Adding a Sense as Device Analysis tool

With a dedicated Sense on a single device you can accurately track that device’s usage and use Goals to set certain alerts. Detection is still relevant in so far as Sense is capable of identifying all the components of the device.

So if Sense knows, say, “I am only monitoring a Ground Source Heat Pump” one can assume that could lead to more fine-grained component detection and, more importantly, reliable usage alerts and failure detection. This is much the same as the potential for failure detection with smart plugs but with the dedicated Sense the analysis is founded in higher sample-rate ground truth data and so one would expect better confidence.

I have been using a dedicated Sense monitor on a hot water heater in this way.


  • Is there an adequate and willing user base to make this something that could be usefully employed?

  • If a monitor is physically configured as “device-dedicated”, could Sense benefit from adding the option to designate the monitor as such and treat the data accordingly? (If only for the Data Team to then know what they are looking at)

  • Can the Solar inputs be exploited in a similar way? e.g. If a second Sense monitor’s main CTs are dedicated to EV charging/chargers and the Solar CTs are dedicated to a PowerWall(s).

Note: Posting this in Data Analysis because this is more of a preliminary call driven by the need for certain data vs a feature request … but increased confidence for alerts and load shifting recommendations would be the obvious outcome of improved device recognition.

1 Like

It’s a good question. I’m guessing I would buy a couple more units if there was a Sense-supported device/circuit-only mode to tackle my 240V loads that are missed or inconsistent today. That said, I think there are a couple of downsides to the existing monitor for use as a probe.

  • too much hardware - the Sense monitor has specialized hardware (FPGA) to do it’s DSP magic, something that runs up the BOM cost, but isn’t needed in a single circuit/device mode.
  • too much software - there’s also a lot more software inside the monitor than needed to just report power usage.
  • challenging form factor - I would prefer a smaller package with inline current monitoring, rather than CTs, for power endpoint installation, rather than breaker box install. My breaker boxes are already crowded. The inline approach also has the benefit of working with much cheaper off-the-shelf power measurement chips.

Agreed, to a certain extent, but I’m talking more about the interim analysis aspects vs using it as an inline power usage monitor. Less of an overkill smart plug (for 240V) and more of a Power User tool.


  • Smart plugs (of any form) are anathema to a “proper” dis-aggregation device (Sense).

  • To get to a smart-plug-free system more quickly there needs to be reliable data that Sense can tap in to for 240V devices (and 120V for that matter!).

  • Existing solutions for 240V (and non-Kasa/Wemo 120V!) outside of using a dedicated Sense are letting good data slip away.

  • Exploiting the 4 existing CTs for 2,3 or 4 unique circuits does not preclude detection (on the Mains at least) and approaches or undercuts the cost of other solutions.

  • Multiple similar devices (e.g. EV chargers; fridges; freezers) on a single circuit or select multiple circuits could still exploit detection hardware (overkill).

  • A Sense, though “installed” in a panel is usually easily reconfigured to “test” particular circuits/devices. Some of us use an HS110 in this way.

*Sense modifications to the UI/data potentially wouldn’t be necessary if there was a device or account nomenclature. Maybe.