Another strange unknown device

I have come to realize that it is tough to get others to help figure something out but here goes anyway.
Let’s start with the basics.
This item showed up as a TV yesterday. It isn’t a TV.
It is using 220-ish watts. About what a big screen draws.
It is not any of these:
Coffee pot, iron, space heater, computer, furnace, well pump, etc.
And strangely, yesterday it was off from midnight till 5 AM, but today it came on at midnight and went for two hours. Then it stayed off until about 9 AM.
So kind of the exact opposite behaviour.
Really weird.
I want to find it.

can you post pictures of the wave form? that will help determine if it has a motor spike or a heater looking draw. Could be a refrigerator or freezer…hard to tell yet.

Here is it coming on at 9 AM, stayed completely flat till 10:06 AM.

are those images from the device’s info page of past usage? cuz if so, they’re prob not 100% right lol. esp since the device is still kinda new to Sense. I would scroll back through the actual meter usage at those times and look at the spike there and screen shot that instead; that’s a true view of the usage spike, and not what Sense “thinks” the usage was, just FYI.

I say that because Sense isn’t 100% right on device identification reporting; sometimes it thinks my EV charging usage has dropped to 600W when I know for a fact that the usage is pulling 3300W, per my OpenEVSE, which has real time usage available. So if I go to the EV device charging usage in Sense, it will show (incorrectly) the 600W usage and not the true 3300W usage.

I am a little confused. It came on at 9:20:38 AM. If I go to all usage at that time you can’t even see a blip, it is only 220 watts out of the thousand of kW I am using at that time (geothermal heating).

A Sense-detected Device Power Meter is always showing a guess.

A good indication of that is a perfectly flat wattage plateau like you have in this device detection: 221W. That is a time-averaged usage that Sense has determined.

The algorithm would be looking for an on/off signature and can sometimes miss either or both.

Some things to watch for, consider, and post info on:

  • How long has the newly detected device been detected for?

  • What stats is Sense showing for it?

  • What has Sense named it? This is often the best hint as to the nature of the device rather than decrypting the waveform.

  • Do you see the detected device in the Mains Power Meter? More often than not this will reveal that the device is a component of something more complex.

  1. It was detected about 36 hours ago.
  2. Don’t understand the what stat question.
  3. Sense named it Samsung TV (it’s not, my Samsung TV is on a Kasa plug)
  4. Don’t understand this question either.

This is the main power meter right now.

See the last point in my previous post but to wrap your head around that consider an idealized single device of, say, 1000W that consists of TWO 500W heating elements that come on at different times … and more often than not together. What would you expect to see and how would you expect Sense to deal with what it sees?

In the real world a hotplate with multiple elements is very much like this.

This is not what you have, but the component signatures (I suspect) of a somewhat complex device are being picked up by Sense.

Given #3 I would correlate the power meter for the Samsung TV on the Kasa plug with what you are seeing for what Sense thinks is “Samsung TV”. Then tell Sense that the Samsung TV is plugged into the Kasa.

Regarding #4: The Mains Power Meter (your overall power) is something you can trust to be showing you actual electrical usage. A given device power meter can only really be trusted if it’s for a smart plug.

Here are some TV stats, an LG OLED connected to an HS110 smart plug. This is available under a Device assuming it’s been detected for long enough.

Your post has prompted me to realize we need more TVs in the Community Device Library. I’ll post mine soon.

If Sense decided it is a Samsung TV, then it probably established that via NDI. That would mean it used your Samsung smart TV networking traffic to correlate power usage with TV activity… More on NDI here.

1 Like

I listed my devices that I own in my Sense app ahead of time. That is why I think it said Samsung TV, because it knew I had one.
At 9:20 this morning, when mystery device came on, the Samung TV was de-energized via my Kasa plug. No power going to it at all.
As you can see Sense is currently showing 174 watts of usage by my Samsung TV (via Kasa smart strip) and it also is showing 221 watts of the mystery device. Screenshot 2020-03-06 12.34.00

Can you post the Stats for “Samsung 4K TV” and “unknown” (if there are any)?

Just to update this situation, this unknown device that was almost on constantly yesterday, hasn’t shown up for over nine hours until just now.
Today, when my wife turned on the AV components it came on. But it shows my Samsung TV that is on my Kasa strip and the unknown device.
Could sense be seeing the TV and showing two signatures? If so do I merge it? Will that say my Samsung TV is using 400+ watts when I don’t believe it is.
I can tell you this, if I turn off the TV using the smart strip, the unknown device doesn’t go away.

I’ll say this once again. Sense likely “found” your smart TV via NDI, while it was already on the smart plug. That would mean Sense is double counting your usage through the two different devices. The Unknown device won’t likely see an off detection when you turn off the smart plug because Sense doesn’t “see” the true smart TV off cycle which includes some networking traffic as the TV tells the network it is going to sleep. That’s why it doesn’t go off in the detected device.

You have three options.

  • Delete the detected device. But it may come back now that it has been discovered.
  • Merge the two devices by going to the smart plug device (Samsung 4K TV) and using the Manage tab to modify “What’s plugged in ?” for the smart plug. You should be able to add the newly detected device to the smart plug device. The smart plug measurements will supersede the detections.
  • If you have high confidence in the detections (which it seems like you don’t), you can also remove the smart plug and repurpose it, using just the native detection to spot the TV. Beware - repurposing a smart plug needs some additional steps, including a reset in the Kasa app, to separate it’s old history from it’s new purpose.
1 Like

Thanks for the suggestions.
I think the biggest hurdle for new users is the false expectation that if you go in and trip a breaker you will immediately see the device go off.
I understand what you are saying. When I energize the TV and turn it on, Sense sees the normal start up pattern. When I say “Alexa, turn off TV” the TV is de-energized and doesn’t display it’s “fingerprint”.
I am going to try your suggestion with managing it inside the Sense app.


You are right. For many devices, the breaker or outside switch on-signatures and off-signatures look different than the “normal operation” on and off signatures. And if done from the breaker, multiple on or off-signatures on the same circuit might also obscure one another. That’s one reason most of the Sense documentation says “look at the Power Meter, not the Bubbles” when playing with the breakers to identify a mystery device.

1 Like

I’ve said it before and I’ll say it again: LAZY WINS!

1 Like

4 posts were split to a new topic: Sense New User Video