ESP8266 Community Forum

- Thu Jun 09, 2016 6:28 pm #48891 I still don't know why the performance was suddenly better. Maybe because the BME was running for a while? Or maybe because I used the I2C bus more frequently with more sensors? I should try to deduce that in a test project. I don't have the time now, but maybe you can test multiple sensors too and try to let the BME run for a few days?

Regards,

Robert

- Sat Sep 24, 2016 8:22 am #55611 I have same problem:

http://www.esp8266.nu/forum/viewtopic.p ... 9419#p9374

- Mon Sep 26, 2016 12:54 pm #55722

I have same problem:

http://www.esp8266.nu/forum/viewtopic.p ... 9419#p9374

I also still have this problem. The strange thing is that using an Arduino uno makes the problems go away...

- Wed Nov 09, 2016 12:12 am #57860 Temperature and humidity are very difficult to measure electronically. Mercury thermometers are generally the most accurate, but even these thermometers need to be calibrated by the NIST (National Institute of Standards and Technology.) Or traceable to a NIST calibrated thermometer. The range of your numbers are quite good for the devices you are using them for measurement, though the "real" temperature may be outside the range, or inside it. To give you an idea of temperature variability, the difference between temps measured over one urban block can be as high as 13F. Unless your accuracy needs are absolutely critical, then be happy with the numbers you have. If necessary you can always calibrate them against a NIST certified thermometer, but I warn you that the procedure is rather expensive.

Just as an aside, the average body temperature of healthy humans is 98.3F. But the MD who did the measurements of what was a large sample used slightly mis-calibrated thermometers. It was easier and far cheaper to just call the average body temperature 98.6 than to change all the thermometers in existence.

I added another BME280 with another I2C address. My output at the moment is:

HTU21: Temp=22.06 Hum=30.91
SHT31: Temp=21.82 Hum=31.05
BME280: Temp=22.66 Hum=33.75 Pres=103272.67
BME280/2: Temp=22.60 Hum=33.47 Pres=103227.92
HDC1008: Temp=21.80 Hum=25.82
MCP9808: Temp=22.31

The temperature is now getting close to the real temperature. I didn't do anything to accomplish this!
Humidity is still higher than the HTU/SHT, but I don't have a reference, so maybe they are off.

To recap, I installed the latest Adafruit library (I am using BME Adafruit breakouts by the way), added two pull-up resistors of 2K2 to both SDA and SCL. I am powering the ESP from my laptop through USB at the moment. I don't use Wifi in my sketch, just printing the values.

Why it is getting better?

Hi RKieboom!

I have been testing like you the BME280 sensor and have had large variations compared to DHT22, SHT21 and BMP180. These three sensors marked similar values, however the BME280 marks always more or less 1 °C more of temperature and about 10% less moisture.
I tried using the library as comets Adafruit but the Serial gives me the error "Could not find a valid BME280 sensor, check wiring!"
I am using the "Wemos D1" board which is based on the ESP8266, and in my case instead of using the breakout of Adafruit, is one purchased in Aliexpress.

I would greatly appreciate if you could help me.
Thank you very much.