Always-on Ongoing Problems (reading high)



I’ve had Sense for a year now and while it’s improved a lot and made great strides in device detection the always-on reading doesn’t seem to have improved and constantly reads high. Many of my always-on devices like my server, modem/firewall, and stereo equipment have been moved onto HS110s to reduce the load but this has little impact on always-on. There are multiple periods of time each day where my power draw hits ~400 watts for an hour or two and if you subtract out the HS110 loads my always-on should be more like ~280 watts. There are a few times over the last month where always-on will drop to 280 watts for a day at most but more often then not it aggressively pushes back up to ~400. There has been discussion in the past about revisiting always-on and I was wondering if any changes have been made. This issue really throws off my monthly reports and makes my “unknown” power use look much better than it actually is. Any feedback from other users or Sense would be great, thank you.


Two thoughts:

  1. Sense has played with the Always On calculation a few times, but I believe the newest incarnation has stabilized since mid-Jan 2019. Here’s why I say that…
  1. Always On is a pure numeric calculation, but to assess it you have to look at detailed Power Meter views of your past 48 hours for Total Usage and for every device on a smartplug. Post those and maybe we can help diagnose. Or look at them yourself and come up with Always On values for each based on the near-lowest value over the past 48 hours. If there are dropout periods that total to over 15 min over the past 48 hours on your smartplugs, those could whack the subtraction factor from the smartplug to zero.
Always on/smart plugs

Did something change with always on very recently @kevin1? I’m talking about within the last week. My always on hovered just below 500 watts until a few days ago where it has almost doubled. There have not been any changes in our hone to account for this. I have a floor heater that is about 500 watts that show in other when it’s on. The heater is on almost all the time and I’m wondering if it’s being calculated twice (other and always on).


Nothing has changed in the past week.

For both of these cases, data outages may have a part to play, but I think it’d be worth reaching out to Support as they can peak behind the scenes at your Always On figure. With HS110s though, in @senseinaz’s case, that wattage should be pretty spot on. Idle wattages can complicate things though as well. Support should be able to help parse out what’s happening here.

1 Like

I redid my plot, just to see if anything was different than previously. See below:

At first I thought something was amiss because I had a bunch of very low readings the past couple days, one even negative ! But then I took a look at my past 48 hours and discovered a tiny band of time when my house was using less power than I had ever seen it use (excluding dropouts), 223W (below).

That little patch will affect my Always On until tomorrow AM, unless another low patch comes along. My mission now is to figure out if one or more of my devices on smartplugs has gone quiet, plus why that little low patch was there.

Morale of the story is that you should look closely at your past 48 hours if you think something is funny with Always On. And if it still doesn’t make sense, contact support.

1 Like


After a little more investigation, it looks like my Always On anomalies are caused by real things…

  1. As I posted earlier, I hit a record low power usage point from about 5AM to about 6AM Tues morning. I’m still not sure why - maybe there was some kind of weird brownout. My Sense and networking gear are on a UPS so they kept chugging along.

  2. But around 6:30AM something started taking many of my devices on my IoT Network offline, including my Sense.

That left gaps in many of my biggest smartplugs Always On consumers.

Those half hour gaps dropped my Always On for these two smartplugs to close to zero from somewhere in the neighborhood of 90W. All around the house, smartplugs could have shed about 190W of measured smartplug-based Always On due to the gap.

That’s why I saw a set of Always on transitions that looked like this…

Unpacking a little further, there were two events that changed Always On (AO) dramatically on Tues morning:

  • About 5:45 AM, about 1/2 hour into the new record low Total Usage, Always On was driven negative because the sum Total Usage (212W) - smartplug/Hue AO (190W) - Any other identified devices went negative.

  • About 7:30 AM, about 1/2 hour after the start of the smartplug gaps, smartplug AO dropped from 190W to 0W, and now the sum Total Usage (212W) - smartplug/Hue AO (0W) - Any other identified devices goes positive again.

Expecting further negative Always On weirdness on Thurs AM when the gaps are exhausted by the smartplug AO calculation and the smartplug AO starts to climb again. A lot depends on whether I see any more low periods of Total Usage.


Gonna add an epilog to this story. A few days later, things are back to a more typical Always On level of 205W. Even though I can’t see the internals of the Always On calculation, I can see that even short disruptions on the low side, in Total Usage or smartplug consumption, take about 48 hours after the disruption is removed to work themselves out of the system. Here’s the very left-most end of my Always On graph above, in much greater time detail:

You can see the two 48 hour cycles it takes for Sense to work the data gap in smartplug data, then a series of record lows in Total Usage, out of the calculation.


Ryan can we add a feature to the smartplug integration to designate the device is always-on?


I have also asked about this in the past. A checkbox on the device manage tab to indicate always on and perhaps a specific color bubble for any device with the check box ticked.

1 Like

We have some Always On revisions scheduled throughout the year, so keep watching.


Always on has been aggravating for me, especially the last couple weeks. Although I know it use a 48 hour average, it still shows much higher than it should. I’ve confirmed this be taking 2 day blocks and moving a kill-a-watt around. I’ll show over 500 watts the majority of the time and it will be mixed in the evening with over 900 watts. I’ve been working hard to get this number to accurately reflect what is always on. It has never been right. By my calculations with other tools, we should be at 280. I think always on and other are double counted at times.


I had an issue after adding a few new smart plugs, I opened a ticket with support. I had to express why it was wrong 2-3 times. Finally support agreed, and the issue magically resolved itself :man_shrugging:

Maybe open a ticket with your findings?

1 Like

How do you know that you’re using the Kill-A-Watt on everything that is actually on consistently during .a 48h period?


I second that question. I’ve done an exhaustive measurements of every always-on item I can think off (with Kill-a-Watt, smart plugs, or turning off breakers and then back on one by one, waiting a little for stabilization – driving my wife crazy in the process ;-). And I get to 178 W, but my real always on seems to be 233 W. Not sure what I am missing. I am sure I’ll figure it out with time, but it can be tricky.


By going room to room and using anywhere there is a 120 receptacle with something plugged in. There are only a couple I cannot get to like a refrigerator but I can turn it off from the front and it powers it down completely.
If not using the kill-a-watt I ha e unplugged things for a day or two. Right now I’m showing 514 always on and it stays right there until evening. Not all but about half of evenings will jump to over 900 or more.


It is possible that Wi-Fi issues you’ve been having (or have you solved those?) can cause issues with Always On, forcing the calculation to restart.

Otherwise, there really isn’t a reason that it would be so inaccurate. If you look at your whole-home Power Meter, what do you see for the baseline wattage, at the lowest point? That should be pretty close to the Always On figure that Sense is calculating. It might not be perfect, but it certainly won’t be hundreds of watts off.

Here’s my home from the last two days. My calculated Always On currently is 53W.


The last 24 is a terrible time for me to use as yesterday was catch up on laundry which had multiple heating elements running all day.
My always on is still at 406 right now.
Look at you show off those numbers, I’ll never get that low.
I’ll go ahead and show the last 24 and the lowest point I found was a bit over 800


What about a view of the previous 24 hours. I can see what look to be a couple short dropouts in the past 24. But Always On goes back even further.


Here is the previous 24.
I think always on has to ha e some of other in it or I just really don’t understand the averaging.
What I’m thinking my always on should have
Dish network dvr on standby 45
Clocks for fridge, oven and microwave
Floor fan 45 watts
4 usb wall chargers charging iPhones
CCTV dvr and monitor 70 watts


That low level li e you mentioned wouldn’t be always on the way I see it because it includes what is curroon also.