MASTER THREAD: 2021 Model 3 - Charge data, battery discussion etc

[Venom]

I can only assume that with every run the battery was heating up more and more. A battery at 35°C cell temperature can still pull well at 30% SoC.

The thing is, that the scenario of low SoC and high battery temperature is not a common situation in day to day driving. Usually you start with a warm battery, which then discharges and gets colder with the new octovalve/heat pump.

I have had a situation recently when at -6°C air temperature the battery heater stopped heating up the battery any further than 10°C cell temp. On that route I couldnt drive faster than 120kph, so I could only look at my battery being stuck at 10°C while discharging. Even at high SoC the car was never gonna show its full potential.

This is all probably great for efficiency...not so great for a performance car.

Ah this makes sense! Is there a way to heat up the battery in that situation where you're battery is cooling down?

 

[clerks]

This is all probably great for efficiency...not so great for a performance car.

Correct and anyone buying an M3P didn’t buy it to be neutered at below 50 SOC. The worst part of all of this to me is the 2019 M3P starts beating the 2021 at as early as 70 SOC if you study Bjorns chart. So you get the claimed performance for 20 percent of the range and the rest is inferior.

 

[AlanSubie4Life]

Curious as to why this video shows hardly any performance degradation from 100% to 29%:

Wow...that video. Doing drag runs in an area with cross traffic, uncontrolled entrances, pedestrians, cyclists, and where speed limit is 35mph. It's not even a good spot for a run (not flat, curved, etc.) And leaves his temporary plate number on the dash. So irresponsible. I hope we can all aspire to not be like him.

Anyway, it was 50 degrees and this is in the US, and it's unlikely that the battery was cold at all (certainly at high SoC he charged at the Culver City Supercharger minutes before). It's unclear what he did to get the SoC to 30%, but it looks like it wasn't done on the same day (both days were in the morning...). There are a number of possibilities for getting to that state (for example, maybe he supercharged to 30% from 5%?), which would impact actual pack temperature during the run.

Seems pretty consistent with other reports out there, anyway. Also you'd presumably want to check at 20% or 10% SoC to actually see significant dropoff, if this voltage limit is a factor in the US. The extracted values above from @eivissa are for cold packs and things only became really evident there around 40%. So with a pack that might be 40-50F degrees warmer (no way to know), it's hard to say what to expect.

 

[AlanSubie4Life]

Ah this makes sense! Is there a way to heat up the battery in that situation where you're battery is cooling down?

One way is if you are in close proximity to a high-speed charger (Supercharger or otherwise), navigate to it. Will waste considerable power of course, so not ideal, if you plan to actually get where you're going, and you're running tight margins. Also requires a high density of high-speed chargers to be workable.

 

[AlanSubie4Life]

The worst part of all of this to me is the 2019 M3P starts beating the 2021 at as early as 70 SOC if you study Bjorns chart.

*At a battery temperature on the 2021 that is colder by 10-20C (~20-40F), FWIW. But for full disclosure, as previously pointed out, you can see that at 10% SoC in this table the 2021 is warmed up a bit (but still 15C colder), and still does extremely poorly when compared to the 2019 (and compared to the LR 2021 at similar temp).

 

[Venom]

One way is if you are in close proximity to a high-speed charger (Supercharger or otherwise), navigate to it. Will waste considerable power of course, so not ideal, if you plan to actually get where you're going, and you're running tight margins. Also requires a high density of high-speed chargers to be workable.

So basically there is no practical way of doing this?

If we find out that the cold temps are to blame for the degradation in acceleration, Tesla could resolve this via a firmware update to warm up the battery when it reaches a certain temp threshold, right?

 

[MasterC17]

I've done comparable tests today @ 30°C Cell Temp Mid with the same results:
-Base voltage at 30% and 20% SoC is identical with Björns video.
-Voltage Sag is identical at all SoC below 40% -> A short burst down to 280 and then steady 295V lower limit.

View attachment 637247 View attachment 637248

Interesting. The data I posted was done at 24C. Pack voltage was roughly the same (340V), but my max discharge was 320kW; the voltage dropped to 278V. What speeds were you driving from/to? It is normal for the pack voltage to climb back up as your speed increases over 100kph due to back-EMF (e.g. it should not stay that low, so it may be completely normal).

Either way, a difference of 40kW isn't that big, but I would be curious to know why. Had you done several runs before that one?

 

[AlanSubie4Life]

So basically there is no practical way of doing this?

If we find out that the cold temps are to blame for the degradation in acceleration, Tesla could resolve this via a firmware update to warm up the battery when it reaches a certain temp threshold, right?

I assume that if this actually persists as a problem and Tesla deems that these batteries will always be worse than the prior batteries in cold conditions, after their analysis of performance in the field is complete, an easy fix for them (if someone brings it to Elon's attention or whatever, or their engineering team decides it is a problem), would be to ensure that Track Mode provides the option of engaging the stator heating (probably unnecessary), or cuts off heat scavenging from the battery (probably sufficient in most cases). As with most Track Mode configuration, it could be fully customizable by the user.

That wouldn't solve the practical limitations that currently exist on surface streets (you can't responsibly use Track Mode there), but I'm assuming there might be a general alignment of the battery performance over time, and this Track Mode solution would close the gap on the prior version of the car (there's no way to avoid that difference even if the pack were identical), which didn't scavenge waste heat from the battery for cabin heating.

It terms of firmware upgrades - what you suggested would affect efficiency results, so I have my doubts about whether they would do that. Cuts both ways however, since a colder pack might cut out a little sooner in the EPA test and provide less energy, so maybe there is room for optimization here, and make it a wash. Who knows.

If I had to guess: I think it's likely that things will get substantially better than they are now, but never perform equivalently to the 2019 Performance in cold weather, because the battery will always be cooler in cold conditions with the 2021, due to the methods used to improve efficiency. It might be so much better than it is now that it doesn't really matter though. (For example: would anyone really care if your 0-60 is only 6 seconds at 10% SoC, when you aren't in Track Mode (there are some claims that Track Mode allows greater discharge rates at lower SoC)?) On the upside, I'd imagine that hot weather track performance/heat soak could be alleviated by perhaps running the cabin heat during your track runs (turn it into a real sweat box!). Someone will probably try that this summer I guess. (Whether this would help depends on what is the limiting factor/bottleneck for heat soak on the track.)

 

[MasterC17]

Correct and anyone buying an M3P didn’t buy it to be neutered at below 50 SOC. The worst part of all of this to me is the 2019 M3P starts beating the 2021 at as early as 70 SOC if you study Bjorns chart. So you get the claimed performance for 20 percent of the range and the rest is inferior.

Just to keep things in perspective, the Performance 3, and all EV's for that matter - have always been "neutered" at lower SOC. As the Voltage drops your maximum discharge power drops (Current = V/R). The difference from 100% to 80% isn't that big, but anything below 70% there is a noticeable difference and it keeps that curve as the SOC drops lower. This isn't new to the 2021.

Temperature is really the next most important factor when it comes to power. The ideal temperature for the Model 3 is 80f-120f. Bjorns chart is misleading, because the battery temperature is well below ideal even at 70% in his testing. In my 2018 at 25% SOC, the Max Discharge is the same as the 2021.

So, to be clear, there is still no evidence to show that the 2021 has less power at the same SOC as the 2018-2020 when both have a warmed up battery pack.

On the other hand, there is evidence to support the theory that the 2021 Model 3 Battery Pack cannot sustain ideal pack temperatures when ambient temperatures are well below freezing.

It is still possible that the NA and EU versions of the 2021 are running different firmware. However, I think it is far more likely that the concentration of Model 3 Performance buyers in the EU are driving in comparably colder weather. Also keep in mind a lot of people on the US are not driving nearly as much as typical.

FWIW, as previously mentioned, you can punch in a Supercharger as a destination to heat the battery pack up. This does consume more energy, but it would also keep the pack warm. So, you're basically giving up the extra efficiency of the 2021 over the 2018-2020 to match the performance of the 2018-2020 in sustained driving in cold temperatures.

I do agree that it sounds like this could all be resolved with a Firmware update, but I think it is important not to confuse normal and expected operation/behavior with the actual concern we have with the 2021.

I may do a sustained drive in cold weather to see if I can duplicate the concerns coming out of the EU.

 

[clerks]

Just to keep things in perspective, the Performance 3, and all EV's for that matter - have always been "neutered" at lower SOC. As the Voltage drops your maximum discharge power drops (Current = V/R). The difference from 100% to 80% isn't that big, but anything below 70% there is a noticeable difference and it keeps that curve as the SOC drops lower. This isn't new to the 2021.

Temperature is really the next most important factor when it comes to power. The ideal temperature for the Model 3 is 80f-120f. Bjorns chart is misleading, because the battery temperature is well below ideal even at 70% in his testing. In my 2018 at 25% SOC, the Max Discharge is the same as the 2021.

So, to be clear, there is still no evidence to show that the 2021 has less power at the same SOC as the 2018-2020 when both have a warmed up battery pack.

On the other hand, there is evidence to support the theory that the 2021 Model 3 Battery Pack cannot sustain ideal pack temperatures when ambient temperatures are well below freezing.

It is still possible that the NA and EU versions of the 2021 are running different firmware. However, I think it is far more likely that the concentration of Model 3 Performance buyers in the EU are driving in comparably colder weather. Also keep in mind a lot of people on the US are not driving nearly as much as typical.

FWIW, as previously mentioned, you can punch in a Supercharger as a destination to heat the battery pack up. This does consume more energy, but it would also keep the pack warm. So, you're basically giving up the extra efficiency of the 2021 over the 2018-2020 to match the performance of the 2018-2020 in sustained driving in cold temperatures.

I do agree that it sounds like this could all be resolved with a Firmware update, but I think it is important not to confuse normal and expected operation/behavior with the actual concern we have with the 2021.

I may do a sustained drive in cold weather to see if I can duplicate the concerns coming out of the EU.

For us noobs what do you make of the post from this am that the voltage sag was the same at 30C? Also, I have read separate from this issue that the 2021M3P has increased performance loss across various SOC versus older models. I believe in researching the 2021 and the various curves on power loss, the 2021 seemed to do worse overall presumably due to the heat pump?Have you experienced this?

 

[MasterC17]

For us noobs what do you make of the post from this am that the voltage sag was the same at 30C? Also, I have read separate from this issue that the 2021M3P has increased performance loss across various SOC versus older models. I believe in researching the 2021 and the various curves on power loss, the 2021 seemed to do worse overall presumably due to the heat pump?Have you experienced this?

I need some more information from that user before drawing any conclusions. Voltage should not sag to a sustained level, and he and I saw the same low voltage points. It will drop immediately to some number, slowly continue to decline until you hit 60mph, and then increase as you exceed 60mph. So, if that user was doing pulls from 40mph-80mph you would see a sharp drop and then a rise as described in their post.

I don't think the heat pump should have any affect on power. Its only real function is cooling/heating and arguably has no measurable affect on battery current draw from the drive units.

It's a wee bit comical because just 1.5 months ago I had several people throwing a fit that I had claimed the 2018-2020 and 2021 had the same exact power and they said the 2021 must be more powerful (for several random reasons). Now, we have people claiming the opposite. The only common thread is that both opinions aren't using a lot of facts to support their claims. Obviously, the more data the better, but we need the full scope of that data, not just pieces of information. I stand by my data which indicates the power between the two at the same battery temperatures and same SOC's is indistinguishable.

Again, that is not to say that the net result to the end user is not a different experience, and obviously if the battery pack is getting colder while driving in the 2021's over long distances than that is something we want to solve, but we should be focusing on what the root cause is.

 

[AlanSubie4Life]

For us noobs what do you make of the post from this am that the voltage sag was the same at 30C

As was mentioned earlier, crudely constructed approximation of what is happening:

Vmotor = Vopen - Imotor * Rint

Pmotor (input) = Vmotor * Imotor

Vmotor is what we see in these data sets. You can't see Vopen unless the battery is open circuited - it's internal to the battery. But it will be related to the SoC.

Rint is the internal battery resistance.

So: if Rint is much lower at higher temperature (and it might be, if you can get those electrons pried off those lithium atoms much faster at high temperature), you can have a higher current Imotor and achieve the same Vmotor at the same Vopen (Vopen being the open circuit voltage - the unloaded voltage corresponding to current SoC).

@eivissa would have to tell us whether the indicated power also aligned with Bjorn's videos - but it doesn't look like it. He indicated 278kW at 28.5% and 238kWh @ 20%, which compares to 220kW and 154kWh in Bjorn's videos at similar SoCs, but I'm not sure these numbers are directly comparable (he could clarify whether they are).

And just as a reminder, you have to remember there is not a direct relationship between 0-60 time and peak power, because peak power (when it's at its maximum at 100% SoC) is only reached at ~45mph - so you don't actually (theoretically) have a hit on your 0-30 time, even at pretty low SoC, since you're not power limited at those speeds (e.g. it requires approximately zero HP to launch massively hard in a 3P+ - you just need the torque!). It's a bit more complicated than that - you need to be able to deliver a lot of current to get that initial torque too, but those are details which we can gloss over here, as it is off topic. Anyway, there are plenty of curves out there showing power vs. speed at different SoCs, so you can look them up and verify all the details of how SoC affects power for yourself. To the point here - it would be much better to see that data for all these cold weather runs, rather than the simple table below, because you could see the overall effect of temperature throughout the run, and we might learn something.

 

[Venom]

This guy just uploaded a video showing 0 - 60 times from 15% to 10% SoC (2021 M3P). Performance degradation isn’t THAT bad at 14c compared to Biorns low temp results. This is a US model.

 

[clerks]

This guy just uploaded a video showing 0 - 60 times from 15% to 10% SoC (2021 M3P). Performance degradation isn’t THAT bad at 14c compared to Biorns low temp results. This is a US model.

Agree. These results seem in line with what one might expect at those low SOC and it wasn’t that warm outside either. I bought an Audi S4 in 2000 that did zero to 60 in 5.5 seconds and I thought that was crazy fast. I guess an M3P pulling sub 5 second zero to 60s at 10 percent SOC is acceptable 21 years later. Thanks for posting this. I’ll sleep better tonight.

 

[AAKEE]

Tesla could resolve this via a firmware update to warm up the battery when it reaches a certain temp threshold, right?

The whole idea with the heatpump + octovalve is to save energy and increase range in cold climat.

The whole idea is to use battery ”waist heat”. Its waist heat in a technical term even if we sometime lile the battery to keep warm.

Any firmware(or other) change to this will considerably reduce range and increase the charging costs. I dont care that much, but I understand Teslas work for energy effectiveness.

I have the ’21 M3P and my heatpump has the known problem with a defect sensor. Sometimes it doesnt work and the heatpump compressor runs in the emergency mode( which is direct 1:1 heat production as the old model 3).
With about -10 to -20 my normal consumption when driving to work(50km drive) is about 170-180 wh/km but when the heatpump goes into the emergency mode( can easily be heared or felt as vibrations in the steering wheel) I get 230 to 240 wh/km for the exact same drive. The difference in consumption comes from producing heat with 1:1 efficiency instead of heatpumping heat from ( among other) the battery.

So we can not rely on Tesla changing the software so it never cools the battery, that will not happen.

But as it is a performance model, it would be possible to get a setting like track mode that not use the battery heat( inhibit octovalve from stealing battery heat. Also, a “battery heat” for-performance-mode function would be great. If everything of these settings/modes are put in the track mode settings( not neccessary needing to activate track mode) it will be clear to anyone using it that it comes with increased consumption/reduced range.

By the way, anyone tested the track mode in cold to see how the battery temp behaves ?

 

[AAKEE]

FWIW, as previously mentioned, you can punch in a Supercharger as a destination to heat the battery pack up. This does consume more energy, but it would also keep the pack warm. So, you're basically giving up the extra efficiency of the 2021 over the 2018-2020 to match the performance of the 2018-2020 in sustained driving in cold temperatures.

Put a supercharger as destination requires either having one close enough(within about 20-30minutes) as it doesnt start to heat the battery until rather close. The closest for me is 55 km away and completely int the wrong direction for anything but driving to my mother in law. So, it isnt really usefull to drive 25km in the wrong direction away before supercharging starts.

It seems my ’21 aims for a 12C tempretarute on the battery pack. The heat pump steals heat from the battery down to about 12C, and there it stays if on highwayas. Its the same if the battery already is warmer due to starting from the garage, the battery cools to about 12C.

I gues it cools to 12C and if driving on a highway and the battery warms from heat losses inside the battery this heat also will be used to keep the coupe warm.

As soon as the outside temps go up this spring and we up north dont need heating any more, our batterys will be warmer and we will also get more power at low SOC.

 

[AAKEE]

Actually, I think this have to be told to be clear:

The heatpump together with the octovalve is a fantastic work from Tesla. Its a big step towards longer range in cold climates and it is really good for the future. A warm battery contains a lot of heat that otherwise would be lost and instead the battery would need to supply the heat from its capacity.

—-> We dont want the software to be changed so these good things do not happen for the daily driving. But we would like to be able to deselect it, and prioritize power over consumption at least in the Performance version.

 

[eivissa]

Interesting. The data I posted was done at 24C. Pack voltage was roughly the same (340V), but my max discharge was 320kW; the voltage dropped to 278V. What speeds were you driving from/to? It is normal for the pack voltage to climb back up as your speed increases over 100kph due to back-EMF (e.g. it should not stay that low, so it may be completely normal).
Either way, a difference of 40kW isn't that big, but I would be curious to know why. Had you done several runs before that one?

That was just a longer drive on the German Autobahn with speed between 120khp and 160kph to warm up the battery for DC charing enroute (didnt work by the way due to bugs in the 2021.4.3, but thats another topic). No repetitive acceleration runs were done.

Max discharge power is using wrong tables for the refresh, so the values are off...usually too low. They are working on fixing this soon.
It is still interesting that you are showing a much higher value at lower cell temp. Makes you wonder if the US models are less power restricted...though the lower voltage limit of 278 peak and 295v seems identical.

I have to check if the voltage increases with speed. Not sure right now.

I did post some screenshots in the past that were showing my maximum battery power peaking at values between 60-80kw when doing longer drives in cold climate at SoC 12-9% on rural roads (Cell Temp 10-15°C). Very very slow acceleration indeed and a top speed, full throttle, of around 155kph.

Also, here is a screenshot were the battery heater stopped heating the battery when 10°C was reached. Cell temp never went any higher than 12°C at constant speeds on country roads. Only high speed driving on the Autobahn helped. If you would continue on country roads and reach lower SoC, that is when Björn Nylands "limp mode" kicks in. Nothing you can do about it, but hope there is a SuC nearby to navigate to, while not actually going there.

 

[AAKEE]

Makes you wonder if the US models are less power restricted...though the lower voltage limit of 278 peak and 295v seems identical.

I have to check if the voltage increases with speed

Cell temp never went any higher than 12°C .

I would be quite surprised if they put limits in EU vehicles only that was’nt needed for either a) regulations or: b) technical reasons.
So a low temp restriction at low SOC is most probable due to a technical reason( i.e “dont kill the battery”).
= Its not very likely ( me thinks...) to really differ for cars having a new battery technology that either needs more data before unleashing power or actually need protection to stay healthy.

For higher speeds, we know that power drops of when the engine speed increases.
I think we should mainly should use speeds below about 100-120 kph or 60-70mph for the reason to stay below the power drop, or at least specify the speed at which the test was done.
( Not to mean that you shouldnt do that higher speed test, it is absoltute still interresting).

Battery temps: If I leave the car out at cold temps for the day, and preheat before driving I dont know the battery temp at the start of the drive(not checked yet) but I know it will end /stay about 12C after a while driving.
I also know if the batt temp are higher than 12 it goes down and stays at about 12 degrees. I think the goal is to use heat from the battery needed, down to 12C.
If the battery is colder, Im leaning to believe(still to be checked) the heating aim is about 12c.

 

[Venom]

@AAKE

So this isn't THAT big of an issue? I have a M3P on order, would it be wise to pause delivery or will it be fine?