There are quite a few options for monitoring your whole house water or power usage, but i’m not thrilled about adding a bunch of extra hardware to my breaker panel.

I’ve known for a while that the city reads my utility meters by driving by, but had assumed this radio technology was some kind of encrypted proprietary data that I couldn’t access. However it turns out most utility meters in the US broadcast their readings in a completely unencrypted format every 15-60 seconds and with the recent advances in Software Defined Radio’s it doesn’t require anything particularly complex to pick it up.

Hardware Installation

I bought a Nooelec NESDR Smart v4 for $40 and then plugged it straight into the USB port on my Synology NAS. It does get quite hot in regular operation, but that doesn’t seem to affect its performance.

SDR Installed

Database setup

I was then able to make a meters bucket on my InfluxDB database

curl 'http://nas-server:8086/query?pretty=true' --data-urlencode "q=CREATE database meters"

Software-Defined Radio Stack

The tools I’ll need to capture the readings are as follows:

  • rtltcp – This is a library that allows a software defined USB radio to be accessed by TCP over the network
  • rtlamr – This tool connects to a radio being exposed by rtltcp and decodes and exports meter readings
  • rtlamr-connect – This takes the readings from rtlamr and writes them into InfluxDB

Fortunately, I found a handy docker image that rolls all that together into one:

docker \                                         
  run \
  -d \
  --privileged \  # Normally docker container can't access USB
  -v /dev/bus/usb:/dev/bus/usb \  # This maps the whole USB bus to this docker container, probably overkill
  --name rtlamr \ # Here's a name
  -e COLLECT_INFLUXDB_USER=user \ # Add your InfluxDB credentials here
  -e COLLECT_INFLUXDB_HOSTNAME=http://nas-server:8086 \
  -e COLLECT_INFLUXDB_DATABASE=meters \  # Here's the bucket we're writing to, don't forget to create it first
  -e RTLAMR_FORMAT=json \  # This is very important
  -e RTLAMR_MSGTYPE=scm,r900 \  # These two message types grab my meters, you can use "all" if unsure
  shoginn/rtlamr-collect:latest   # The image to run

Let’s let it run for a while and look at the data in influx:

from(bucket: "meters")
  |> range(start: dashboardTime)
  |> filter(fn: (r) => r._measurement == "rtlamr" and (r._field == "consumption"))


Well that seems to work, all in all I can see pretty reliable transmissions from about 50 utility meters. Figuring out which endpoint_id belongs to my meters required a bit of detective work but I found something that mostly corresponded to my endpoint on both my electricity and water meters.

The next challenge will be to use InfluxDB to get those numbers into a usable format for monitoring.