Another Mystery Device - 213W

Hey everyone. New here and just got Sense installed last week from the company about to install my solar. This is a fascinating device, but I have a very frustrating problem in that I cannot find a device that cycles on roughly every few mins for about 1m20s and pulls exactly 213w each time. What makes this even more frustrating is i’ve turned off every breaker in my house and it still cycles on and off (my sense is installed in my meter base panel). If anyone has any ideas, I’d gladly welcome them!

Is the breaker that Sense is on dedicated only to Sense? If you’ve flipped off everything else, then that should be the only breaker turned on (and any usage in that panel would be coming from that breaker). Are you certain that nothing else is on that breaker? Does your networking equipment sit on a UPS (so that its breaker could be off), because you would obviously need both Sense and your wireless network+internet connectivity to be working to see anything in the app.

FWIW, it wouldn’t be Sense itself, as that is far too much consumption.

Yes, Sense is on it’s own dedicated breaker. My network gear is on a UPS which is how I was able to turn the house off and still monitor. I’ve toggled all my breakers 3 or 4 times trying to find it and nothing makes it go off. This is why it’s driving me mad.

Are you turning off the breakers when you see this mysterious usage? Or are you seeing the mysterious usage appear even after all breakers are off?

If the former, that’s due to how Sense determines whether a detected device is on: it looks for specific on/off signatures. Flipping a breaker typically results in a different off signature than the device turning off normally, so Sense doesn’t see that device as having turned off. Since you’ve shut off the rest of the house during that time, regardless of whether it sees the off signature (to turn off the device bubble), you would see the overall usage drop (probably to only what Sense itself is using: about 5W).

1 Like

Can you describe your install a bit? Where does power enter the house. Where is the Sense installed etc…
While device detection can be wonky, I’ve never heard of the main power meter being wrong.
So while already asked, are you 100% sure there aren’t any other breakers or circuits that are on a different panel or anything?
That description of the usage / wattage makes me think some sort of small electric strip heat. So a small engine block warmer, or rice cooker or something like that.

1 Like

Sorry my responses all have to be approved which is delaying my replies. Here’s my setup in detail:

Power enters the house at the meter base outside. Sense is installed in the meter panel on the outside of the house on a dedicated two-pole breaker. I only have one panel in my garage which has breakers for the entire home. I have no other sub panels.

Network-wise, Sense is connected via ethernet to my network closet that is all backed up via a UPS. This allows all of my network functions to fully operate in a full power outage which is how I was able to monitor Sense with all the breakers in the off position.

@ciphershort, as @ben suggested, if you are looking for a mystery device bubble to disappear when you flick the breaker off, that might not happen, even if you flick all the breakers off. If Sense sees a different kind of transition for a breaker turning off vs. the mystery devices normal off transition, then the bubble’s going to state up. Often, turning a breaker off, turns off multiple devices, making the off transition look different from the single device alone.

You should be set with post approval permissions moving forward

@ciphershort Can you take a screenshot of the live power meter when this happens while you have all the other breakers off?

@kevin1, can you split this to a new topic?

Thanks everyone for working with me on this. I’ll be able to kill power to the house again this afternoon and will see about leaving it off and getting screen shots of the live power meter. So what you all are saying is that when I’m watching the device-specific graph, turning off breakers may not actually affect the behavior of that specific device’s “voice”? That would definitely explain why I’m not seeing it die but I’ll give this a go and see what happens. Would it help if I left power off for a longer period of time?

1 Like

I couldn’t post pics yesterday but here’s some screenshots of the device specific screen and graph.

Hm, so this doesn’t look like a motor.

At first glance, this looks like a resistive heating element, but I’m curious as to what others here think.

Question for you: do you live in an apartment/townhouse/condo? I only ask because we’ve heard from customers in apartments that they’ve had their neighbor’s AC unit feed into their panel mistakenly. Here’s the story :slight_smile:

Sense first identified it as a heating element of some sort which I would agree by looking at it. No, I live in a single family dwelling.

1 Like

That’s correct. Turning a device off via a breaker can create a different off signature than the device’s normal switching mechanism. You should always look at the main realtime power meter when doing these sorts of tests. That data is “straight from the horse’s mouth” and doesn’t have any signal processing applied.

Eventually Sense will time-out a device that’s been on for longer than it expects, so if an off event is missed by the ML, the device will eventually be reset to its off state. But that could take many hours.

1 Like

Interesting. Thanks for explaining this behavior. I’ll definitely watch the main power meter when doing this and see what it shows.

1 Like

Not sure if this will help or not… But your chart that shows a short spike every few minutes looks similar to my Jura SuperAutomatic Expresso/Coffee maker… Except my spikes are 1200W every 90-120 seconds…

My coffee maker is ready to dispense “on demand” which means that it needs to be ready to output hot water at a moments notice… The appliance has a thermo-block that needs to be kept hot so that the water passing through gets heated in real-time… So every 90-120 seconds or so it gives a burst of 1200W to keep the thermo-block hot…

I only mention it because maybe you have an appliance with a similar use case, but at a lower wattage burst…

I have a Keurig K-Duo Plus coffee maker that I thought might be the culprit but I unplugged it for a night and it didn’t stop the cycle…

I’ve gone through my devices and pulled some screenshots of what it is not, and what does look similar. That signature looks very similar to my microwave and AC condenser. I initially thought a light I have that is 200w, but here is that:

The light stays exactly at 205 watts. It is some form of incandescent, but I haven’t investigated.

Then I thought about my toaster oven since talk is a resistive heating element. My toaster over is a very bouncy power meter.

The toaster oven also uses significantly more wattage than your mystery device.

I also looked at my kitchen fridge defroster:

It too has a very active power meter.

Next I thought about my Keurig since coffee makers were mentioned above. I don’t leave mine on, and when it is on, it’s on for 12 seconds at a time, then turns off, and on again about every 70 seconds.

My microwave actually has a very similar signature, but again, voltage is way off and I think you’d know if you microwave was running this frequently.

This is two, three minute reheats a minute apart.

I know motors were ruled out, but here is my circulator pump for radiant heat:

It looks like a motor and nothing like your device.

Sticking with motor, I looked at my AC compressor, and that stays super steady:

But it still has the surge at the start that is a tall tell sign of a motor.

Lastly was my basement fridge, which has no ice maker or fancy anything.

This again looks very much like a motor, so not your device.

Not sure if any of this helps, but maybe they’ll trigger some thought from someone else?

1 Like

Well I spent a bit of time yesterday doing some testing. I let the house sit with all the breakers off for a good 30 mins and the device did not show up so at least that is logical. I’m having a hard time getting sense to load data lately as it keeps timing out when loading device graphs so once that clears up I’ll have a better idea. From what it did load, while I was leaving individual breakers off for long periods of time, I’ve ruled out fridge, coffee maker, dishwasher, water heater, and several electronic devices.

It sounds like the gaps in data you’re seeing are related to the open API issues we’re investigating. You can follow along here: [Bug] - Network problems