Tesla detected, but incorrect power

So today my Tesla was finally recognized and shows up in sense while it was charging. The tesla app, my Wallbox charger are both telling me charging is happening at 11 kW. Sense’s bubble display is showing total power consumption of 11,550 W. Without the car charging I typically have a number of about 500 there, so I am figuring indeed 11 kW for the car.
However, the Tesla bubble consistently shows 9, 618, so about 1,400 missing. I found this in the “other” bubble. This is wrong, but also odd. Sense is somehow concluding that the Tesla cannot be responsible for more than 9,618.

Then I got this thought. My charger is hardwired and can go as high as 48A (and it does on a 60A circuit). What if Sense assumed the max is 40 A, which is true for many home chargers? 9618/4048=11,541! Another way of looking at it, 9618/40 = 240V which is sort of the max voltage on a circuit. My Tesla app is currently reporting 230V so true current might be 9618/240230=9217. Now apply up conversion to 48A and come to 9217/40*48=11,060. Add my “typical” just under 500 usage and it almost matches up perfectly.

Hmm. I am tentatively thinking there is a bug there assuming max current is 40A. Perhaps sense should introduce some kind of setting for devices identified as EV that allows setting the charging current ceiling so it doesn’t have to make this assumption? Alternatively it could not typical averages on other days and go with that?

2 Likes

Interesting mystery. A few things to think about:

  • Sense doesn’t make any assumption about max amperage, though it will attempt to make a call on full charging usage based on the power ramp it sees.
  • That means Sense does not continuously “monitor” the EV charging usage, so if your usage goes up after the initial ramp, Sense won’t see it (unless you have your charger on separate direct current monitoring CT).
  • I’m also wondering if you have you Wallbox in a solar output following mode, which some can do. That could lead to a case where the charger in the car is initially current limited by solar production, and only later able to charge at the full 48A (that would be a good sized solar install).

Here are screenshots of the start and end of ramp up:
CleanShot 2022-04-13 at 11.20.48 CleanShot 2022-04-13 at 11.21.09

You can see the ramp up goes from 883 W to 11572 W in 26 seconds in two recognizable steps with a brief plateau at 4555 W, and 8176 W. None of these date points seem to correlate with the indicated 9168 W.

The overall charging looked like this:
CleanShot 2022-04-13 at 11.24.16

So as you can see there is no major variation during the 4.5 hr charging session, ending at 11233 W

  • Sense doesn’t make assumptions: As that an official statement by sense, or conjunction on your part? The blog you quote (and which I had read before) does not say anything on this subject.
  • Not continuously monitoring. OK, I can see that, but it does imply that if the original determination of wattage attributed to charging is wrong, it stays wrong for the whole session.
    • Now that my charging is done I can tell you that the Tesla app reported 50.7 kWh added/used
    • Sense reported 46.1 kWh
    • Charging time was 4h36m
    • So 46.1 / 276 * 60 = 10020 kW is what it effectively thought
    • Tesla says 50.7 / 276 * 60 = 11086 kW
    • So an effective difference of over 1 kW as I originally observed
    • I do not have the charger separately monitored. I use 2 CT for my main feed, and the other 2 CT on my solar circuit. No more CT possible with a single sense device
  • The Wallbox is hardwired, set to 48A max on a 60A dedicated circuit. It is not in solar output mode
    • As can be seen on the screen shots I charged from a little after midnight to about 5AM
    • I am on net metering so when I charge does not really matter within the off-peak period, which for me is 9PM - 4PM
    • So no real benefit to solar following anyway
    • All my grid energy is 100% green so timing makes no difference there either (either that or my own provided green energy)
    • My chosen timing is mostly for (a) convenience, and (b) best times for overall grid load
    • A side effect of the chosen charging timeframe is that most of my other load is quite even throughout: except for my computer use at the very beginning, we were sleeping so overall load from other factors is fairly constant. It is less than 400 W so generally is dwarfed by the Tesla anyway. Refrigerator on off, etc. are the bumps on top of the charging, mostly flat, level.

So all that said, I am not (yet?) convinced sense is not making an assumption. They state in the blog that they optimized for detecting specific make/models. Given that these cars can all charge at different levels and that they have to make sense out of the overall usage picture it is not completely unreasonable they assumed some maximum (although in this case it would have been wrong).

Now overall my goal is tracking the usage in home assistant and through sense (at this point) that is not feasible/correct. However, I also discovered the Tesla integration, which accesses the Tesla API to get this (and other) data.
image

As you can see that provided an accurate measurement, so I’ll be using that. Yet I hope sense will eventually get this right.

1 Like

Thanks for the screenshots and answers. Responding:

  • My answer (no assumption) is not official. It’s an inference based on how I know Sense to operate, using on/off transitions rather than levels to determine power usage of detected devices, plus knowing that neither of my Tesla’s show a 40A assumption (Sense gets both usages values about right when it detects - 48A and 80A)
  • I’m in the same situation as you - Sense with solar and NEMS2 TOU EV2A electricity rate, so there is no benefit to to charging based on solar output.
  • I do use a second Sense so that I can watch the usage of one of the HPWCs in my garage (I have 2) as well as watching all my electric heating. I only use that second Sense account in conjunction with Home Assistant to integrate additional data.
  • Since you are using Home Assistant, you could track Sense vs. Tesla Charger with a few minor adjustments to a YAML file. I’m seeing a better match between Sense usage and Tesla usage, when Sense doesn’t miss a detection. Here’s how to overlay both Tesla and Sense power usage info.