shelly uni Problem with Shelly Uni + 3x DS18B20 temperature sensors

[Luke.Starkiller]

Hi
Just bought and set up a Shelly Uni with 3 DS18B20 temperature sensors. Had no problem adopting it, and when adding the 3 addon sensors I could see them all and add them. The problem is that they each lose connection on and off. I can see that they show up as N/A in Shelly Cloud for a couple of seconds and then the value comes back. They don't do it at the same time, often just one of them. But all 3 are affected. I use Home Assistant and I can see the gap in the history chart.

The other problem is that they out of nowhere report about half the correct temp for a couple of seconds then jump back to the right temp. All 3 of the sensors have been lying on my desk for 48h and it's about 20-22c in my office, but they all have some reports of 5-10c over the past 24h.
The higher value 25c+ is when I'm holding them in my hand.

Does anyone have a clue what's going on?



 In shelly cloud shows that temp sensor 1 and 2 reports lower temps than it should from time to time. And the 3rd picture shows whats shelly cloud shows when a sensor loses connection for a couple of seconds before coming back.

Home assistant 15min graph showing the gaps that are reported by all 3 sensors.

In Home assistant 24h graph showing the frequency of the 10c measurements wrongly reported.

[terae]

Some of my clients and -teme- from this forum (when testing) have had also some similiar things happening... It's possibly reason of lack of some filtering of data, when there's some times peak values as kind of status code from recently restarted 1-wire DS18B20 or so... another possibility is some issues with most new batches of DS18B20. How ever error values should be allowed to be filtered out if not automatically done.

Edit: Forgot to tell, that if I haven't got it wrong, it's under inspection to do fix at Shelly (?)

[mschaeffler]

@Luke.Starkiller Do you have an Uni or a PlusUni?

I have seen similar things with PlusUni.

[Luke.Starkiller]

On 4/6/2024 at 3:36 PM, terae said:

Some of my clients and -teme- from this forum (when testing) have had also some similiar things happening... It's possibly reason of lack of some filtering of data, when there's some times peak values as kind of status code from recently restarted 1-wire DS18B20 or so... another possibility is some issues with most new batches of DS18B20. How ever error values should be allowed to be filtered out if not automatically done.

Edit: Forgot to tell, that if I haven't got it wrong, it's under inspection to do fix at Shelly (?)

@terae These are the 3 wire DS18B20.

 

On 4/6/2024 at 3:46 PM, mschaeffler said:

@Luke.Starkiller Do you have an Uni or a PlusUni?

I have seen similar things with PlusUni.

@mschaeffler It's the PulsUni.

Edited April 8 by Luke.Starkiller

[Dimitar]

In next firmware release we implement a filtering which solve such a problems. You will have it in 2 weeks. 

[terae]

11 hours ago, Luke.Starkiller said:

@terae These are the 3 wire DS18B20.

 

@mschaeffler It's the PulsUni.

The technology is called 1-wire. You can feed these from any DC source locally and just transfer data by one wire, another one has to be same ground but not exactly in same cable.

[terae]

10 hours ago, Dimitar said:

In next firmware release we implement a filtering which solve such a problems. You will have it in 2 weeks. 

Great! Thank you for info. Many are waiting to have that solved.

[Luke.Starkiller]

17 hours ago, Dimitar said:

In next firmware release we implement a filtering which solve such a problems. You will have it in 2 weeks. 

Hi Awesome! will it fix the gaps in the measurement or when it sometimes report 10c lower temp and then jumped back?