Very large negative power values received a few of times a day

I have local MQTT working fine, but occasionally get an electricity power value of -8388.608 kW. Is anyone else seeing something similar?

This happens a few times a day and messes up my graphs. It also breaks the graphs in the Bright app. I can just ignore these values so they don't get added to my local database, but was wondering what causes it.

Whenever one of these bad readings comes in, the cumulative import and export values are null when they are normally a number (even if it is 0.000). The signal from the meter seems ok, LQI is always between 240 and 248 and RSSI -38 or -39.


{"electricitymeter":{"timestamp":"2022-10-09T01:00:38Z","energy":{"export":{"cumulative":null,"units":"kWh"},"import":{"cumulative":null,"day":0.446,"week":31.237,"month":39.994,"units":"kWh","mpan":"MPAN SSWG","supplier":"---","price":{"unitrate":0.33020,"standingcharge":0.44120}}},"power":{"value":-8388.608,"units":"kW"}}}` from `glow/1234935253D4/SENSOR/electricitymeter


{"electricitymeter":{"timestamp":"2022-10-09T01:00:48Z","energy":{"export":{"cumulative":0.000,"units":"kWh"},"import":{"cumulative":8999.761,"day":0.450,"week":31.241,"month":39.998,"units":"kWh","mpan":"MPAN SSWG","supplier":"---","price":{"unitrate":0.33020,"standingcharge":0.44120}}},"power":{"value":0.194,"units":"kW"}}}` from `glow/1234935253D4/SENSOR/electricitymeter


  • I have still been getting this bad power reading once or twice a day, but everything else is working fine so I finally decided to just ignore it :) I use telegraf to insert the data into influxdb, so to stop it messing up my graphs I used the json_v2 data_format and set the GJSON path to only return data if the power value is over -8000.

      servers = [ "tcp://"]
      topics = ["glow/+/SENSOR/electricitymeter"]
      data_format = "json_v2"
          path = "[@this].#(electricitymeter.power.value > -8000).@this"

    telegraf does log an error message when the path ends up empty, but that's better than inserting bad data to the database.

  • I've had one of these readings so far, similar magnitude negative value. I assumed it was an erroneous power calculation from the meter as it happened at the exact point of the fridge inrush current and just deleted the value. Interesting you spotted the null values, this may suggest an actual bug.

  • And another (~every 48 hours), this time without any inrush current. My cumulative energy values are also missing from that record like yours.

  • Since ignoring any value < -8000 I haven't had any more problems with that :)

    My meter is a SMETS1 Honeywell one.

  • Hi all,

    A new customer also reported this issue, Clive replied to them and said the reading is not a spike per se but corresponds to all F hex and is a meter saying 'I dunno'

    We filter these at the ingress - but everyone has requested that we send on mqtt the raw data warts and all - so, as @af7567 has done, you need to code around it.

    Please note that we find enrolled SMETS1 meters, and particularly Honeywell, have been quite problematic for us in many ways (in case you've got that combination).

Sign In or Register to comment.