As far as I know, we’ve had pretty good luck with model S and X. 3 is still pretty new so we just don’t have enough data to make major advancements on that yet. Have you worked with support at all on detecting your S?
@vespula, I can’t speak to specific models, but the data science team is working very hard on improving overall EV detection. It is a major undertaking however since EV signatures look very different from normal, quick-transient household loads.
Thank you RyanAtSense. 
From conversations reported by the data scientist since March 2017, it is clear that EV signatures are difficult. Currently there are over 40 different EV models in the US, so the issue is going to mushroom. I don’t want to sound impatient because I love your product, your staff, and approach to solving electrical signatures. It looks like this issue has been studied for over a year and could be another year as so many EVs are being added. When l mention CT clamps, I am not looking for a quick fix or band-aid. I just wanted to comment via the forum that if it has to be the answer, in the end, it is acceptable to me. For the general Sense concept, it probably feels like a cop-out. I am just hoping that by this time 2019, my Sense sees my Volt and Model 3.
Thanks, Mark. We’re confident that the rest of 2018 and 2019 will bring great things for Sense. Thanks for the support.
No, not all to the poll. Thanks! I fit into a Niche that this article I posted on IEVs will explain.
https://insideevs.com/yes-your-ev-can-charge-on-sunshine/
Thanks again, Ryan. I am in it for the long hall with you and your product.
With 8,700 views to the thread I would say you are correct.
@markhovis73 created a polling tracking users with EV devices and solar arrays. Check it out: How many Sense users have both EVs and Solar Arrays?
Ok…figure it’s been a couple of years…checking in when sense will be able to detect my leaf EVSE?
I don’t care if it knows what it is, I just want it to recognize that daily pattern. I’m happy to identify it. Would it help if I sent in a timeline picture when that is the only thing happening?
I have the same problem. I have 2 Teslas and non of these gets detected. The pattern should be very easy to detect. Quick ramp up to 12 Kw followed by a drop of 12 kw 1-2 hours later. I also have a similar problem for my AC units. It looks to me like sense is tuned in for lower power devices and is missing the high powered ones. I love me sense but the consumption reports are pretty much useless as the cars and the ACs are my represent the majority of my consumption. It we would be great to see some focused effort on EV’s and AC detections
Based on my experience, Tesla detection is currently liimited to models with the 80A dual charger. We have 3 different models and that’s the only one ever detected, but Sense does identify its charging cycle reliably.
It hasn’t found my Chevy Volt’s EVSE in 1.5 years!
It does seem that Sense should be able to pick up heavy loads such as clothes dryer, oven, and EVSE devices being that they are 240v and not 120v and their startup currents and operational currents should have unique signatures easy to pick out from other “noise”.
I think this is what frustrates so many of the Sense users that, on the surface, one would think that it is relatively easy to recognize these high power devices. The pattern recognition software of Sense doesn’t seem to be as adaptable as the documentation implies. I think most feel that there are devices that should be easy to discern and others, like lower power devices like LED bulbs, that would require careful tweaking of the models.
AI depends on large amounts of sample data. How many samples do you think they have for these newer EV’s?
@miracj,
There’s a thread somewhere here in the community in which the Sense rep acknowledges that the original identification software (likely a set of recurrent neural nets - RNN) was set up to find rather short on/off power signatures, maybe on the order a second or two. And that was premised on immediate identification, where the bubble pops on immediately after the device turns on.
But the defining characteristics of EV chargers are a longer ramp up to a fairly high level of power for a long period, easy for us to see, but out of range for a RNN with a 1 second history. Sense can reliably identify the charging of just one of our three Teslas, so they have some knowledge on how to tackle longer windows. The other half of the Sense message was that they were working on infrastructure to deal with longer identification cycles, where the bubble might pop up 10 min into the car charge, but the whole charge cycle would be accurately post-processed into the recognition history.
AI does not necessarily require that much data, and only for those objects that fall into the “noise”. High power devices don’t have to fall into the noise. Besides, the data being collected has been for almost 2 years now.
I think there’s probably 500 repetitions of the same pattern at my house - enough for the software to say “Here’s this pattern I keep seeing, and when I see it. Any idea what it might be?”
That’s not how it works. You should really do more research into the product.
This is pretty spot on (I was probably that rep 
)
Basically, we understand how EVs can be different, it’s just a matter of turning that knowledge into actionable insight for Sense and to do that we need data from our users. And, to address your questions @miracj and @robinson, we need a lot of data. Sense is based on pattern recognition and we cannot recognize the pattern of a specific EV without seeing it (1) repeatedly over time and (2) in various contexts, i.e., different homes.
@RyanAtSense anything us users can do to help Sense HQ identify these patterns? Perhaps we can self-identify as owning an EV?
I’ve only had my Sense for about 2 days now, but I would love to help sense find my two EV’s, and others in the sense community. Would it be better to plug the EV in at the same time every day? or is it better to be random? I can be flexible and do either. I have a JuiceBox 40 Pro 240v charger that I share between a Chrysler Pacificia PHEV, and a BEV BMW i3. The JuiceBox produces great data on charge cycles, but it can’t distinguish between the two cars, so all of my data is lumped together.
Whatever I can do to help, I would be more than happy to do it.
Thanks