The top breakers in your service disconnect go to you main panels through the wall. The lower breakers are going somewhere else, they are not being monitored (by Sense). Your three condensing units makes perfect sense. Does the missing usage coincide with their use? Was it warmer on the 14th than the 16th? Something else could be tied into the units circuits, follow the cables over to the units. Before you move your CT’s flip the breakers off for a day or so and see if everything matches up. If it does you could always dig in more and turn the breakers on, crank up the t-stats so the units don’t turn on. Then you would know their standby wattage (if any).
You have a couple options.
You could move the CT’s to the meter panel and monitor panel 1 and 1 condensing unit, and the other CT’s monitoring panel 2 an 2 condensing units.
If your not concerned about Sense monitoring (or ever discovering) those units you could always leave everything alone and know how much those units are consuming by the difference between Sense and your utility.
He should be able to put the CT’s around the conductor coming from the meter to outside panel 1 and also do the same for the conductors in outside panel 2 with the solar CT’s. This will meter everything. If the ct cables are long enough, or get extensions, then he should be able to leave the sense monitor inside and run the ct cables to the outside boxes.
Great advice. I’ve turned the breakers off and wrote down the time. It’s a pretty mild day so I may leave them off for a while (at least until my wife threatens divorce). I care less about monitoring the condensing units. I really wanted to know Sense was accurate and where the discrepancy was. This community has been extraordinarily helpful in that regard. I would like to see in the future Sense to have the ability to add more dedicated circuit CT’s or be able to combine data from multiple Sense devices. Thanks for everyone’s help. I’ll let you know how the data looks after this test.
Extension cables for Sense current sensors are available in 4’, 12’, 25’, and 40’ lengths. If you need more length, you can attach two of the extension cables together.
@whalleye Keep in mind if you run the CT’s to both outside panels then you will be capturing the load on the outside AC units and both indoor panels. You can still power the sense from the inside panel if you wish. The other option would be to knock a hole through the brick and run a conduit to both outside boxes and mount the sense on the inside wall. The extension cable is probably the easiest/cheapest option. An electrician should have a fiberglass rod/fish tape to pull the ct cables through the conduit.
Wanted to let you know that those breakers were the problem. Uploading the charts to show you. I shut the breakers off at 7am and turned them back on at 8pm. Thanks again for everyones help!
That is great. You should be able to put CTs at the top of the feed after the meter. I have the same situation as you. Can you upload pictures of the inside of first subpanel after the meter?
I finally moved my Sense outside. It was installed about 2pm on May 6. I’m uploading the chart and pics of the installation (although they may not be very clear). Bottom line is it looks like it is catching all the power and is very accurate except for the first 2-3 hours. Couple of questions for the group…
Any idea why, shortly after installation, a few times Sense saw more power than the utility? Does it take a while to “get adjusted”? Seems like it was an anomaly and never happened again.
Some devices that were found (and consistent) prior to the move are now coming on when the device is actually not on (like a hair dryer). Should I delete the device and just let it find it again over time?
Great to see Sense working. I don’t think the bump up and down on the Sense side is due to the monitor “settling in”. With Sense showing numbers above and below your utility during a few hour stretch, I would put my bet on the difference stemming from “estimated reads” on your utility’s part. That’s when the utility misses a few meter reading transmissions and is forced to estimate based on past history. Eventually the meter succeeds in updating the central office again and the estimation syncs up with the real usage. This gives rise to a few hours where the data diverges, but the net delta over that stretch of time is zero.
You might want to look at the entries in then.cave from your utility around that point in time to see if there is any indicator of “estimated” - for me, it’s in the notes column, if/when it happens.
Thanks for the reply. Any idea why it is not recognizing some devices it has previously found? Is it because of the additional power it is seeing (a more crowded room)? Should I just leave the found devices as they are or delete them and have Sense go through the identification process again?
If the Sense device (Hair Dryer) is turning on when it’s not actually on, I might expect that you added additional loads that look similar to the hair dryer. You might have also inadvertently switched phases (L1, L2) - that could cause Sense to lose the actual hair dryer detection and possibly associate a similar device, on the opposite phase, with the hair dryer detection. If you didn’t swap phases, and Sense is missing the hair dryer, it might be the more crowded room effect…
I’m not sure if I switched phases or not. If that is the case should i delete the device and let Sense find it again? I had an electrician do the work and it took several attempts before he got the CT’s around the correct wires where the load looked reasonable.
Quick question. Not a big deal but does anyone have any idea why Sense and the Utility wouldn’t match up during a one hour time period? Has anyone seen anything like this? Attached is the graph which shows perfect alignment between Sense and the utility except for one hour. I am thinking Sense is correct and the utility is wrong.
There are many possible reasons for discrepancies between Sense and a utility revenue-grade meter.
Sense and the utility meter viewing house loads / supplies from different points of reference in your wiring. You had this kind of issue at the start of this post and have since fixed. Keeping the the same metering topology as your home meter can be especially tricky when you have multiple energy sources / sinks (solar, generators, batteries).
Systematic Sense errors - for example, a loose CT, will cause Sense errors vs the utility that are proportional to usage on that particular CT. That doesn’t seem to be your issue.
Random Sense errors - for example, a short-term data dropout from a Sense monitor, will cause a drop in Sense usage vs. the utility for the affected hour. Occasionally you will see short late-night (US time) dropouts when Sense updates monitor firmware. You can see one of these here:
Utility systematic error - usually the utility revenue-grade meter should be regarded as golden, but there are cases where the messenger (meter-reader human, smart meter wireless connections) can cause systematic errors due to missing, early or late meter readings back at the utility. In the case of a smart meter with a spotty connection, the utility will fill missing readings with “estimated” readings, but eventually the utility will connect again and true up with the accurate meter data. From what I have seen, the utility usually flags these missing hours / reads as “estimated”.
I would look closely at the Sense Power Meter for the hour where the usage diverges.
