Internal tempsensor

[EveWes 0]

I have not figured out if 990.0.1 and 990.7.1 is the same internal temperature sensor or two sensors?
990.7.1 i know reports overheat when >= +85C.

 

 

Please login or register to see this image.

/monthly_2020_12/image.png.24c59fe953405e5298959c1b61c4be11.png" /> 

[Tim__ 112]

It's a good question!  My take is that the .7.1 is the internal device temperature.  Not sure what the .0.1 represents but keen for some one to explain to me!

[akatar 207]

WHat is the purpose of the temp sensor? they are 12 degrees higher then the actual temp.

in my case: actual temp is 15 degrees, temp sensors implant are 25 and 26 degrees

 

why????

[pnutp0wer 22]

The internal temp sensor is for 'safety' and 'overheating' purposes. I.e. if device is to hot than ........

[Martin_N 143]

I preferably use an internal sensor (.7.1) for temperature monitoring even without the use of a 1-wire sensor. Just calibrate and set the offset. It is an orientation measurement, but fully sufficient in building applications. If the device is stably powered, there are no temperature fluctuations during the warm-up at start-up.
If you do not use outputs on the device, but only inputs, this is an even better option, because there is no warming during switching.
Another important thing. The higher the voltage you supply to the SmartImplant, the higher the internal temperature of the module! I recommend supplying the module with a reasonably low voltage, depending on the line losses. (simply because its operating temperature may exceed and unpredictable behavior may occur)

[bonhomme 3]

I realise this is an old thread, but what the hell, might help someone...

 

I did some analysis of this because I wanted to use an SI to measure temp somewhere (for a different reason ... basically I didn't have another sensor and temporarily rigged up an SI I'd bought for a different purpose). I found the internal sensor gave ridiculously high readings, searched for the issue and found this thread.

 

Since then, I connected a DS1820B sensor to the SI which enabled me to compare both readings. 

 

Please login or register to see this attachment.

 

I think it would be difficult to set a working offset (or it would be very rough) - since the required offset seems to depend on tempterature. I pulled the data from the Fibaro API and did a regression - you can see that the required offset would depend on the ambient temp, and also that there is quite a bit of variance.

 

Please login or register to see this attachment.

 

Please login or register to see this code.

 

Hope that's of use to someone

Edited May 4, 2023 by bonhomme

[Fabir 64]

@bonhommeThank you for the analysis. I've noticed the same issue, the SI internal temp sensor isn't reliable.

 

Therefore in this use case, I always connect DS1820B sensors to the SI.

[bonhomme 3]

Looks like I posted too soon! After leaving the setup a few more days and collecting more data, it seems like it levels out and the slope stops being significant (p-value 14.7%).

 

Please login or register to see this attachment.

 

Please login or register to see this code.

This suggests you could use an offset of 7.25°C in the case of my setup (it will depend on lots of variables like, how busy your Z-Wave network is and how much the SI is working as a repeater, the input voltage to the SI, etc).

 

Oops! 😳

 

 

[Fabir 64]

@bonhommeThank you, can you tell the specs of the power supply you've used on the SI for this test? V? VA? A?

[bonhomme 3]

Sorry, I missed your question.

I initially used a 12V DC power supply, which I think is max 1A (but not relevant, the SI will never draw that much).

Then I needed that for something else so now I have it running on a 19V DC supply (an old battery charger, I chopped off the charger part and stripped the wires...). It does run a few degrees hotter on 19V.

Having gathered a lot more data since this post, the variance of that difference is huge, I certainly wouldn't recommend using the internal temp for much.

[Fabir 64]

@bonhommeThanks, Cheers!