Tesla VS. GM: Battle Of The Battery & Electronics Cooling Systems


Tesla’s system is more complicated but adds flexibility, more operating modes and failure back up

We just shared another good video from Weber Auto and John Kelly detailing the Chevrolet Bolt EV cooling systems.

Tesla, like GM, has two main cooling loops: one for the battery and one for the high voltage power electronics. The Bolt EV’s system is simple. The two cooling loops operate independently. They are not connected. Tesla’s system in both the Model 3, S, and X allow the motor/ power electronics cooling loop and the battery cooling loop to be connected in series via a fancy four-way valve (ref and see figure below).

The fact that there’s an additional valve connecting the two loops in series OR in parallel is a bit more complicated. However, that added complexity allowed Tesla to eliminate the battery heater in Model 3. With the two loops connected in series they can use waste heat from the motor and power electronics to heat the battery in cold weather. GM cannot, but as my partner Keith Ritter has pointed out it could be easily added to GM’s system via the addition of another heat exchanger between the two loops.

On the other hand, GM has a whole dedicated glycol loop for cabin heating while Tesla does not. So there are three glycol loops in the Bolt EV but only two in the Tesla. Tesla heats the cabin directly with a high voltage resistance heater that just dumps heat into the cabin air. GM on the other hand uses the resistive heater to heat the glycol loop and then uses that warm glycol to heat the cabin air.

Professor Kelly offered some explanation for GM’s using an additional glycol loop and heater in his video @ 28:59: better control. We think there’s another explanation. GM didn’t want the high voltage heater exposed directly to the cabin for safety reasons. (via GM engineer Wop on Tour). So, while GM’s system is simpler in that it doesn’t have the four-way valve connecting the two glycol loops, it’s more complicated because it needs a dedicated battery heater and because it has another glycol loop dedicated to cabin heating. Tesla dumps the heat directly to cabin air and eliminates the glycol loop.

Here’s a diagnostic mode schematic of Tesla’s Model S cooling system operating with the power electronics loop and the battery cooling loop interconnected. In this schematic the front radiator is bypassed because we are scavenging heat from the motor and power electronics:

Tesla Model S cooling systems connected

With the cooling loops connected in series, the glycol coolant goes through the battery first, then the motor and power electronics. Since the Bolt EV cannot run with the battery loop and power electronics loop connected, it can’t scavenge heat from the power electronics. GM’s system is not as efficient.

What other operating modes does connecting the 2 loops allow?

  1. In the event of air conditioning compressor failure, Tesla can cool the battery and power electronics with the front radiator. To picture this mode, look at figure 1 and imagine the front radiator NOT being bypassed. In the Bolt EV, failure of the AC compressor would result in battery over temperature in hot climates. About all GM could do in the event of AC compressor failure is keep recirculating the battery coolant. So, at least GM’s battery would be uniformly hot. Please note that we don’t have confirmation from Tesla that this mode actually exists. We have postulated that does based on Tesla’s schematic only with front radiator NOT bypassed.
  2. Track mode in the Model 3. Again, we are speculating a bit here. Tesla has not verified the specifics but it can be postulated from figure 1. In this mode we DO NOT bypass the front radiator. With the two loops connected, we run glycol first through the front radiator and then through the AC chiller to super-cool the glycol prior to going through the battery and power electronics. In addition, we know that when track mode is switched on, all the fans in the cooling system come on and pre-chill the pack and power electronics.

In summary:

Tesla’s cooling system is a bit more complicated because of the added four-way valve that allows the power electronics and battery loop to be connected together. However, this added bit of complexity opens up additional operating modes that make the vehicle more energy efficient (waste heat used to heat battery), allows elimination of the battery heater and provides back up in the event of air conditioning compressor failure.

Which system is better?

Let us know in the comment section.

Categories: Battery Tech, Chevrolet, Tesla

Tags: , , ,

Leave a Reply

67 Comments on "Tesla VS. GM: Battle Of The Battery & Electronics Cooling Systems"

newest oldest most voted

My understanding is the Tesla vehicles spend tons of energy (less than when S was originally released but still comparatively high) keeping their battery in a particularly narrow temperature band when parked. That consumes far more energy than saved with the added complexity to use waste heat from the electronics.

For some, this is preferred. For others, it’s a lot more electricity they’re paying for. Whether for or against the approach, I wish the EPA had “some” way of listing it on the stickers. Maybe it’s a couple examples of standby energy usage as a function of ambient temperature, separate from the MPGe figure, or some entirely different approach.

But including that across all vehicles versus its current omission would be much better. It would let people make more informed decisions.

That was a common misunderstanding, but the infamous vampire losses have nothing to do with thermoregulation of the battery.

If you leave a Tesla off and outside long enough, the battery temperature will approach ambient temperature, no matter how cold or hot that may be.

With energy saving on, my S will use about 1 kWh per day recharging the 12V battery. This is due to running the onboard computers and modem.

So about 12c of electricity a day, so not a great loss, though they did work at that as it was much worse at one time.. It’s true what you say about ambient temperature, since when the battery is cold you get warnings about regeneration not being available, later as the battery warms up, that warning goes away.
What I do, during cold weather is put my battery on a charger once a weak for an hour, mostly just to check its level, or state of charge. There seem to be more problems with the 12v battery, maybe because they sit around too long before they are put in the vehicles. You want to get those batteries installed with a month or so after they come out of the factory. Don’t every buy a battery from big box stores, where they may be sitting on the shelf for months.

I’m still surprised that an EV needs a 12v battery is it a lead acid battery.

Neither EV nor hybrids need lead acid battery. It is legacy from old times.

Hyundai Ioniq Hybrid eliminated it. Although it still has separate section of Li Ion traction battery dedicated to 12 V circuits.

alot of the electronics, like power windows, radios airbags, all expect 12v.

Right, and that holdover may last a long time. Similarly, we’re still using QWERTY keyboards altho Dvorak keyboards allow faster typing. Inertia applies to economics and customs, not just to physics.

Tyrany of the installed base.

1kWh / 24 hours = 42 watts. That’s like my laptop on full blast. Tesla really should work on power management aspect of this. Use a netbook, drop it down to 5 watts.

It is less than 42W, but the recharging of the 12V battery is inefficient. It might take 1 kWh from the main pack to add 0.25 kWh to the 12V in several increments.

No way. It’s charged with a DC/DC converter. Those are over 95% efficient. The one that charges my converted car is.

Apologies that I incorrectly equated the energy use to battery TMS, but I think the concern remains valid. The fact you use 1kWh per day with energy savings turned on remains disconcerting to me. That’s a large amount of power usage over a year or over the fleet. And it’s even higher when energy saving isn’t on.

To add some perspective… With 200,000 Tesla sales in the US so far, they consume 200,000kWh every day just sitting parked. That’s just US sales, and assumes they ALL have the energy savings settings enabled, so a “best case” number of how much energy is being wasted.

Put another way, every day enough energy is wasted to provide roughly 200 people with the power they would typically use/need over a full year.

That’s pretty terrible for an “environmentally friendly” vehicle and brand. 200MWh per day, or 73GWh (73,000,000 kWh) per year! Awful.

Yeah, and what if those 200’000 “terrible” and “awful” Teslas were petrol cars at just 8 l/100km? It takes 3 kWh of energy to refine 1 l of petrol, so your fleet would eat up 10 times more electricity on top of its consumption. (15000 km/yr, 8 l/100km, 200’000 cars = 720’000’000 kWh) I guess that is a pretty decent win for EV, Eric, even if I agree that 42 W is too much and that figure should be 10 W or less.

Sure, it’s better than petrol cars, but it’s worse than every other EV. Tons of Tesla people here tout the MPGe of the Tesla over other EVs, shouldn’t we strive for less energy use with Tesla too? If our bar to compare against is gas cars, well, that’s not very ambitious IMHO.

Also, as an aside, unless that number is updated, the energy to refine petrol includes heat energy that comes as part of the refinery process, rather than pure electricity input. I still agree that it’s better than petrol/gas cars (when those cars are running, when they’re off they consume no gasoline, infinitely less than the energy Tesla consumes), but the bar I was comparing Tesla against was other EVs.

Eric, even petrol car batteries drain when the car is parked. They don’t consume fuel while parked but as soon as you start the engine the alternator starts charging the battery. So in the end extra fuel is consumed to offset the phantom drain. I do agree that trying to lower phantom drain is in everyone’s best interest.

Petrol cars only use a small amount from their 12V battery. We’re talking 100Wh or so regardless of whether the car is parked 1 day or 1 month.

I just wish Tesla employed the same standard practices. I’m glad you seem to agree, but it still sounds like I’m in the minority given the voting.

Let’s put that in perspective:

If the worldwide fleet of ~1 billion cars consumed just 1 gallon per day, that would use ~33.7 billion kWh worth of fuel per day, not per year… of which about 70-75% is wasted due to the thermal inefficiency of gasmobiles.

If the average car “wasted” as much energy per day as a Tesla car, then the world would be much, much better off.

Perspective?! HA! Let’s talk perspective.

For starters, Nobody here likes gassers. That aside…

All the ICE cars in the world use ZERO gallons of gas (or energy) while parked and off. Except Tesla.

All other EVs use ZERO kWh for running computers. Except Tesla.

Don’t equate apples to oranges. You can’t simultaneously boast about Tesla’s high MPGe compared to other EVs and then ignore this energy consumption. That’s ludicrous. It’s as silly as EV naysayers claiming Tesla cars are useless because they can’t go as far as majority ICE cars on a charge.

Cherry picking data for or against a vehicle is silly. Not that I expect anything less from you.

“All the ICE cars in the world use ZERO gallons of gas (or energy) while parked and off. Except Tesla.”
“All other EVs use ZERO kWh for running computers. Except Tesla.”

Strictly speaking These two statements are not true. All recent vehicles continue to use power to keep radio settings, the clock going, newer automatic transmission settings, remote entry systems ( key fobs ) and alarm systems operating. The 12 volt lead acid batteries also have a rather large self discharge. All this power in an ICE car needs to be replenished by burning gas on start up.

Yes, they use some tiny amount of power limited to a couple hundred Watt hours or less, and have their systems completely shut down to prevent further drain. So if they are parked for a day or a month, it’s still a tiny fraction of what a Tesla consumes every day.

I just hope Tesla eventually takes that energy consumption more serious as they gain increasing market share.

Evaporation still exists, as do leaks.

Tesla has eliminated the battery heater in the model 3. So connecting the 2 loops is the only way tesla can heat the battery. Tesla has figured out a way to energize the motor even when the the car is parked. The motor IS the main battery heater. Re cycling wasted heat from the power electronics is a side benefit.

Thanks George, and my mistake. I’m still concerned about the amount of energy a Tesla uses when parked and “off” but it is clear now that the cause is more the computing left on in the vehicle rather than the TMS (and therefore a less appropriate digression for me to make on this thread).

Multiple modes of resting as well — idle, sleep, deep sleep. 3rd party apps like TeslaFI.COM have options to allow your car to go into a sleep (vs idle). This is working well on my son’s Model 3.

That’s good to know, Scott. It should still be the norm. I really expected Tesla to have an awesome user experience WITHOUT all the vampire drain by default. Musk can put rockets in space, he can’t figure out a way to get the functionality AND not use energy?

As an embedded systems engineer myself (turned program manager now, but I digress), I know this is possible – without sacrificing functionality – through proper coding practices and embedded hardware implementations. Which is perhaps why I am critical of the lackluster performance on Tesla’s part in this (small) regard.

Call Elon he loves problem fixers.

I think to answer the question about which one is better I think you have to understand why GM and Tesla went the paths that they did. Maybe the electronics in Bolt don’t get hot enough to warrant the added complexity of scavanging heat. With the Volt GM did more mixing of the cooling systems due to the ICE being involved. So, unlike Tesla that uses the same type of system for every vehicle it looks like GM is using the system that works for the vehicle at hand to manage cost and complexity. Tesla on the other hand uses a similar system on all their vehicles to manage cost.

In the end I don’t know if one system is better than the other – they are just different. We need to look at Kia/Hyundai, Jaguar, Audi, VW, etc… to see what trade-offs they make.

On the GM system the heater for the cabin loop is outside the car cabin, in the engine compartment, like it would be in a gas engine. This means GM’s system is heating up the world in addition to the cabin. Not efficient. Go cheap, design twice. But, GM will probably cancel this car too at some point soon.

And you can consider the electric resistance heater in the Tesla a Luxury Item as it allows for rapid ramp up in temperature, far faster than a glycol loop.

However, it is likely shared with other ICE cars, so maybe cheaper than you think (using something different when the car is built on an ICE vehicle line might be more expensive).

I think the GM heats the Glycol loop very rapidly. My Clarity PHEV has a similar heater and is blowing hot air within half a block of my house. It really is a safety consideration as an air to air electrical resistance heater could be a fire hazard in an accident, although a PTC heater should reduce risk there (they don’t get as hot). The Clarity BEV/FCV use a PTC heater heating the air directly as far as I can tell.

So luxury! It works fast and so does the air conditioner. They both seem great compared to idling a truck.

I haven’t heard of anyone complaining about how quickly the Bolt heats up. Mine appears to head up very quickly. You’re talking about a very small amount of liquid. Also, electric resistance heaters can be fire hazards in the cabin. Dust can accumulate on them and pose a fire risk. Not saying they would, but you’re increasing the chance.

Hmmm, I converted a car to electric over 9 years ago. It has two PTC heater cores in it. I just bought two 1.5kW heaters at Lowes and removed the cores, installed them in a frame in place of the original heater core. I live in high desert with a gravel driveway about 200 yds long, and lots of dust. No fire yet.

According the the Bolt’s Owners Manual, the Bolt’s cabin heating glycol loop contains about 2.1 quarts of glycol. That is about 5 pounds. I haven’t found a hard firm rating, but several people on a Bolt EV forum estimated the heater to be between 7.5 and 9 kW based on measured-kW demand during pre-heat and a 30A/350 VDC fuse, plus info from a “knowledgeable GM employee”.


Assuming the heater goes to full heat of 7.5 kW during warmup, the loop would warm from an ambient of 40F to 120 F to get substantive heat out of the heater core would be in just under 1 minute. That is pretty darn fast to me and is faster than the heating loops in all current ICE vehicles, which have to warm up the entire engine block and several gallons of glycol to get to heating temperature.

Calculations for heating time are “ideal” and ignore other realities of the system. There’s other valves, hoses, etc. that need to be heated beyond that Glycol, and they require more energy to heat up per unit volume than the Glycol requires.

Still, I agree it’s faster than most standard ICE vehicle heating loops.

The heating loop includes the other stuff. The heater ‘radiator’ is in the conditioned space therefore is not the loss everyone here is suddenly worried about.

I have a 2018 Bolt and a 9 year old car I converted to electric which has 2 PTC 1.5kW heater cores in place of the original heater core. The converted car is blowing out warm air 10 seconds after the heater is turned on. I’ve counted 1000, 2000, … a number of times to estimate this. The Bolt takes longer, but it is fast enough. Not an issue.

Yeah you got that right. Of all the silly things to worry about, this ain’t one of ’em.

Just head temps, coolant exits from the cylinder heads.

Well said, Bill711Coffee!

Very strange that you got some down-votes for your well-informed and very relevant comment. O_o

That is funny. The reason he got the down votes is the ‘outside heater loop’ loses a trivial amount of heat, and if anyone is worried about it, they can wrap a bit of fiberglass insulation around the hose. But the heat loss is trivial.

It is nothing compared to the typical vampire drain loss.

The LEAF’s battery cooling system is the least complex, most reliable and needs no failure modes what so ever. It’s clearly the best.


I think the simplest system that works well enough will win. The Leaf’s is fine for many climates, but not hot climates. To be fair, the Leaf’s system will be less trouble used where you might end up with coolant leaks from the Bolt EV or others. Living up north the Leaf system might be a better bet, but will never work well for DCFC.

As was pointed out in the Weber Auto video these coolant systems are a lot different than what’s on an ICE vehicle. They are low pressure (5 PSI) systems. And the temperatures are nowhere near the hundreds of degrees ICE cooling systems have to deal with. Coolant leaks should be rare given the pressures/temperates and they should last the life of the car.

That is promising, but it is still an added complexity. As you say, an old BMW might have brittle plastic coolant pipes in a large part as it is dealing with much greater temperature extremes. Still, many issues with the Volt have been coolant loop related (and nothing to do with thermal).

That’s very funny, Derek! Thanks for the laugh.

The LEAF battery pack wilts in the HEAT of the Southern USA. 10-20% a year. It’s terrible and the newer lizard chemistry doesn’t help. The capacity is lost forever. Very poor. Some day LG who now owns the battery system will have to fix it for future LEAF.

Derek’s joking (and pretty funny joke at that). If there’s no coolant, there’s no coolant issue. Same is true with Bolt, simply drive it without coolant.

Many years ago, I had a guy working for me doing some chip work. Some things didn’t fit, so he just commented them out. Well, it easily met timing and fit well since the functionality wasn’t there. I laughed so hard, I thought I was going to die.

You must be joking. Are you?

The LEAF is a good exemple of what not to do with a powerful electric car battery.
It loses an insane amount of range in cold climates and it deteriorates abnormally in hot climates.

If you’re outside these extremes, the absence of battery temperature management is less debilitating. Until you use a fast charge station…

I read the comment as highly sarcastic 😉

Yes, he’s joking. Some comments need a smiley to alert the reader, but that one didn’t need one.

Heh. Oh, the Leaf’s “cooling system” needs failure modes all right. It just doesn’t have any, as Rapidgate shows.

Actually the leaf has a cooling loop to dissipate hear from the charger and controller I had to replace one of two pumps on it to keep it going

Direct air resistance heaters tend to smell when they first kick on due to the high temperature surfaces on the coils. Water to air heaters have lower surface temperatures and so don’t tend to smell.

Yep, but if you design them with larger surface area they can run at lower temperatures. It is just a matter of design.

Does this mean that Tesla cannot use excess heat to heat the cabin? That doesn’t sound great for operating in a cold climate. Or is the heat waste very small?

Tesla production vehicle does not use waste heat for cabin heating. Tesla’s PATENT did though.
It’s explained here:

Interestingly, the patent included a glycol loop for the cabin JUST LIKE GM. but that glycol loop was eliminated. So that’s why HVACman has suggested that GM could easily add it to the BoltEV. All we need is a heat exchanger between the power electronics loop and the cabin loop.

I think the author forgets there are 5 different ways to accomplish the same thing. I have no problem with either system’s performance. A standard hydronic heater in GM cars seems to me to be just an economization since the same part could also be used in ICE products. The fact that the BOLT ev works with just 2 small water pumps and only one brushless fan motor seems to me pretty good. The suction gas cooled 3-phase inverter and motor on the air conditioning compressor (semi-hermetic) makes for an easily installed/serviced unit – only 2 power wires, and, ideally it should last longer than on ICE cars since there is no mechanical drive shaft seal to leak refrigerant. The one ‘performance’ issue I have not seen adequately investigated is, what is the ‘vampire drain’ in Tesla products during cold weather? (In GM products from 2010 the VPD has been almost non-existent – I say ‘almost’ since the car wakes up every few days to recharge the 12 volt battery if needed). In cold (0-10 deg F) temperatures the first model S’s were a continuous 1840 watts… (such as NYT’s Broder’s test drive). That is so much the ideally 1400… Read more »

I’ve seen some say that datalogging apps like Teslafi cause some of this because keep “waking” the system when it should be in sleep mode.

That may explain a bit of it, but the early ‘s’s drew 1840 watts. That is a huge amount of ‘vestigial’ power to be constantly drawn. I assume it had something to do with keeping the battery warm, and wouldn’t be surprised if the more modern Teslas had insulated battery packs. And, NO, the cabin heater was not on.

When cooling systems battle… is that anything like a mud wrestling competition? 😉

“Tesla’s cooling system is a bit more complicated… However, this added bit of complexity opens up additional operating modes that make the vehicle more energy efficient…”

To me at least, that’s the overriding factor. The system which helps one car be more energy-efficient than the other, is the winner. Better energy-efficiency means better range and faster charging (as measured by miles added per minute of charging).

Well one thing we know for sure is that the Tesla’s can DCFC at twice the kw as the Bolt so that is evidence that the Tesla TMS is better at cooling during high kw DCFC.

That’s why you pay twice the money for a Tesla, you get twice the cooling system.

It also puts the lie to those people who say that cars should fast charge at 1000 kw, or 2000 kw or higher – explaining that it is trivially easy to do. I say, ok do it!

In practice in such a case I’d think it would be cheaper to have the refrigeration at the Coral, since there is no need for making such systems portable at those high charging rates.

They also assume batterys last 40k miles… hmmm. Related?

Watched the weber auto bolt pack disassembly, imbalances, and hotspots galore! And thats without taking 3 cell modules apart. I presume 1 left, 1 right, and 1 stuck between them not that the plate along side will conduct heat that well. So all cooling on 1 edge.

Also insulation only under the cells leaving long strips with brackets bolted to them to transmit heat out of pack… they’d been better spaying on insulation on the bottom of the plate.

Simple Above & Complex below has long been a winning Formula – Strange the article was written by one of the Subjects – GM ?