Env Tails readings are different

Hardware
1

I have 3 env tails all sitting right next to each other, firing at the same time.
The temp and humidity levels are all different. Is there anyway to calibrate these devices?

Time - Temp - Humidity - Pressure - Voltage - Reading - voltage reading
1/1/2016 3:42:55 PM - 75.5475 - 39.8717 - 1010.38 - 3.291 - 23493 - 1.17976
1/1/2016 3:42:55 PM - 74.6788 - 41.5807 - 1011.57 - 3.282 - 22565 - 1.13006
1/1/2016 3:42:52 PM - 73.9645 - 42.1606 - 1011.88 - 3.263 - 22757 - 1.13308

I am going to be comparing temps in different areas of my house and with these inconsistent readings, the devices will be useless.

2

Here are the accuracy specs of the tail:
Silicon Labs: Si7021-A20

Precision Relative Humidity Sensor
± 3% RH (max), 0–80% RH
High Accuracy Temperature Sensor
±0.4 °C (max), –10 to 85 °C

We don’t know the “actual” RH when you took measurements, but let’s pretend
the real RH is 41%.

All 3 would be within spec: 39.77 - 42.23%

The temp is +/-.4C … that’s about .7F
Again, we don’t know the “actual” temp.

Looking at a few other sensors sold by Sparkfun …

Sparkfun RHT03 …
specs: +/- 2% RH, temp +/- .5C

Sparkfun HIH6130 …
specs: +/- 5% RH

Sparkfun SHT15 …
specs: +/- 2% RH

You need to get a hold of an NIST traceable standard for RH and temp. Put all of your sensors in a box with the traceable standard and note the difference from the standard. That would be your value for error (compensate for the difference).

Otherwise, just know that your sensors are not reading any more accurate than what they are supposed to read. Basically, you get what you pay for.

3

Temp readings in particular can be very easily skewed by air currents and any sort of local heating no matter how slight it may seem to human perception.

The tails use MEMS sensors which are @mlseim pointed out have accuracy specs. If you want better accuracy you need to pay significantly more for calibrated sensors (likely hundreds of dollars for something with a calibration certificate).

Wall thermostats are often much less accurate than these sensors…

4

Hugo and Mlseim thank you for your notes. After the sensor’s have “calmed” down and been running for 24 hours, they are getting more even. All 3 with in .5 F. So thank you again. The units are performing just right.