How Accurate is Sense? Summary of Sense Accuracy Measurements vs Utilities

Have you compared Sense usage against your utility usage ? That’s the measurement this topic is focused on.

Detection, as several other topics highlight, is very much a Your Mileage May Vary Depending on Household Devices (YMMVDHD) kind of thing with a whole lot of pointers to things like smartplugs if you want to see what your outlets strips of electronics are doing.

You do have a larger group of than usual Always On, but I presume you have already explored the components that contribute to that ?

I am very pleased with Sense and it’s extremely accurate in my case. I have only had Sense since Dec 30 2022, already found 9 devices.

I have took pictures of my outside meter, went to work came home compared the results from meter to sense, first time showed my meter should be 13900 the meter was 13899, the next two times, I test, Sense said meter would be 13910 meter was in fact 13910, the next time Sense said meter would be 13938 meter was exactly 13938.

I take a picture of meter, add what sense says I used, add that total to the picture I took, so when I get home, I know what my meter will display.

1 Like

I just looked at 2022 annual usage and generation from Sense and our local utility bill. The total usage (11413 kwh) and solar generation (8772 kwh) are much higher from Sense than from the utility (usage 8583 kwh and generation 6096 kwh), but the net usage is reasonably close within 154 kwh (. Given the accuracy of Sense as noted in this discussion thread, any suggestions on why the large difference in the annual numbers? As near as I can tell, I’m comparing Jan to Dec 2022 in both cases. Also, the Enphase solar energy summary for 2022 shows 9200 kwh generated. Any suggestions or thoughts about these differences?

Thanks @blamb.kt ,
Interesting ! A bigger than 25% difference in usage and production numbers, though the net is within about 1% of your yearly metered total usage. I have a couple of questions that might help sort out:

  • Sounds like you have 2 meters from your utility , one for the house total usage and one for the solar production ? Maybe you can capture a picture of the meter, just to see what kind they are ? (fyi-my house has a single net meter, so my utility has no idea of how much solar I produce, just the net , plus how much I export to the grid).

  • Do you have a way of pulling more detailed data from your utility ? This is a great way to remove all the possible “measurement interval issues”. For me, I can download 15 min usage data from my utilitity’s website - that’s very useful for sorting out differences.

  • Do you have single, split-phase power, like most residential in the US, or 3-phase (some large multi-unit town-house developments and condos use this).

  • A few possible explanations are that:

    • You aren’t lining up intervals correctly
    • Your utility isn’t reading in line with the times stated
    • Your meters are reading something different than Sense (requires deep discussion on Real Power vs. Apparent Power, phases and power factors).
1 Like

Thanks for the suggestions. We’re traveling this month, so I’ll have to wait til we’re back to check on the meter type. It was installed new when we put in the solar panels in 2021.


I solved the issue of utility vs Sense. The utility data were for power supplied FROM the grid (usage) and power supplied TO the grid (generation). These are different that the actual usage and generation measured by Sense due to the mismatch between when solar was available vs when we were using power or generating solar power. The good news is that the net power measured by Sense and reported by the utility agreed to within 3% and the total annual solar generation measured by Sense and by the solar panel system (Enphase) agreed to within 6%. If we didn’t have Sense we wouldn’t really know our actual usage and solar generation. Kudo’s to Sense!

1 Like

@blamb.kt , great sleuthing ! I 'm going to add your experience to the table above, because it shows yet another thing to be aware of (from grid and to grid meters) when trying to do an accuracy comparison.

You have nailed one of the great things about Sense for a solar user - Sense breaks out the data into Total Consumption and Solar Generation, the two most meaningful values for most user. Most users only see Net Usage (in my case) or From Grid and To Grid (your case) from their utility.

ps: My experience is that my Sense net usage is within about 1% of my utility, but my Enphase is optimistic to the tune of 3%.

pss: The closeness of the Net of your utility numbers to the Sense’s Net (154) should have been a clue. From Grid - To Grid = Total Usage - Solar Production

1 Like

Did new reading today, Sense said it should be 14249 it was 14250, So to me very accurate.


I have three months of Sense data to compare with my electric bill since moving into my first house. So far, there’s a huge descripancy. Average usage on my electric bill for the past three months has been 21, 19, then 21 kWh/day. Compare that to the average usage reported by my Sense meter over the same three months- 10.8, 10.6 and 11.7 kWh/day so far in May. Either my Sense is signicantly underreporting my usage, or Chicopee Electric Light is charging me for a lot more electricity than I’m using. Either way, I’d like to get to the bottom of it.

1 Like

Welcome, @o498957 . Congratulations on having your own house! The big discrepancy you report is worth following up on. I suggest:

  • Make sure the CT clamps are tightly closed. CT clamps are the two crab-claw shaped donuts with wire leads which allow Sense to measure energy usage. If your CT clamps are not tightly closed, they take incorrect readings. Sometimes clamps can be jiggled open by the act of closing the electrical panel back up, even if they were closed when you last saw them with your eyes, so double check clearances if you are comfortable working inside your panel box.
  • Is there a sub-panel that is connected to your meter but not included in what the CT clamps measure? Some users post photographs of their panel box here to allow the community to diagnose issues, which is surprisingly effective. You’ll automatically be allowed to post images after a 3 day waiting period for new users, beginning with your first post on the forum.
1 Like

@o498957 , I’m betting on one of the same two scenarios that @jefflayman is suggesting as the cause - either incompletely closed CT clamps(s) (sensors) or a hidden “secret” second subpanel or split mains bus. Both have been seen on the community forum before.

Both meters are fully clamped on the power mains. Checked clearance when covering the panel back up, does not appear to be interfering. The clamps look a bit dirty, but they’re definitely closed.

100% sure there’s no subpanel. I spent a day mapping the breakers to every light and outlet in my home with wireless circuit breaker finders, I can account for all breakers and corresponding lights/outlets on the main panel. Also paid for a comprehensive professional inspection before finalizing the home purchase, was present on the walkthrough, and spent most of it asking electrical questions, so if it existed, it would be identified.

1 Like

@o498957, you’re right - all looks good. The next thing to do is to look at a more detailed comparison between Chicopee’s kWh numbers and Sense’s. I tend to recommend looking at daily or even hourly data if your utility offers it, because that generally helps more readily pinpoint where discrepancies might be stemming from.

Here’s a step-by-step approach from the quickest and easiest to the most detailed.

More thoughts on your installation:

  1. I helped my brother install his SENSE unit. For some unknown reason the past homeowner wired the HVAC directly to the outside meter. The only location to place the CT clamps so that he could measure all his power was above the meter on the incoming feeds. Check your outside installation and see if you have a similar issue where someone may have tapped the service cables before the breaker panel. You will need a professional electrician to install the CT clamps.

Could you post some pictures of your electric service as it enters and leaves the electric meter.

  1. I believe you when say that your CT clamps are fully closed, but if you want to make certain, then tape the ends closed. I used masking tape to ensure my clamps remained closed.

  2. You have two loose unterminated wires (a black and a red) on the left and 2 unterminated wires on the right (a black and a white) with copper showing. Please add a wire nut or tape the ends for safety. You do not want those wires coming in contact with any live part of the panel. This is a safety issue.

  3. I don’t think you have the space, but I found when I was able to center and place my CT clamps at a 90-degree angle to the incoming wires, that I got a slightly closer reading to the power company numbers. I documented my experience about 5 years ago in this forum.


The most detailed data from my electric company is for 30 day billing periods. I can check my meter every night at midnight to get daily totals, and theoretically could monitor over shorter time intervals to test under a specific, focused group of usage variables, cutting off power except for a small number of steady, known power loads to compare the meter and Sense data to the predicted usage. Would require some planning and precise execution on my end, but is doable.

1 Like

Thanks for the tip on wire caps for those unterminated wires. They’re routed to an old 240v through the wall a/c that’s no longer in use. I freed up space by removing the breaker when I moved in, but neglected to cap the wire ends. Will do.

I believe the incorrect Sense monitor readings issue is solved. Sense support suggested that my monitor wasn’t getting solid wifi connection, and wifi dropouts were the cause. That was not the case. The router is about as close as possible, 15 feet max, it’s a decent router, Asus Rt-ac86u, and my wifi connection test in the app showed no issues, and I saw no time gaps in monitoring history. What worked for me was doing a data reset. Registering the Sense monitor from scratch, letting it complete initial tests, and immediately I can see a huge difference. While before, it was reporting 300-400 watts total with two window A/Cs running, the live data it’s reporting now with the same approximate electrical load appears to be right in the ballpark of where it should be, somewhere between 1000-1500 watts, and moreso when the dishwasher, washing machine and/or dryer is running. Don’t know why it was underreporting significantly before, but alls well that ends well, and seems to be reporting correctly now.


@o498957 , thanks for the follow-up. Will be interesting to see how the next full-month / billing cycle comparison goes, after the data reset. If you see good accuracy, we can add to the table under “Accurate after data reset”

I need to dig up all these commends and references, I see the “Other” bubble on mine with a lot of consumption for stuff that should have been already detected like portable AC’s, so it looks like in normal/working cases we shouldn’t see really the other bubble showing up all the time correct? …pardon my ignorance :slight_smile:

@mbicca ,

First off, this topic is mostly about accuracy of Sense total usage vs. the measurement shown by your utility. Sense has been shown to be very accurate, at least when users are comparing apples-to-apples. There have been a number of cases where Sense seemed off, but the initial user comparison was missing a key factor, that caused the discrepancy (see the table for more details).

When it comes to detection of devices like AC units, the answer is that “it depends” on many factors:

  • How long has Sense been operating ?
  • How long has the AC unit been operating ? Has Sense seen enough cycles ?
  • Is the AC unit producing clear on/off cycles ? Single stage simple AC units do, variable speed AC units and mini-splits do not.
  • Do you have a lot of noise (other devices switching on and off in your house) ?
    All those and other factors play into how easy or hard it is for Sense to detect your AC and other devices. The best thing you can do is to start a separate Topic under detection and post some very zoomed in views from the Sense phone/tablet app showing the AC turning on and off.

@o498957 ,
Have you had any luck sorting out your Sense accuracy vs utility ?