More Tesla Model 3 Battery Info Than You Knew You Wanted: Part 2


Two Bit da Vinci returns as promised, with more truth about Tesla Model 3 Batteries.

A few weeks ago, we shared a video about Tesla Model 3 batteries. Tesla has a reputation for long-range vehicles, mostly due to battery cells with very high energy density. In fact, Tesla batteries are more energy dense than the competition. The Model 3 battery pack uses 2170 lithium-ion battery cells, which provide more power at a comparable cost to previous 18650 cells.

The first video in the series dove into the Model 3 battery tech for about 12 minutes and included tons of valuable information, especially for those that yearn to understand the technology. This follow-up adds another ~20 minutes and loads of additional data from a multitude of sources, which are all included below in Two Bit’s YouTube video description.

There’s no doubt this YouTuber is dedicated to the task and it’s outstanding that all the sources are shared. However, with the myriad of information out there (albeit some conflicting and misinformed), these deep dives are often a work in progress. Hopefully, as interest in the segment grows, we’ll continue to get a wealth of new data and deep dives about this ever-changing technology.

What do you think of Two Bit’s analysis? Do you have anything to add and/or corrections to point out? Please share your wisdom in the comment section below.

Video Description via Two Bit da Vinci on YouTube:

Tesla Source Data:
Teslanomics Data:
Jack Rickard Battery Teardown:

This is part 2 of our Series on the Truth About Tesla Model 3 Batteries. Today we’re going to discuss how Tesla takes these 2170 cells, and create their world class battery pack modules.

Categories: Battery Tech, Tesla, Videos

Tags: , ,

Leave a Reply

32 Comments on "More Tesla Model 3 Battery Info Than You Knew You Wanted: Part 2"

newest oldest most voted
F150 Brian

Great video!

Anyone spot that he used the picture of a clutch disk for the DC/AC converter?

I like how he compares EV to ICE – overall fair but for two things: ICE efficiency of 20% is very outdated. Most modern direct injection gas engines are 30-35% efficient with the record for a production engine at 54.4% (a MAN diesel). The other is cold starting. Even my wife’s 2011 Caravan never gets plugged in, even at -40 ad it starts just fine.


I too noticed both of the issues you’ve raised but in general he’s right, some older cars have problems cold starting and the 30-35% efficiency is maybe when the engine is brand new, also the fuel efficiency varies with the outside temperature, so … it’s a gray area.


The ICE in the new Toyota Camry, Avalon, and RAV4 is 40% efficient, while the ICE engines in the hybrid versions of these vehicles and the Prius are 41% efficient. These cars get practically the same MPG well past 100,000 miles if they are properly maintained, so their efficiency is the same as when they were new.

Once an ICE warms up to operating temperature in cold weather, it gets virtually the same efficiency as in warm weather. The greater density of colder air and the increased rolling resistance of cold tires are what lowers the fuel efficiency of an ICE vehicle in cold weather.

Get Real

And all ICE no matter how efficient pollute.


Those are some fairytales right there. Peak efficiency in the lab? Maybe. In real life? Ahahahaha.


Your ICE will never approach the efficiency of an EV specially in stop and go traffic.


“…ICE efficiency of 20% is very outdated. Most modern direct injection gas engines are 30-35% efficient with the record for a production engine at 54.4% (a MAN diesel).”

I’m fairly sure the energy efficiency figures you’re citing for ICEngines, 30-54%, are maximum efficiency shown in bench tests when the engine is run at its most efficient RPM speed.

If we’re talking about real-world average efficiency when actually being used to power a car driving down the road, then 20% efficiency is probably about right, or perhaps even a bit overstated. ICEngines in gasmobiles spend most of their time running at fairly low speeds, not the very high RPM speeds at which fuel efficiency is highest.


“. . . not the very high RPM speeds at which fuel efficiency is highest.”

That’s flat out wrong. You’ve got it backa**wards. ICE engines achieve their highest fuel efficiency when running at the lowest possible RPM. ICE vehicles have a wonderful device called a transmission. Perhaps you’ve heard of it?


You’re correct that low rotation speeds result is less friction loss. However, high pumping losses occur at low throttle openings which are typical in low power situations. Better efficiency would occur at low rotational speeds and wide open throttle (i.e., low pumping losses). This isn’t a common situation for an ICE.

Our 1st-generation Honda Insight had a lean burn mode that helped in the low power, low rotational speed situation that’s common with ICE’s. By running at 22-25:1 air-fuel ration (very lean), the engine produced less power, so the throttle had to be opened wider to maintain the necessary power which reduced pumping losses. However, a nasty side effect was the greater production of NOx gasses in lean burn mode. The Insight had a NOx trap that prevented NOx from exiting the exhaust, but its capacity limited the time that lean burn mode could operate.

The increased complexity required to increase ICE efficiency is increasing their cost and is likely reducing their reliability. The end of the ICE road is drawing ever closer…


alohart said:
“However, high pumping losses occur at low throttle openings which are typical in low power situations.”

Not anymore. ICE engine tech has advanced. Modern IC engines long ago eliminated the “butterfly” throttle valve, which was a major cause of pumping losses, and instead control engine breathing by varying the lift and timing of the intake valves with much, much smaller pumping losses.

A modern ICE vehicle traveling at 55-60 MPH is spinning only a couple hundred RPM over idle speed. At this RPM, without the pumping loses from a “butterfly” throttle valve, the ICE is running at its most efficient sweet spot, resulting in higher MPG.

Also, under the same load, a slower rotational speed uses less fuel than a higher rotational speed.


It doesn’t matter whether you restrict airflow into the cylinder with a butterfly throttle valve or specially controlled intake valves, you still get pumping losses.

You can reduce pumping losses by “Atkinson-izing” the cycle, keeping the intake valves open during part of the compression stroke. This sacrifices power, though. Hybrids accept this trade-off because they get extra power from the e-motor.

” the ICE is running at its most efficient sweet spot,”

If that were true the 4cyl Camry LE would be rated at 53 mpg highway instead of 39. The Camry Hybrid LE gets 53 from the same engine, just with heavily Atkinson-ized valve timing. (The improvement is even more dramatic on the city cycle, because the LE engine is even less efficient at lower power plus regen also helps the Hybrid).

“under the same load, a slower rotational speed uses less fuel than a higher rotational speed”

True down to about 1500 rpm (under load). Below that frictional losses drive fuel consumption up.

20% overall thermal efficiency is about right for the US fleet today, with the mix of highway/city operation and the amount of cold starts.

First off, the new 2.5L in the Camry actually runs on both the Otto cycle and the Atkinson cycle. It runs on Atkinson cycle under light load and switches on-the-fly to Otto cycle under heavy load (stomping on the gas pedal). So the difference in MPG between the hybrid & non-hybrid Camry comes down to the extra 1% thermal efficiency of the dedicated hybrid engine (41% vs. 40% due to higher compression ratio), and the power-split CVT as opposed to an 8-speed auto. Second off, are you sure there’s no difference in pumping losses with and without a butterfly throttle valve? I’ve read elsewhere that removing the butterfly valve does indeed significantly reduce pumping losses. Even the Wards link I posted above strongly implies pumping losses are reduced. WardsAuto: “[Valvetronic] eliminates the usual ‘butterfly’ throttle valve — and with it a major cause of pumping losses.” As mentioned above, the 2.5L engine in the non-hybrid Camry does have the ability to run in Atkinson cycle, so pumping losses would at least be reduced that way without the butterfly throttle valve. For the record, I now have an ICE only as a range extender for my PHEV, and that is my… Read more »

Final comment – peak thermal efficiency is generally around 2000-2500 rpm at almost full throttle. US cars never really run at this point – it’s way too much power for steady cruising and if you floor it (e.g. for a steep hill) the automatic transmission will downshift and your engine rpm will jump far above 2000.

This article has a simplified generic engine map. It also has a chart with cruising speed efficiencies for a Honda Civic and Ford Focus. I don’t know which Civic they tested, but it’s unusually efficient at low speed. Even the Focus is more efficient at low speed than most US cars.


Below is the thermal efficiency graph for Toyota’s new Dynamic Force engines at different engine RPM and engine Torque.

comment image

Graph found on this webpage:

Great find, Fonzie. A couple notes. This map only goes up to 3200 rpm, which is fine because that’s the high efficiency region. But the engine redlines at 6800, so it’s just the left had side of the full graph. Also, this graph shows a max torque of 180 Nm, which doesn’t match the graph below it and the spec table value of 250 Nm. I thought perhaps they grabbed a graph from the 2.0L engine by mistake, but a look at that page shows similar strangeness: If we use the Model 3’s 60 mph cruise power requirement of 12 kW, then 2000 rpm = 57 Nm of torque. That’s ~30% efficiency on that engine map. At 12 kW, 53 mpg = 31.5% efficiency while the LE’s 39 mpg = 23% efficiency. Hmm. My 12 kW may be low, after all the Camry lacks the Model 3’s smooth underbody. On the other hand, the EPA highway test averages less than 60 mph. So it’s possible that engine map is for the Hybrid. Anyway, both engine maps will have similar peak efficiency, but the Hybrid’s map will have high efficiency at much lower torques. The regular engine DOES use Atkinson-style… Read more »
I’m pretty sure P-Pullyu is correct in the general assertion that ICE engines (let’s say gasoline engines) have an efficiency peak at some, moderately high RPM. See Wikipedia – Engine Efficiency : ” Engine efficiency peaks in most applications at around 75% of rated engine power, which is also the range of greatest engine torque (e.g. in most modern passenger automobile engines with a redline of about 6,000 RPM, maximum torque is obtained at about 4,500 RPM, and maximum engine power is obtained at about 6,000 RPM).[citation needed] At all other combinations of engine speed and torque, the thermal efficiency is less than this maximum”. In addition, his/her reference to ≈20% efficiency when “used to power a car driving down the road” simply acknowledges that normally varying engine speeds will cover a gamut of efficiencies which necessarily average out to less than the peak value. This is fundamental and not open to dispute (although the actual 20% number may be). As for transmissions, their purpose is to shift he torque curve around to crudely match conditions at the currently operating wheel rotation speed and torque demand. As such, there may be a side effect of improved average efficiency over, say,… Read more »

It’s usually closer to 50% of max power than 75%. Peak efficiency is close to full throttle but at a fairly low RPM. That’s not 75% of max power, except for engines with low redlines and/or poor airflow.

MikeM said: “. . . simply acknowledges that normally varying engine speeds will cover a gamut of efficiencies which necessarily average out to less than the peak value. This is fundamental and not open to dispute (although the actual 20% number may be).” Actually that “gamut of efficiencies” for Toyota’s new Dynamic Force Engines is a decrease in thermal efficiency of ONLY 1% to 2%, from 40% to 39% or 38% over practically the entire RPM range of the engine, so long as the engine torque (as opposed to wheel torque) is 110-Nm or more. IIRC, Toyota has mapped the their transmission shift points to hold a gear so that the engine torque and RPM stays in the 38% to 40% thermal efficiency range while cruising at highway speeds (i.e.: low engine RPM and transmission in overdrive or high gear). So in other words the Toyota DF engine operates at “near peak” thermal efficiency over practically its entire RPM range so long as there is a sufficient demand for torque (i.e.: not while driving down a long fairly steep decline). Below is a Toyota graph that reads like a “contour map” showing the thermal efficiency of a Toyota Dynamic Force… Read more »

Note my comments above. The graph shows less than half the engine’s rpm range. And 110 Nm is almost twice what you need to cruise at 60 mph.


You’re just claiming number that are achieved in a relatively short span of use.
Most ICE still run when you’re braking, or idling at a stop, a street light, or when park just running the AC.

So 20-25% REAL efficiency is probably close to the actual figure.

You might achieve your 40% at a steady highway speed in a hypothetical empty road with no traffic jam ever and leading to heaven, but in reality, it’s not happening often.

Timothy Hughbanks

I would be interested to know what “efficiency” means in this context. For example, does near-100% efficeint mean “near-maximum Carnot efficiency” – i.e., nearly as efficient as that allowable by the 2nd Law of thermodynamics, or does it means all the energy of the combustion reaction is converted into work – which, of course, is thermodynamically impossible. Of course, this is the huge fundamental advantage of battery-powered electric vehicles (or even fuel cells) – Carnot efficiency is almost irrelevant.


I read this whole thread and I just want one question answered. What is the efficiency and MPG of an engine sitting at a red light vs. any BEV?


It’s been about 15 years that an engine can stop when sitting at a red light. I don’t even think that a new car can be sold without this system in most European countries.


only 6% of the batteries had to be replaced, all from 2013 and all within warranty. If one reads the mainstream media it may seem that tesla’s cars are often on fire and their batteries die on a regular basis. Most people I’ve talked to about electric cars know very little about EVs but are already worried that the batteries won’t last. The propaganda (oil, competition, shorts, UAW, etc) works, at least for now, hopefully in time people will be more and more informed and choose accordingly.


Well made video.
Thanks for sharing.
Another reason why cellphone batteries don’t last is the SOC voltage. The BMS typically allows 100% at 4.35V per cell and 0% and less than 3.0V. Very bad for battery life.

BEVs are set to 4.20V or under for 100% SOC and around 3.3V for 0%. Much gentler.


It’s a great video. My only quibble is the suggestion that seat heaters should be cut back in order to extend range. In fact, seat heaters are the best way to extend range instead of using the very wasteful cabin heater. Seat heaters use very little wattage, and are direct body heaters, rather than heating the air.


Au contraire, freezing your ass off while also freezing your face is the best way to increase range 🙂

Empire State

Arguably, the shivering-distracted human operator may negatively impact range.

Bill Howland

These videos always make the comment that unlike GM cars with a separate heater for the Battery, Tesla just makes the heat by turning down the efficiency.

That is true enough, but the dinky little resistance heater in the VOLT and Bolt ev is such a trivial piece of an insertion heater that it is hardly worth mentioning. The two companies warm up the Glycol, by modifying the controls in the Tesla, and by turning on a very dinky heater in GM’s case… It is such a small item that it is hardly worth mentioning. I prefer the GM system since it requires no change to the parameters of the car’s essentially ‘Drive By Wire’ system. I’ve never seen much of a safety issue with the GM system – and the way they continue to do it allows an unmodified propulsion system.

Empire State
Bill, I agree that the magnitude of the operational efficiency difference between the different manufacturers’ coolant heating systems as a result of the heat element used here may be small compared to the amount of attention given to the systems’ difference in the video. I, however, prefer Tesla’s approach. In keeping with “the best part is (the absence of) a part,” heating the fluid with an existing required component rather than an additional component added for that express function has advantages. It also reveals a different kind of thinking going on. Rather than maintain, as you say, an “unmodified propulsion system” and add a heating element requiring its own simple control, Tesla’s approach considers the arguably higher reliability solid state control system as that which should be made more capable rather than adding another item from the ever-improving but still fallible “hard parts bin” that auto engineers are accustomed to using. Not to exaggerate the importance here, but it’s more akin to the way a spacecraft might be conceived rather than the way a car’s propulsion system is usually conceived. I know an engineer who was once chastised(mildly) for asking about the specifications of a satellite’s 12v bus when presented… Read more »

Unfortunately, the representation of the module architecture is pretty off. We know that the coolant pipes in the Model 3 modules do not actually snake back-and-forth, but rather just run in one direction. We also know that no cells are flipped in the Model 3 modules. These are two major departures from the Model S architecture.

Also, while it’s hard to find any reliable information on this, the consensus from most sources seems to be that discharging below 30% doesn’t damage the battery. (The last couple of percent probably does, though.)