Model S Plaid Battery Details, sourced from EPA Docs, Vehicle Observations, & Supercharging/Charging data

[AlanSubie4Life]

It's too early to conclude that. There are only a couple of indications of Wh/mi equivalent comparisons. It seems possible that the Model S in both variants might actually be more efficient than is a P3D in terms of Wh/mi.

It’s less efficient. Obviously tires make a huge difference and my original statement makes that clear (you can't compare vehicles with much different tires - I'm talking about the powertrain and the aero fundamentals). However, the new Model S is fantastically efficient, and a huge improvement over prior models. It has more efficient aero and more efficient drivetrain than Model 3 - and that almost makes up for its weight and size. Hopefully the new motor technology will trickle down to other vehicles in the lineup.

The Plaid and the new S LR are fantastically efficient and the power train is much better, but we have the apples-to-apples comparison:

AC Numbers (Wh/mi):
US06
Plaid: City Portion: 295.6. Highway Portion: 274.5. https://iaspub.epa.gov/otaqpub/display_file.jsp?docid=52948&flag=1
Model S LR: City Portion: 254.1 Highway Portion: 232.5. https://iaspub.epa.gov/otaqpub/display_file.jsp?docid=52949&flag=1
Model 3 Performance: City Portion: 262.3. Highway Portion: 258.6. https://iaspub.epa.gov/otaqpub/display_file.jsp?docid=51301&flag=1
Model 3 LR: ???? (It's not in the document that I could see...unfortunate!)

Note the S LR is substantially better than the P3D (not surprisingly, given the impact of aero and wheels). Obviously tires and aero wheels matter a lot!
As mentioned above, you have to convert these to DC numbers by the charging efficiency - I'm not going to do that here, but the numbers to use are the efficiencies mentioned above for the Model S, and for Model 3, it is 86% for Performance, and 88.8% for the LR. This won't have much impact on the relative positioning here, in any case.


So, let's go to straight highway numbers (note, less aggressive than the US06 cycle, but the only way to "calibrate" the Model 3 LR:
These can also be done directly in DC numbers:
Plaid: 99.3kWh/443mi = 224Wh/mi https://iaspub.epa.gov/otaqpub/display_file.jsp?docid=52951&flag=1
Model S LR: 98.3kWh/512mi = 192Wh/mi https://iaspub.epa.gov/otaqpub/display_file.jsp?docid=52949&flag=1
Model 3 Performance: 80.8kWh/400.2mi = 202Wh/mi https://iaspub.epa.gov/otaqpub/display_file.jsp?docid=51235&flag=1
Model 3 LR: 78.6kWh/447mi = 176Wh/mi

So you can see once again that the S LR is more efficient than the Performance Model 3, but if you equip the Model 3 Performance with the wheels and tires from the Model 3 LR, it will be comfortably equivalent to the Model 3 LR, and about 8% better than the Model S LR. This is a low speed test so the gap is probably smaller than this at speed.

So Model 3 LR & Performance, when appropriately equipped, are perhaps 5% more efficient than the Model S LR.

But the new Model S is spectacularly efficient given its size and weight. It is impressive.

You can compare to the old Model S Long Range Plus: 103.9kWh/538mi = 193Wh/mi. (So similar to the new Model S LR...but again the tires & wheel combo may not be comparable! The old Long Range Plus is surprisingly efficient, and not sure how much of a role the tires play in that (I think a lot). I'm not bothering to look up the old Long Range numbers...)

Anyway, lots of numbers out there, and it makes the picture clear this is a very efficient vehicle.

Clearly in all the numbers above, it all depends on those coefficients that Tesla uses. They could sandbag them a little. Really hard to know. I'm assuming they are just using the "correct" values.

 

[AlanSubie4Life]

I meant the refreshed LR is 98.3 and the Plaid is 99.3 but both batteries are 450V max/410V nominal.

It's just variation in the test. Cars are run until they shut down and that varies from vehicle to vehicle. Also packs of course vary in initial capacity.

The two vehicles have the same pack, at least according to all information provided in the document.

 

[AlanSubie4Life]

I guess now it's time to guess on exact pack size! Based on the info provided in the other thread (sorry there are two...that one seemed more general/speculative, so I started this one, to be based on specific known info...though this post is now veering towards speculative...), it's likely 110s:

Looks like the Model S Plaid is ~460V at 90% SOC

https://www.reddit.com/r/teslamotors/comments/ob8uhk

Maybe 112 cells in series? My LR Model 3 reported 401V at 90% SOC when Supercharging...

I'm impressed that they are allowing voltages that high - V4 SC cabinets appear to only be rated for 410V.

From above: I'd guess this is more like 110s (4.2V*110 = 462V), which tracks well with the 403/401 at 100%/90% datapoint provided here (we know 4.2V per cell from lots of Model 3 SMT data). 112s would be 470V max which seems too high for 460V at 90%. (Since Model 3 at 100% is ~403V, and apparently 401V at 90%.).

On the other hand, 110 is divisible by 5, but not by 4. 112 is divisible by 4, of course, and the pictures earlier here suggest four modules, not five. In Model 3, they didn't have identical series cell counts in each of the four modules, so divisibility by 4 isn't required, though...even if there are four distinct modules. (Could have 2x26 and 2x29 or something). 112 being the actual series count would imply significant undercharging to me , so I guess I don't think it is likely - the curves we've seen don't fit that hypothesis correctly (it's way short of 470V, and short enough that it means way more than a few % short of max capacity, I think).

So with that and the ~99kWh capacity vs. the prior ~103kWh (96s86p)...the question is whether the individual cells are different. The density of the pack went up considerably (see above), but perhaps that is just packaging and not cell changes?

If no cell changes then: 96*86*99kWh/103kWh/110 = 72.

So maybe 110s72p? 7920 cells instead of the prior 8256?

Or if there has been a few % density improvement in the individual cell (like Model 3 2170L cells), then maybe 110s70p? This would potentially limit supercharging speed though (depending on the C limit of the cells and whether that scales with the density), so I kind of doubt it.

Maybe we should start a poll!

 

[AlanSubie4Life]

192 vs my 290 Wh/mi.

Gotta be careful about comparisons like that. 192Wh/mi is the raw efficiency on the very slow highway EPA test cycle. You'd have to look at the same data from your model year Model S to see how it compares exactly. But I'm fairly sure it won't be 290Wh/mi. For sure the new one is more efficient (see the apples-to-apples comparisons above). Probably by at least 5%, not sure how much more vs. much older vehicles (with the AC induction motors, etc.).

 

[Dave EV]

So maybe 110s72p?

That would be my guess...

 

[razh]

I guess now it's time to guess on exact pack size! Based on the info provided in the other thread (sorry there are two...that one seemed more general/speculative, so I started this one, to be based on specific known info...though this post is now veering towards speculative...), it's likely 110s:



From above: I'd guess this is more like 110s (4.2V*110 = 462V), which tracks well with the 403/401 at 100%/90% datapoint provided here (we know 4.2V per cell from lots of Model 3 SMT data). 112s would be 470V max which seems too high for 460V at 90%. (Since Model 3 at 100% is ~403V, and apparently 401V at 90%.).

On the other hand, 110 is divisible by 5, but not by 4. 112 is divisible by 4, of course, and the pictures earlier here suggest four modules, not five. In Model 3, they didn't have identical series cell counts in each of the four modules, so divisibility by 4 isn't required, though...even if there are four distinct modules. (Could have 2x26 and 2x29 or something). 112 being the actual series count would imply significant undercharging to me , so I guess I don't think it is likely - the curves we've seen don't fit that hypothesis correctly (it's way short of 470V, and short enough that it means way more than a few % short of max capacity, I think).

So with that and the ~99kWh capacity vs. the prior ~103kWh (96s86p)...the question is whether the individual cells are different. The density of the pack went up considerably (see above), but perhaps that is just packaging and not cell changes?

If no cell changes then: 96*86*99kWh/103kWh/110 = 72.

So maybe 110s72p? 7920 cells instead of the prior 8256?

Or if there has been a few % density improvement in the individual cell (like Model 3 2170L cells), then maybe 110s70p? This would potentially limit supercharging speed though (depending on the C limit of the cells and whether that scales with the density), so I kind of doubt it.

Maybe we should start a poll!

Its realistic to assume pack density decreased to accommodate increased cooling infrastructure. During the delivery event, Elon claimed the Plain can be run from 100% -> 0% on the track without seeing heat-related performance degradation. I can't imagine all of this was achieved just by installing a slightly larger radiator.

Even though pack capacity decreased, the overall efficiency improvements (aero, motors, heat pump) more than made up for it. A nice side effect is faster charge times (both because a smaller pack is faster to fill and and from improved cooling during charging)

 

[MP3Mike]

Its realistic to assume pack density decreased to accommodate increased cooling infrastructure. During the delivery event, Elon claimed the Plain can be run from 100% -> 0% on the track without seeing heat-related performance degradation. I can't imagine all of this was achieved just by installing a slightly larger radiator.

Even though pack capacity decreased, the overall efficiency improvements (aero, motors, heat pump) more than made up for it. A nice side effect is faster charge times (both because a smaller pack is faster to fill and and from improved cooling during charging)

According to the recently released EPA test data while pack energy capacity decreased, pack energy density, by weight, increased. (Assuming that the pack is the same physical size, pack energy density, by volume, decreased.)

 

[MarcG]

During the delivery event, Elon claimed the Plain can be run from 100% -> 0% on the track without seeing heat-related performance degradation.

I don't remember him saying that (and I was there in person). Can you please point me to the timestamp in the event video below, at which point he states this?

 

[razh]

You can find it at 12:00 in the video you linked, or just click this:

 

[Zoomit]

You can find it at 12:00 in the video you linked, or just click this:

Going from it “doesn’t get winded” to “Plain can be run from 100% -> 0% on the track without seeing heat-related performance degradation” is too specific for me. But your original point is valid. The cooling system overhaul likely involved more than just doubling the radiator size.

 

[razh]

Going from it “doesn’t get winded” to “Plain can be run from 100% -> 0% on the track without seeing heat-related performance degradation” is too specific for me. But your original point is valid. The cooling system overhaul was very likely more than just doubling the radiator size.

Unplugged Performance Pikes Peak run also confirmed the absence of heat-related performance degradation, but I guess those runs far from hour+ long track sessions. Going uphill for 10 minutes + flooring it out of most corners is a pretty good test though.

 

[Zoomit]

Unplugged Performance Pikes Peak run also confirmed the absence of heat-related performance degradation, but I guess those runs far from hour+ long track sessions. Going uphill for 10 minutes + flooring it out of most corners is a pretty good test though.

Randy purposely used “maintenance throttle” when possible to reduce the heat load caused by high discharge or regen. The cooling system is vastly improved, and those runs are a serious stress test, but it’s not accurate to say the PPIHC runs “confirmed the absence of heat-related performance degradation.”

I’d like to see the results of a 25-min session at Buttonwillow after recent Supercharging (on site!) in 99F temps. That will be more telling than a 7-min Pikes Peak run at 45F temps.

 

[MarcG]

Unplugged Performance Pikes Peak run also confirmed the absence of heat-related performance degradation

I wouldn’t go that far. In the video below, Ben (president of UP) talks about how power is abundant at any SoC, but even though cooling is much improved, it still has limitations that ”you *almost don‘t have to worry about cooling, but we’re learning where the limits are”

 

[MarcG]

Randy purposely used “maintenance throttle” when possible to reduce the heat load caused by high discharge or regen. The cooling system is vastly improved, and those runs are a serious stress test, but it’s not accurate to say the PPIHC runs “confirmed the absence of heat-related performance degradation.”

I’d like to see the results of a 25-min session at Buttonwillow after recent Supercharging (on site!) in 99F temps. That will be more telling than a 7-min Pikes Peak run at 45F temps.

Yes, this exactly.

 

[Zoomit]

I'd guess this is more like 110s (4.2V*110 = 462V)

Strong corroborating evidence for a 110s pack has been provided by @JWardell with his initial take on Plaid CAN bus data. YT video at 19:00

He's seeing the log data report max pack voltage of 461.76V. As was predicted, this infers a pack configuration with 110 bricks in series. This is 14.6% higher than the typical 96s configuration used in many of the prior S/X and all the 3/Y. While many other BEVs also have a 96s configuration, I'll note a few that have higher than 96s:

Polestar 2 - 108s
I-Pace - 108s
e-tron - 108s
Taycan & e-tron GT - 168s or 198s

462v apparently remains compatible with the Supercharger network as well as other global ~400v DCFC standards, but provides higher power while mitigating the heat loss that is a squared function of amperage. Stated another way, if Tesla stayed with a 96s pack, the same power output would require 15% higher amperage with a resulting 31% higher heat load into the cooling system (assuming no change in wire resistance).

This should also kill any speculation that the Plaid packs were secretly doubling voltage to ~900v.

I do expect that Tesla will be staying at 110s (462v max) for Cybertruck and Roadster V2 packs which should have a baseline design using 4680 cells. It's also possible that the 3/Y could evolve to this higher pack voltage with a version of the 4680 cell packs, though I don't know what specific ramifications that would have for their motor or inverter designs.

In short, 462v is the new standard replacing 403v; similarly, 110s will replace 96s.

 

[JWardell]

Strong corroborating evidence for a 110s pack has been provided by @JWardell with his initial take on Plaid CAN bus data. YT video at 19:00

He's seeing the log data report max pack voltage of 461.76V. As was predicted, this infers a pack configuration with 110 bricks in series. This is 14.6% higher than the typical 96s configuration used in many of the prior S/X and all the 3/Y. While many other BEVs also have a 96s configuration, I'll note a few that have higher than 96s:

Polestar 2 - 108s
I-Pace - 108s
e-tron - 108s
Taycan & e-tron GT - 168s or 198s

462v apparently remains compatible with the Supercharger network as well as other global ~400v DCFC standards, but provides higher power while mitigating the heat loss that is a squared function of amperage. Stated another way, if Tesla stayed with a 96s pack, the same power output would require 15% higher amperage with a resulting 31% higher heat load into the cooling system (assuming no change in wire resistance).

This should also kill any speculation that the Plaid packs were secretly doubling voltage to ~900v.

I do expect that Tesla will be staying at 110s (462v max) for Cybertruck and Roadster V2 packs which should have a baseline design using 4680 cells. It's also possible that the 3/Y could evolve to this higher pack voltage with a version of the 4680 cell packs, though I don't know what specific ramifications that would have for their motor or inverter designs.

In short, 462v is the new standard replacing 403v; similarly, 110s will replace 96s.

Plaid has 108 bricks (108s)
Discovered that a few hours after making the video, I'll have to do another soon

 

[AlanSubie4Life]

Plaid has 108 bricks (108s)
Discovered that a few hours after making the video, I'll have to do another soon

Really? So they are really taking these cells to over 4.25V or is there something we are missing here with the data (or is the 460+V datapoint incorrect)?

 

[wayne]

Nominal Volt: 450V ==3.6*125 ==3.7*120

From Owner manual

 

[JWardell]

Correction: It *IS* 110 bricks! Damn DBC didn't go high enough.
So your prediction is correct @Zoomit !

 

[Zoomit]

Correction: It *IS* 110 bricks! Damn DBC didn't go high enough.
So your prediction is correct @Zoomit !

@AlanSubie4Life did the prediction. Thanks for following up though!