1-wire value (Vad) corruption?

[Rob_in]

Hi all,

This is an odd one so wanted to ask if anyone has seen similar. It could be related to 1 wire extension online, but sensors reporting UNKNOWN status or maybe not, but...

We have Unica 1-wire temperature humidity sensors in each room. Been working flawlessly for couple of years. Then this afternoon just the humidity values started to play up. I record these in InfluxDB & display in Grafana, and here's this afternoon (temperature & humidity):

As you can see, the temperature looks fine, but Humidity is getting a lot of 'out of range' values. It's not just 1 sensor, this is happening to them all.

Checked the 1-wire bus voltage and it's 4.92v - plenty close enough to 5v.

Checked all the connections are nice and tight (besides, I haven't touched anything in the cabinet).

I refuse to believe that every one of these sensors has developed the exact same fault at pretty much the exact same time.

These devices are represented in Loxone as DS2438 devices and the humidity values are read from the Vad input and scaled appropriately. When I look at the 1-wire monitor I can see Vad being reported as spurious. Ie. in the graph above I can leave the Loxone 1-wire monitor running and wait for a bad value to appear, and when it does can then go look in the 1-wire monitor and find the bad Vad value that was read.

I refuse to believe that somewhere in the wiring there is a fault that causes only the Vad value in transmitted packets to be corrupted and somehow Loxone is still accepting said packet (surely a CRC error would occur in any transmission error and Loxone would reject the entire reading).

Given all this it looks to me that there's something wrong in the Loxone Extension or the Miniserver itself. Either hardware or software, that is corrupting these Vad values. I have no idea why, but that seems the only logical explanation. I did changed the config this morning but hours before this fault started and absolutely nothing to do with these 1-wire inputs.

Any ideas?

Cheers,

Robin

[Tico]

I've had sensors reporting corrupt data before. Despite going through a range of fault-finding measures, I could never find anything mechanical or electrical at fault.

The issue turned out to be a worn SD card. Old or nearly worn-out SD cards can present some weird and wonderful error types.

It's now happened enough that changing the SD card is my first 'go-to' when fault finding.

[Rob_in]

Well, this is rather odd... as quickly as this issue appeared, it has now vanished.

I am highly suspicious of this kind of thing but there was more more clue: our outside humidity sensor did not seem to return errors in the same way as all the internal ones. So this leads me to the further conclusion of this being something to do with our interior environment at this time.

As unlikely as this seems, I did seal some flooring around that time with acrylic sealant so while it seems a long shot, maybe something in the chemical composition of this sealant was in the air and caused temporary failure of the interior humidity sensors.

We do have air quality sensors (MQ135) that detect various nasty gasses and I have seen these go 'off the scale' when using other building products, but they weren't showing anything odd during the time the humidity sensors were.

I'm going to attribute these odd stats to environmental factors then, but will report back if I see this again.

Cheers,

Robin