Calibration of DHT sensor AM2302

Projects Programming
1

I have a project where I use an ESP8266 unit (ESP-07) as the controller for a house monitoring system.
So far I have used DHT sensors (AM2302) I bought several years ago. I have checked these before using the humidity calibration method of having a container with humid salt and all 4 of these sensors showed expected readings.
Now I have connected a newly bought sensor (I got a batch of 10 AM2302 units) and its temperature seems OK but humidity is way off, like about 20% too high...
So I got the calibration chamber out and put both an old and the new sensor into the box.
Result after about 15 minutes:
Old sensor: 27.1 C and 69.5%
New sensor: 27.0 C and 91.0 %

So temperature seems to track but humidity is way off!

Is there an explanation for this and is there something that can be done to fix it?
Maybe there is some calibration constant one can adjust?

2

Have you tested all 10 new sensors? Maybe that first one you tried is bad. Did you buy on eBay? I think sometimes eBay sellers buy rejected batches from factories at much reduced cost. The batches are rejected because more than a certain % of units from a sample from the batch fail a test. The eBay sellers test the units infividually and sell on the working units. I suppose their testing may not be very thorough or they might just make a mistake, or be lazy and hope you will not complain if one in ten units are bad.

3

I will make a breadboard where I can easily change the units to test the remaining ones.
Currently all is soldered in place...
And, yes, I bought them on eBay.

The strange thing is that I have recorded the data from the logger and at first the two sensors were fairly the same, like:
(3rd column shows the sensor number)
2018-09-01 14:12:23 1 27.5 43.7
2018-09-01 14:12:23 2 28.5 41.6
2018-09-01 14:17:23 1 26.9 45.8
2018-09-01 14:17:23 2 25.9 46.3
2018-09-01 14:22:23 1 26.8 46.2
2018-09-01 14:22:23 2 25.9 46.9

Then I tried to make the new one register a different value so I moved it a bit:
2018-09-01 15:43:28 1 25.9 50.4
2018-09-01 15:43:28 2 25.5 69.4
2018-09-01 15:48:28 1 25.8 50.6
2018-09-01 15:48:28 2 25.5 69.5
2018-09-01 15:53:28 1 25.9 50.1
2018-09-01 15:53:28 2 25.6 69.6

Finally I pulled out the moisture chamber and started that test (target humidity is 75% if I remember correctly):
2018-09-01 17:28:28 1 27.3 70.2
2018-09-01 17:28:28 2 27.1 91.1
2018-09-01 17:33:28 1 27.2 70.3
2018-09-01 17:33:28 2 27.1 91.1

Maybe I broke something when I tried to get different values on the new sensor...

4

Is it possible the "faulty" sensor got a drop of moisture stuck in it? Maybe disconnect and leave in a warm dry place for a few days and then re-test.

I use am2320b sensors on my weather station. Two have failed so far and I have replacements on order. They don't survive the winter outdoors on a high moor in Yorkshire!

5

Drop of moisture?
Well, that might have happened since I tried to breathe on it to raise the temperature and humidity...
I have now tested one of the remaining 9 sensors and it seems to be OK.

My use case is to have 3 sensors in my newly built attic of my home in order to monitor temp amd humidity to check that there will not be conditions for mold to develop on the inside of the roof. So I have placed 3 such sensors in each end and in the middle of the attic. They will be subject to outdoors temperatures which means down to -20 C or so.
Hopefully the sensors will survive since they are inside the attic. The attic is very low, only 50 cm high so it is impossible to enter it after construction. Sensors were placed during construction but I have not connected these yet.

6

I would not have thought that breathing on the sensor would damage it. I was thinking more along the lines of an Atlantic storm.

In the relatively protected environment of an attic, your sensors should have a much longer life than my sensors do outdoors.

50cm is not what I would call an "attic". What is the construction?

7

PaulRB:
50cm is not what I would call an "attic". What is the construction?

Well my house originally had a flat roof but this resulted in numerous rain water leaks and we (and our neighbours with the same house type) got fed up so this spring we modified the roof by adding a new top with 6 degree slant from the center. We had to keep it low due to building regulations in our area...
So that is why it is only some 50 cm high. But rain water now drains off of it.
Since it is hard to inspect I added the sensors inside to monitor the climate.