

So I'm working on a project to monitor the temperature and humidity (mostly!) on my Humidors (the lockers you put your cigars in) with an ESP8266, including web interface, OLED, mail notification, etc.

For the programming, everything seems to work as it should, but the sensor values are a huge issue.

Based on many recommendations I went with a BME280 via I2C (mounted via flat ribbon cable, so it is away from the ESP).

While the temperature seems ok to me, the humidity value is simply wrong (also the approx. altitude is bullshit).

So I own a calibrated hydrometer so I have a reference.

At 75% humidity the MBE280 shows 53%, at 52% it shows 41%, so unfortunately it is not a fixed offset...

So I assume that the BME280 is simply not the sensor I need.

Following this List the SHT31 would be good, but the SHT85 would be perfect:

Compare different i2c temperature and humidity sensors (SHT2x, SHT3x, SHT85, ..)

zoomx: BME280 temperature overestimate surprised me but I discovered that datasheet talk only about pressure and humidity and not temperature and you don't find information about temperature accuracy. Did you found sensors that are out of specifications? I do not trust in accuracy numbers from datasheets (in general). The only uses for datasheets are i2c port and commands. One of the reasons is I use (mostly cheap Chinese) modules, not sensors itself. AHT15 and SHT85 are only exceptions whe…

Can I address them nearly the same via I2C?

So that I don’t need to completely redo the programming.

Or do you have other recommendations?

I'll read the sensor about every 2-3 seconds and a difference of about 2% would be fine for me.

Thank you and regards,

Felix

[xfpd]

@Crix1990

not a fixed offset

Gather many measurements from both devices, graph their difference at each percent. If the graph is not linear (or does not follow a math curve) then hard-code the translation from BME to REAL at each percent. I would guess that temperature will also effect both readings, so you might need a matrix of temperature vs humidity-difference at every temperature inside your code.

[robtillaart]

• Compare DHT22, AM2302, AM2320, AM2321, SHT71, HTU21D, Si7021, BME280

An elaborated comparison of humidity sensors, which give plenty of information and although it is a bit dated (some dead links and older sensors) It is still very informative.

As far as have used humidity sensors the SHT85 has the best numbers (datasheet).
It is unfortunately not in the above comparison.
It is not the cheapest sensor and might be hard to get.
Wrote a library for it, see - GitHub - RobTillaart/SHT85: Arduino library for the SHT85 temperature and humidity sensor

I have not worked with the newer SHT40, SHT41 and SHT45 from Sensirion yet, these might even be better than the SHT85, and/or cheaper. See - Adafruit Sensirion SHT45 Precision Temperature & Humidity Sensor [STEMMA QT / Qwiic] : ID 5665 : Adafruit Industries, Unique & fun DIY electronics and kits - for library and more information.

[Crix1990]

Thanks for your replies.

So I checked the BME280 a little more and unfortunately, it seems like the temperature has also an effect on the measured humidity.

At 30°C 75% was measured as 53%.

At 25° 75% was measured as 61%.

So yes, I thought a transfer table would help, but I think this might be a too complicated, since it is already not linear at the same temperature and has also a dependency.

I'll try the mentioned SHT45 next, since it seems to have better I2C support the SHT85 and is available in Germany to an accurate price.

Regards,

Felix

[xfpd]

@Crix1990

not linear at the same temperature and has also a dependency

The difference could still be linear, but with a different slope. Get x readings at x temperatures. Graphit.

[Crix1990]

@xfpd

r, but with a different slope. Get x readings at x temperatures. Graphit.

Yes, it still looks kinda predictive, but I pulled out a complete second set of hardware with the same config and both BME280 have a drivt about 3% to each other, and even this is not canstand.

If I want to scale the project in the end (I plan up to 5 sets of hardware), I really don't think that a individual programmed calibration for each sensor is worth it for me.
I think I prefer then to spend more in a better sensor.

Hopefully the SHT45 is more reliable...

[Crix1990]

Hi together,

to give you feedback on this, I built now a second hardware set with the SHT45 and it works drastically better.

It is always about 2,5% behind my analogue hydrometer, but honestly, I would doubt now if the SHT45 or the hydrometer was tolerances (or if there is still a difference in the humidity.

But still, both BME280 are about 10-15% away from the others (even in the salt test…you know the 75% one) and have a drift of 2% to each other.

And you can not tell me that in this test a difference of 10%+ can have a “not faulty” reason.

As a last question, and maybe I simply not found it, does anyone know the use of the heater function of the SHT45?

Regards,
Felix