WeMos D1 Mini - DHT22 shield

[patvdleer]

I've been playing around with monitoring temperatures and it seems the WeMos D1 Mini DHT22 shield is kinda crappy and unreliable... Due to the heat generated by the ESP beneath the DHT22 slowly heats up. I.e. making the DHT22 shield useless... The second column below is the shield mounted directly on top of the ESP.

Cold boot reading, (after a night on my desk)

Code: Select all

DHT11   22 #1   22 #2   MCP     MCP
18.00	20.40	20.20	20.25	20.25 
18.00	20.40	20.20	20.31	20.25 
18.00	20.40	20.20	20.25	20.25 
18.00	20.30	20.30	20.25	20.37 
18.00	20.40	20.30	20.25	20.31 
18.00	20.40	20.30	20.25	20.37 
...
18.00	20.40	20.30	20.31	20.37 
18.00	20.50	20.30	20.31	20.37 
18.00	20.50	20.30	20.31	20.37 
...
18.00	20.50	20.30	20.31	20.37 
18.00	20.60	20.30	20.31	20.37 
18.00	20.60	20.40	20.31	20.37 
18.00	20.70	20.40	20.31	20.44 
18.00	20.70	20.40	20.31	20.37 
18.00	20.70	20.30	20.31	20.44 
18.00	20.70	20.40	20.31	20.44 
18.00	20.80	20.40	20.31	20.44 
18.00	20.90	20.40	20.37	20.50 
18.00	20.90	20.40	20.37	20.50 
18.00	20.90	20.40	20.31	20.44 
18.00	20.90	20.40	20.37	20.44 
18.00	21.00	20.40	20.37	20.50 
19.00	21.00	20.40	20.37	20.50 
18.00	21.00	20.50	20.37	20.50 
18.00	21.10	20.40	20.37	20.44 
19.00	21.10	20.40	20.37	20.50 
19.00	21.10	20.50	20.37	20.50 
19.00	21.10	20.50	20.37	20.50 
19.00	21.10	20.50	20.44	20.50 
19.00	21.10	20.50	20.37	20.50 
19.00	21.10	20.50	20.44	20.56 
19.00	21.20	20.50	20.44	20.50 
19.00	21.20	20.50	20.37	20.56 
19.00	21.20	20.50	20.37	20.56 
19.00	21.20	20.50	20.44	20.50 
19.00	21.20	20.50	20.44	20.50 
19.00	21.20	20.50	20.37	20.56 
19.00	21.30	20.50	20.44	20.50 
19.00	21.30	20.50	20.44	20.56 
19.00	21.30	20.50	20.44	20.56 
19.00	21.30	20.50	20.44	20.56 
19.00	21.30	20.50	20.44	20.56 
19.00	21.30	20.50	20.44	20.56 
19.00	21.30	20.50	20.50	20.56 
19.00	21.40	20.60	20.44	20.56 
19.00	21.40	20.50	20.44	20.56 
19.00	21.40	20.60	20.44	20.56 
19.00	21.40	20.60	20.44	20.56 
19.00	21.40	20.60	20.44	20.56 
19.00	21.40	20.60	20.44	20.62 
19.00	21.40	20.60	20.44	20.56 
19.00	21.40	20.60	20.44	20.56 

From left to right, DHT22 #1, DHT11, MCP9808 #1, MCP9808 #2, DHT22 #2

[deshipu]

This is a well known problem with circuits including temperature sensors in general. The solution to that is careful isolation between the sensor and the heating parts, and also heatsinks etc.

Have you tried to put the shield *under* the ESP? I wonder if that would have helped.

[patvdleer]

I wondered the same thing so I did, currently letting it sit for a few hours to let it cool down and get back to room temp. I will take new measurements later today.

[patvdleer]

Same shit...

Code: Select all

19.00	20.70	21.10	21.00	21.06 
19.00	20.70	21.10	21.00	21.06 
19.00	20.70	21.10	21.00	21.06 
19.00	20.80	21.10	21.00	21.06 
19.00	20.70	21.10	21.00	21.12 
19.00	20.80	21.10	21.00	21.06 
19.00	20.80	21.10	21.00	21.06 
20.00	20.80	21.10	21.06	21.06 
19.00	20.80	21.10	21.00	21.06 
19.00	20.90	21.10	21.06	21.12 
19.00	20.90	21.10	21.06	21.12 
19.00	20.90	21.10	21.00	21.12 
19.00	20.90	21.10	21.06	21.12 
19.00	21.00	21.10	21.06	21.12 
19.00	21.00	21.10	21.00	21.12 
19.00	21.00	21.10	21.00	21.12 
19.00	21.00	21.10	21.00	21.12 
19.00	21.10	21.10	21.06	21.12 
19.00	21.10	21.10	21.06	21.06 
19.00	21.10	21.10	21.06	21.12 
19.00	21.20	21.10	21.00	21.12 
19.00	21.20	21.10	21.06	21.12 
19.00	21.20	21.10	21.00	21.12 
19.00	21.30	21.10	21.06	21.12 
19.00	21.30	21.10	21.00	21.12 
19.00	21.30	21.10	21.06	21.19 
19.00	21.30	21.10	21.00	21.19 

[deshipu]

Good to know, thank you for testing. I guess is a real use case you will want to have some forced air flow in there.

[patvdleer]

Sadly my attempt at plotting is failing horribly...

http://stackoverflow.com/questions/3790 ... -animation

[patvdleer]

You can clearly see the DHT22 #1 rising way more than the rest

[patvdleer]

Dropped the DHT11

[deshipu]

It seems that all of them are growing, maybe not as fast, but still.

[patvdleer]

Slightly yes but not as much as the shield directly on top of the ESP8266, 19.3 to 21 so ~1.7c rise vs 0.3/0.4c with the rest of the sensors.
I'll pick it up tomorrow and try to edit the code so that one of the MCP9808 will go into sleep/shutdown mode. The difference should be 0.1mA vs 200mA so I guess it should be visible in the temperature deviation as well.