My Frigidaire freezer was not being detected and it has very similar cycles and wattage(150±watts) as my Samsung fridge. I bought a sense HS110 to help me see it’s energy useage.
What I discovered is the freezer has a 1watt vampire load.
Based on some other comments I’ve read in various threads. This ‘always on’ device has a little to no chance of being naturally detected by sense.
Even though this device will never cycle on/off of its own accord (due to that 1watt Vampload) It still has, at least to my view, a distinguishable power surge and drop with a regular pattern that can be otherwise construed as power on and power off.
Basically instead of a 0watt baseline it has a 1watt baseline.
I am curious what are the chances of my freezer and or other devices ever getting a natural sense detection?
Sense looks for (mostly) 1 sec and shorter power transitions, not a drop or rise from 0W, for on and off signature detection. So that 1W baseline Always On shouldn’t affect Sense’s ability to detect the fridge. You can learn more about detection from this video, that answers lots of questions about how it works.
The Sense (founder Mike) interview @kevin1 points to is great but perhaps a little hard to wrap one’s head around the details and how they might apply to what most Sense users consider “easy” detection: “I can see it so why can’t Sense?”.
A simple device scenario I’ve used in the past comes to mind again for this question: Imagine two identical 100 watt light bulbs (remember those!) being switched on an off pretty-much randomly throughout the day.
In theory a clean 100 watt jump up or down in the overall electrical waveform is a clear indication to Sense of “it’s a 100 watt light bulb going on or off”. That seems straightforward enough.
Now you need to ask: But which bulb are we talking about??
Now suppose one of those bulbs is always on.
If you apply this logic to the vampire load aspect (“always on”) of many devices you are presented with a similar detection scenario.
The most interesting thing for me in Mike’s video was the added “uniqueness” constraint. Sense has to be able to see a consistent, repeatable on and off signature plus it has to be unique. Given the “same sized lightbulb” scenario you raise, or the “same model garage door opener” conundrum I have in my house, it’s probably impossible to always get that “uniqueness” constraint right. I could see the “uniqueness” constraint often being the holdup, in what would seem to be a simple detection.
The freezer in my example is much like the fridge. Perhaps it is struggling to find the unique mess between the two. But then why did it detect the fridge?
The freezer cycles no less than 20 times a day. Thus it is pretty consistent and regular. And through all parts of the day. It is also a motor cycle.
In the 2 bulbs example the devices are clearly not unique but if one is on all the time then the other becomes unique because it’s behavior is unique: it comes on and off!
In the real world this is an uncommon situation. A couple of similar fridges/freezers or AC units are easily conflated. Here is where a human (not cheating by knowing which device is on or off) can empathize with the Sense algorithm by looking at the waveform.
