Temperature sensor accuracy comparison

[mfalkvidd]

Great work @wes, thanks!
Here is some additional information: https://forum.mysensors.org/post/38344 (follow awi's "study" link)

[korttoma]

@wes Thanks for the list of sensor accuracy. But for for example battery nodes there are other parameters that are at least as important as temp accuracy and you left out the Si7021 that is really good for battery driven nodes.

Si7021
Precision Relative Humidity Sensor ± 3% RH (max), 0–80% RH
High Accuracy Temperature Sensor ±0.4 °C (max), –10 to 85 °C
Wide operating voltage (1.9 to 3.6 V)
Low Power Consumption 150 µA active current 60 nA standby current

[zboblamont]

@wes Nice.... Perhaps this could be expanded to operating range also?

My reasoning is my own case where I had already plumped for a chain of DS18B20 on a long Cat5e in the house, one-wire to rule them all :+1: .

The problem I had was requiring two external sensors capable of -10 down to -25c at opposite ends of the house, the only solution I could see being two dedicated nodes.
Only after re-reading the spec did it dawn on me the chip range was -55 to +125, but at reduced accuracy of +/- 2. Made a few enquiries, and was advised accuracy at -25 was typically less than +/- 1 so I saved on two nodes by extending the existing chain both ends and living with the lesser degree of accuracy, all the same devices....
YMMV...

[wes]

Added Si7021-A20 - thanks @korttoma

@zboblamont agreed, and actually it would be great to have a comparison of ALL the specs like operating range, voltage, standby power consumption etc, and cover other types of sensors besides temperature, but it's probably a bit hard to maintain all that in a forum post. Thoughts on best place to set this up? (Wiki page? Google sheets?)

@zboblamont my house actually has Cat5 ports in every room but I never thought of using them to set up a OneWire bus. Genius!

[zboblamont]

@wes Perhaps a Table pinned in Hardware might be the most logical location...?

It was geniuses here that this noob picked it up from, including crimping the DS18B20s into the plugs.
In my own case I used small white low profile telephone sockets to blend into the decor unobtrusively, and crimped into RJ11 not RJ45, but the same principle applies...
If you can line the legs up right and allocate the 3 cores required, swapping out is easy and uniform.

So long as the one-wire layout rules fit your network as installed, you're good to go....

[NeverDie]

@wes said in Temperature sensor accuracy comparison:

DHT22: ±0.2℃ @ ?

Really? That doesn't sound right.

[wes]

@zboblamont said in Temperature sensor accuracy comparison:

crimping the DS18B20s into the plugs.

That's even more genius, love it!

@neverdie said in Temperature sensor accuracy comparison:

@wes said in Temperature sensor accuracy comparison:

DHT22: ±0.2℃ @ ?

Really? That doesn't sound right.

Updated to ±0.5℃. I originally looked at this datasheet which says ±0.2℃; I've since found two other datasheets which both say ±0.5℃. I suspect whoever wrote the first datasheet was getting confused between accuracy, repeatability, and resolution.

[remis]

Good topic :Accuracy, and voltage range.

I don"t want to use any more the DHT22. Humidity results are not comparaible with sensiron and other captors results.
We can add humidity and accuracy...
Lot of people forget to filter power supply, and check voltage range with measurements.

[NeverDie]

Humidity accuracy will really sort them. Vast differences.

[hek]

@wes
If you create a project/page on openhardware.io about the temp sensors, we can embed it on mysensors.org.