Improvment of deleting wrong very high statistics value

[genemars]

Hi @dmayard I'm experiencing same issue sometimes.... do you know if is it due to bad Z-Wave communication? If so, I wonder why is the checksum valid for those bad values.

[dmayard]

I think perhaps it's due to bad ZWave communication. But in that case why the checksum would be ok ? I look at some logs already, and in the case of error, the message is different. In my case it's principaly with AEONLabs HEM V2. Meter and KilowattHour messages have 2 parts of values, I analyze that it's the precedent value and the new one, in that case the value is always good. But with the error messages, the new value is specified and the precedent value is zero. I can send you some samples.

Cheers Dani

[genemars]

just trying to figure out if there's a way to discard those bad messages.

[dmayard]

Here is one error trace between good messages :+1: SPI > 01 14 00 04 00 0F 0E 32 02 21 64 00 00 1B 6F 01 2C 00 00 1B 47 9E (7,023 Kw)

SPI > 01 14 00 04 00 0F 0E 32 02 21 64 00 03 8D 53 00 00 00 00 00 00 46 (232,787 kw) --> Error

SPI > 01 14 00 04 00 0F 0E 32 02 21 64 00 00 2A D4 01 28 00 00 2A 09 6F (10,964 Kw) SPI > 01 14 00 04 00 0F 0E 32 02 21 64 00 00 2D 3D 01 24 00 00 2B A7 22 (11.581 Kw)

[dmayard]

It is easy to discard messages with very high value. But in that case 232 Kw is not a very high value. Some times I get some wrong temperature value so.

[dmayard]

And some Sensor with Humidity for exemple on one that doesn't have these capability. But I don't have that logs

[genemars]

Well, let's leave it as is for now. Discarding bigger values is something I've already tried, but as you said it's not a reliable approach.