Peak Usage Data Field Instead Of Just A Notification

It would be useful (for me, anyway) to have a data field with peak instantaneous use along with the date and time of the peak. This would be much easier to use than scrolling back through months or years of notifications. A list of running devices and their draws at the peak would be useful, but I can see some programming challenges.

My context is that NEC calculations show that we would have to upgrade our service from 200 Amps to 400 Amps if we replace our pool heat pump and add an EV charger (or two). So far Sense is showing our actual use to be far below what the NEC calculations would indicate, but a specific peak value would be very useful in deciding whether we will need to upgrade our service or not. (Which in turn could even factor into our decisions about purchasing electric vehicles.)

An electrical contractor told me that quite a few people are preemptively upgrading their service when adding an EV charger. Having definitive information about peak usage might enable people to avoid this expensive service upgrade and also facilitate their decisions about buying electric vehicles. These benefits could be added to the Sense marketing material.

(It would also be useful, and a simple bit of additional programming, to also have a data field with peak solar output or dedicated circuit draw.)


I can see how this would be useful. Yet the concept should be more specific about certain details.

Would this be an all-time peak value? That might unfairly benefit people who are new users, since they won’t have a long history. Picking a constant value, such as the most recent 2-year period, might be more equitable.

How long should the load be sustained? The quote below seems to allow a few minutes of overload, within limits. The startup current of a motor could really increase the peak usage value, but just for a second. This might not “count” for sizing service panels.

Does anyone know of an official answer on how to do this calculation? If the NEC or some other authority defines it, that would be great as a reference standard. Otherwise whatever is implemented will be second-guessed to no end by people with different opinions about the answers to questions such as these.

1 Like

Good points. (And maybe the fear of second-guessing and nitpicking is why this hasn’t been implemented.)

Sense has to have a “last peak” value in their database in order to determine (each second?) if the current reading is a “new peak.” They would have to keep every reading (once a second?), however, in order to be able to implement a sliding 1 or 2 year window.

I was thinking instantaneous peak. I don’t know if Sense would/could define “peak” using a combination of draw and duration that somehow relates to what it would take to trip the breaker on the mains.

One more thought…it might be useful to have a “Clear” function on the peak value so that folks could see the benefit if they have implemented some energy-saving technology. If their consumption has dropped, they will never see a “New Peak.” E.g. I just installed a new super-efficient heat pump HVAC, so I doubt that my current usage will ever exceed the prior peak. This suggestion applies to both the existing “New Peak” notification and a proposed display field.

1 Like

I don’t care if new Sense user or old Sense user, I’d just like to be able to be given my peak without playing CIA Analyst to find it.

This is such an obvious missing detail.

In fact, give us the 5 highest historical peaks, time and date, similar to what is done in the Power Quality Tab

1 Like

I like the idea of Sense providing a range of statistics of the instantaneous (1/2 second) power readings for the past month or year. Max is useful, but median and a histogram might also be useful.

BTW - when I’m searching for max values, I just expand the app Power Meter for the full time period I want to look at, then measure the peaks.

1 Like

One more note, here’s a view of the distribution of power usage in my house, sampled every 2 mins from Sense via Home Assistant. Note that there are many distinct peaks in the distribution, including a small peak way at about 32kW (two EVs charging at the same time).

Thinking about it a little more, @Chuck_Davis . I’m thinking that this kind of plot might be far more useful for investigating the need for a service upgrade, sizing a generator or evaluating a solar upgrade.

1 Like

And this one even tells more - distribution by hour. We can see that the two tiny sets of maxes in the circles are concentrated in the midnight to 6AM hours, all EV charging. The peak around 20kW is Model S charging and the peaks around 12kW is the Model 3 charging.

Interesting data, but I doubt that a significant number of Sense users have the chops to understand it. It’s helpful for me to know that with two EV’s charging late at night that the total draw for the house is only 133 Amps. Net of that info and our new whiz-bang HVAC system, I’m feeling better about not having to upgrade our service from 200 Amps in the future.

As an old tyme systems designer, I prefer to approach from a problem to be solved or a question to be answered. In my case, the question is “If I add Device A and/or Device B is my main breaker likely to trip?” A recorded peak, with some sense of when/why the peak occurred, plus the rated draws of Device A and Device B, should get me in the ballpark of an answer.

1 Like

Something like this would be useful. I’ve got 150 amp service and just got an EV. Will be getting quotes from electricians soon about running a 240v line in the garage, but it’d be nice to know ahead of time if a service upgrade will be needed.

Absent any other information, the electrician would have to use the NEC calculations. Getting an eyeball estimate of your peak(s) from the Sense Power Meter, along with a sense of time-of-day of those peaks, plus the rated draw of the charger in the EV, should help you make an informed decision.

Luckily tripping the breaker on the mains is considered a nuisance issue and not a safety issue. It’s a pretty big nuisance, though.

You raise a good point @jefflayman. I did a quick look at NEC suggestions. They seem to separate out continuous load (3 hours or more) from non-continuous (3 hours or less).

NEC sizing rules. Secs. 210-22(c), 220-3(a), 220-10(b), and 384-16(c) all relate to the sizing rules for overcurrent protective devices (OCPDs). The first three all specify the same requirement:

OCPD size = 100% of noncontinuous load + 125% of continuous load.

From a circuit breaker perspective UL 489 and 1077 rated breakers have trip curves that are a function of current and sustain time. I’m looking at a 200A breaker that is similar to the main one in both of my subpanels. It looks like it takes about 1000 seconds of continuous current at the rated (200A) current before it would trip. Alternately, it can handle 10 continuous seconds of 4x the rated current before tripping.

@Chuck_Davis , Combine those two and you can get an answer to service sizing, but it’s really likely a function of both current and sustain time.

1 Like

Great information. Thanks!

Seems like Sense would need to provide a fair amount of coaching to folks trying to use their Sense data to determine if a service upgrade is necessary. Might be another reason they have avoided this particular patch of weeds.