Why am I always having issues with bubble math and devices stuck on

Since I’ve had my sense meter in 2019 every so often it will show a device is on that’s off. It will stay off until I shut the meter off and power it back in. The reporting it’s off does nothing. I will have days where it will show a device was 400% of my daily usage.

I also have issues with it showing bubbles showing a higher wattage for that device than normal. Like a 110v device showing 4500w but then my total wattage say 1200w.

In the one photo my HVAC fan is 1779w, and the AC (compressor) is 2484w … but my total is 1447w. If I would add my % usage total, it never equals 100%. Obviously once a bubble gets stuck on, it’s way over, but even without the stuck on device you can see my #2 and #3 are adding up to 95.6%

Is there a way to recalibrate this?




1 Like

Hi @ccook - have you already reached out to Support about this issue? I’d like to check in on your ticket if so.

I had a few tickets a while back… same type issues. I had someone from sense look thru my data and say they couldnt find anything. I would guess the last one was in 2020. At one point …If I had a device that would keep getting “stuck on” , I would delete the device. But it never seems to relearn the devices that I delete.

Here is my current bubbles.

It would be great if you could reiterate what you’ve overviewed into a new support request here: https://help.sense.com/hc/en-us/requests/new. I can follow up with the support lead following your case later today and make sure they take a deeper look into this.

I have completed a new support request.

I have documented this multiple times.

@JustinAtSense have you had an opportunity to look at this? I think my request number is 257815

Is there a re-calibration step I can do? I think every device I have … eventually get stuck on, but the devices on are still over reporting the total wattage. The AC is the easiest to show since the draw is so high. But even things like my water dispenser sticks on, but sever devices seem screwed up. Her is my gas furnace which detected as shows it’s alway 841w to run the HVAC Med fan. Randomly in Aug this year… it just started showing that’s it’s coming on at ~100w

Then here is what I mean by it doesn’t add into the total for that day



I’m not as worried as the wrong device shows up … such as a toaster for a hairdryer. But it currently seems very inaccurate lately. If I could import my total usage/ solar … I would just do a full reset.

Hey @ccook, thanks for the reminder. I looked in at your ticket and saw the response from Support. Unfortunately, I don’t have alternative advice beyond what the Support agent shared:

  1. Give the devices a few weeks to see if accuracy improves
  2. If you’re still seeing inaccuracies, delete the devices and wait for a redetection.

As the support agent mentioned, based on the type of detections we can’t make adjustments to the device models. I know this is less than ideal, but let me know if I can send you over a smart plug (not sure if any of the devices you mentioned are on a plug instead of hardwired) to aid with detection here.

If I delete every device I have, will it relearn them all?

If I would use a smart plug. Is that something that I could use on “Device A” for several weeks for sense to learn Device A better then move it to “Device B” to do the same… and Device A would still work correctly?

It’s not guaranteed, but if it redetects the device it should be a more accurate model.

Smart plugs don’t influence native detection directly, so this wouldn’t work in the way that you’ve outlined. I’ve heard some users have had a device detected after leaving it on a smart plug for a while, but that’s pretty circumstantial/unrelated. This would just be able to get the devices into the app for visibility (as long as they’re connected to the smart plug.)

1 Like

Bottom line as I’ve unfortunately found - Bubbles are not accurate and lagging.

They are not to be trusted.

There seems to be no momentum to correct either.

Not true. There are some projects going on, in testing, that will hopefully address some of these scenarios/issues with offline and real-time event dissociation.

1 Like

@JustinAtSense I think Im ready to try a smart plug.

1 Like

Go ahead and PM me your shipping information.