10-80 charging analysis post 2023.30 - underwhelmed

[inconceivable]

Hey all. I hope you are doing well and enjoying your Rivians (or eagerly awaiting them). i know I'm enjoying mine!

In the 2023.30. release Rivian indicated that they improved the thermals to improve the overall charging times. The exact release note was:

• Improved DC fast charging speed, by up to 2.5 minutes, at high battery temperatures. Thanks to Kyle, Out of Spec, for your videos!

This inspired me to evaluate the charging curve. I was hoping to see some solid improvements, but unfortunately I was underwhelmed. The overall charging times doesn't seem that different then what folks like @OutofSpecKyle and Branden Flasch found months ago. (Slight changes in the curve, but overall times to get to high states of charge are about the same).

Anyway, I thought others here might also find this interesting, so I figured I'd share.

Here are my quick takeaways:

• 10-80 was completed in 41 mins
• The truck started reducing charging rate much earlier than I expected. Peaked at 205 kw at 32%, then started to decline
• Going from 10-70% averaged ~5mins per 10%. But going from 70-80 took 9 mins. So... you may want to unplug at 70% since charging rate starts dropping pretty low around then
• The 'session summary' information is likely inaccurate. It indicated that the energy going to the battery was 100.9 kWh and total energy onboarded was 102.9, but EA said it only dispensed 98.7. Theoretically possible EA is wrong, but other calculations based on this data seems to be suspect (see the SoC vs Time chart below and see what the resulting SoC would be if you inferred it from the session summary info)
• The SoC on the dash is also likely inaccurate. During the first 5 mins of charging in indicated it only gained 6% despite spending most of the time onboarding ~200 kw of energy. I believe some 'BMS sway' (as @OutofSpecKyle coined it) is still happening
• If you infer the SoC from the 'session summary' information, there is a pretty big gap between that and the SoC on the dash. I'm guessing the real SoC is somewhere in between, but ?‍

Here is a link to my data. I didn't video the session, but I did take a bunch of pictures (1 a min) which I based the data off of. If there is interest, I can try to include the pictures in the spreadsheet so you can see more of what I saw.

Below are a couple of charts I thought were interesting.

Sponsored

 

[mroe]

I generally aim to charge from approximately 10-70% and spend ~30 minutes charging. Here's two charge sessions on 2023.30 (RAN and EVGo).

 

[miasm]

30 minutes to go from 10-70 is pretty bad. They really need to figure out a way to leave it uncorked at 200kW until 50% or more. I honestly thought it held 200kW until 55 or 60 in the past, but can’t really remember.
I just always wonder if they’re being conservative, or they just need more / better cooling.

 

[mroe]

I don't think we're going to ever see something significantly better until either Rivian re-architects the battery to address the thermal problem or moves to 800V.

 

[SeaGeo]

The charging time isn't only because of the thermals. They have a very odd and inefficient quasi-constant voltage phase starting after roughly 45% to 50% once the pack hits the voltage target. That's what the steps are that we see starting in that range. Instead of slowly ramping the current down to keep a constant voltage, they drop it significantly and then hold it until the voltage slowly creeps up again, and then drop it. This stair-step approach has always kind of bothered me, and it does even more-so knowing that these are the same batteries Lucid has in the GT performance and Dream Edition.

Recently I did a comparison of the Lucid's curve and Rivian's curve normalized for cell count (see below). The "rivid" curve is just a the Lucid curve scaled up for cell count, and capped at 290kw. I did 290kw because I was lazy and didn't want to track the voltage change curing a constant current phase if they opened the pack up to say 680A.

Even without increasing the peak speeds at all, it looks like Rivian is leaving about 5 to 6 minutes on the table from 10-80% on the table by having this stair-stepped CV phase compared to what Lucid would be doing with the same batteries. So that 41 minute charging time could really be something like 35 minutes. You'll noticed the big chunk of area between the blue and orange curves starting at about 65% SOC up to 90ish percent. Even with the smaller battery pack Lucid is still pushing more power in that range (yellow curve).

The bump to 300kw only resulted in another ~3 to 4 minute savings or so, and that's without accounting for any additional charging losses. Bumping to 350kw was about another half a minute of time saved.

*note - I plotted up Kyle's 450A curve because his 500A curve seemed to have some oddities with power at some of the mid-SOC points. That the old 450A curve matches the "Rivid" curve almost exactly in the 50% to 65% range is a nice little piece of confirmation of the cells. If you plot up a 500A curve it matches pretty well from about 45%, his 500A curve just wasn't as clean.

Beyond the higher SOC charging issues, they're still being pretty cautious with temps IMO. We can all guess as to whether the cooling in the pack isn't up to the challenge, but they're still not pushing the packs terribly hard, and the cooling system is similar to Lucid's. I'm not convinced the pack has a thermal issue as much as they're just a more conservative company.

 

[SeaGeo]

30 minutes to go from 10-70 is pretty bad. They really need to figure out a way to leave it uncorked at 200kW until 50% or more. I honestly thought it held 200kW until 55 or 60 in the past, but can’t really remember.
I just always wonder if they’re being conservative, or they just need more / better cooling.

Even if they were able to open it up to 350kw, it would still be something like 25 minutes as far as I can tell. It seems to be tied to the cells as much as anything.

 

[inconceivable]

30 minutes to go from 10-70 is pretty bad. They really need to figure out a way to leave it uncorked at 200kW until 50% or more. I honestly thought it held 200kW until 55 or 60 in the past, but can’t really remember.
I just always wonder if they’re being conservative, or they just need more / better cooling.

I agree. I also think I’ve seen the riv hold 200kw+ up until about 50%. Makes me think they’re still leaving a lot on the table by de-rating early (in my case above at around 32%) likely due to thermals as I can’t think of anything else that would cause it (anyone else know of another reason?).

I would think if they could hit their ‘ideal charging curve’ without thermal de-rating, they can likely shave minutes more off the time. Based on the release note, I was hoping that was exactly what they did, but I guess not. ?

 

[HaveBlue]

100kw delivered in 40min Is an average of 150kwh isn't it? That seems better than previous averages of 120kwh. Maybe wasn't 10-80 though.

 

[zefram47]

Looking at what I thought was a pretty good RAN session in May that I documented, I saw 43 min from 10-80% with the truck adding 50 kWh in the first 15 min just sitting at 200-215 kW. Going from 43 to roughly 41 min is consistent with what Rivian claimed in the update.

 

[Dark-Fx]

Instead of slowly ramping the current down to keep a constant voltage, they drop it significantly and then hold it until the voltage slowly creeps up again, and then drop it. This stair-step approach has always kind of bothered me, and it does even more-so knowing that these are the same batteries Lucid has in the GT performance and Dream Edition.

It's still in CC mode at this point, they are adjusting the curve due to temperature. If you plug in at high 30's it'll stay at 500A much deeper into the charge.

 

[SeaGeo]

It's still in CC mode at this point, they are adjusting the curve due to temperature. If you plug in at high 30's it'll stay at 500A much deeper into the charge.

Before about 50% I agree - unless you manage to get it toasty.

After ~50% it's not a true CC as they're really just stair stepping down a CV line in really large steps (unless you get it really hot before hand). For example, the large steps down at 70% and 85% have been pretty fixed, and those steps aren't caused by heat. It's been basically the same series of steps for the last year and a half.

I've added a curve I had from April and @zefram47's. That smooth drop we both had in the 40% range is them reducing speed and managing the temps. The stair steps we both start to see in the 65 to 70% range are what you get in that range once the temps have been reduced back to what they are targeting, but it's at pack voltage.

 

[SeaGeo]

I agree. I also think I’ve seen the riv hold 200kw+ up until about 50%. Makes me think they’re still leaving a lot on the table by de-rating early (in my case above at around 32%) likely due to thermals as I can’t think of anything else that would cause it (anyone else know of another reason?).

I would think if they could hit their ‘ideal charging curve’ without thermal de-rating, they can likely shave minutes more off the time. Based on the release note, I was hoping that was exactly what they did, but I guess not. ?

Back when Branden had his truck we was regularly seeing 44 to 45 minute 10-80 curves. 41 minutes is a savings. The fast charging behavior just isn't great.

Personally I'm not actually sure I've ever seen anyone hit the 140 miles in 20 minutes statement using the original EPA range that Rivian has used as their benchmark, which would have required picking up about 45% in 20 minutes. I've seen it get close, but never quite there. The problem is that metric is so vague and nebulous that they could say they were basing that off 21s in conserve mode rather than the rated range.

 

[SeaGeo]

On a related note, I haven't been able to log a good session since this spring because the EA chargers in my region derate in the heat before the truck does (the old ABB chargers). I really appreciate it when I get the change to use the EVGO Delta units and am excited to have a reason to use the RAN charger SE of Portland next month.

 

[azbill]

It's still in CC mode at this point, they are adjusting the curve due to temperature. If you plug in at high 30's it'll stay at 500A much deeper into the charge.

This is my experience as well. I have seen it over 200kw up to at least 50%.

Ford does something very similar, the curve adjusts after a certain time to avoid thermal issues.

 

[azbill]

On a related note, I haven't been able to log a good session since this spring because the EA chargers in my region derate in the heat before the truck does (the old ABB chargers). I really appreciate it when I get the change to use the EVGO Delta units and am excited to have a reason to use the RAN charger SE of Portland next month.

It is not just the old ABB chargers, I have seen the brand new dispensers at EA limit as well. I have watched them go above 200kw for a few minutes, then suddenly drop to 45kw. Most of my experience with them is in the desert heat.

I had one new unit start at 32kw and stay there, and another one crap out after 30 seconds. The new units are no better than the old ones.

Sponsored