ESP8266 Community Forum

I'm working on a module that uses an ESP12E and some DHT22's for temperature and humidity reading. I use 2 sensors on each side of the enclosure to calibrate the reading better.

I initially tested the setup on a breadboard, with no enclosure obviously, and the readings were fine. Once I 'cramped' up everything in a relatively small plastic box (7cm x 4cm x 1.5cm), the readings started going up a few degrees compared to the previous breadboard test and I have reached the conclusion that it's because of the ESP emitting heat when on. (I'm working on having the readings with 30-60 seconds deep sleep in between so this should not be a big issue going forward but still...).

The DHTs are not inside the enclosure, I glued them on the sides but there is still some heat transmission it seems.

I could do a 10% temperature subtraction of the reading before sending it over to the database but I'm not very comfortable with that as I would rather the sensors read the right stuff.

Another option would be to add a 3rd sensor on the board and come up with some sort of algorithm there to calibrate the other sensors. I would like to avoid adding more sensors as the I/O pins are few, as you all know.

Thanks,
Andu

I had the same issue with sensors I made using DS18B20. Even mounting the sensor on the outside of the box were up by 2C on reality.

My solution was to have the sensors on a short stub cable and now there is no offset.

https://github.com/roberttidey/ds18b20- ... Sensor.jpg

The box here is 36mm x 22mm. The stub (65mm) shown is much longer than actually needed to reduce the offset butI use different lengths so I can hide the box and have the sensor 'poking' into area monitored.

btidey wrote:I had the same issue with sensors I made using DS18B20. Even mounting the sensor on the outside of the box were up by 2C on reality.

My solution was to have the sensors on a short stub cable and now there is no offset.

https://github.com/roberttidey/ds18b20- ... Sensor.jpg

The box here is 36mm x 22mm. The stub (65mm) shown is much longer than actually needed to reduce the offset butI use different lengths so I can hide the box and have the sensor 'poking' into area monitored.

Appreciate your feedback. Indeed, the deviation I also see is of ~2 degrees.

So you had to position the sensors further away to get the right reading. OK. They look good in your setup but I actually hoped to have the entire thing in a small box and hung it on a wall. More of a nimble solution.

Not to mention I also wanted to make this fully wireless and when the battery is charging, the heat form that will give me more headaches..

I was reading online that the NEST thermostat 3 has like 10 sensors in it. I guess that's how they manage to have all of that in a box, by constantly correcting the readings..

PS. I'm thinking about replacing the DHT22s with the Bosch BME280s. Not that they would be a fix in this case but because they are said to be more reliable?

FYI for others who may land here...

Seeing about 7 degree Farenheit increase from breadboard measurements when placing the DS18B20 in an enclosure with the 8266... the enclosure is 75x54x27mm and the 8266 is a Wemos D1 Mini

Enclosure also contains buck converter, capacitive touch switch, passive buzzer/transistor, and RJ-45 keystone for two reed switches