Battery Degradation Scientifically Explained

[darth_vad3r]

It might follow the supercharging profile curve or something a bit more conservative for some reason?

@AlanSubie4Life have you experienced limited regen after a long-duration regen when you are at a lower SoC like 20-30%, or only when you are up in the 70's SoC?

One thought ... it could have to do with the ramp up. SC-ing does a slower ramp, not an instantaneous start.
Since regen is like a one-second ramp every time, they might have to limit it more than SC-ing?

Another thought ... we could be back to the trap of thinking of power in kW only, instead of its voltage x current constituents.

Perhaps regen hits a current limit (or voltage limit) that is embedded in the supercharging profile that makes it appear to have a more restrictive profile when in fact it is the same underlying current and voltage limits being respected for various conditions (temperature, SoC, etc)?

 

[rainforest]

Does 80% max charge mess up anybody's rated range meter? I thought that many people had problems when charging to less than 90%, and that's why Tesla asks owners to charge to 90% nightly so that the BMS can recalibrate.

 

[AlanSubie4Life]

What allows the continuous Supercharging?

or only when you are up in the 70's SoC?

Cool battery (50-55 degrees or so but no less) and right around 69% SoC. (It was around 210 miles.) So not near where you would normally see a bunch of regen dots. But enough to see a minimal amount.

Long downhill...probably around 8-10% grade...I did the math on how much regen...I think about 30kW at 40mph...more and more regen dots. Got up to perhaps 72-73% SoC. Regen dots clearly not in proportion to the SoC increase. You’d have to look at the Supercharger curves to see where they are at at 70% SoC - but I think generally the Supercharger is above this rate at 70% SoC (LR).

Supercharging does a lot of heating! And obviously at a higher SoC it is limited.

I think the behavior is more or less consistent with supercharging - sounds like for similar reasons. Needs to be warm to make sure rapid diffusion into the anode can occur. Probably gets harder and harder as the thing gets more crowded with Li ions at higher SoC. But heat helps a lot it sounds like. It really seems that when the battery is cool they allow a decent amount of regen for a bit, but there are cumulative effects as the lithium ions accumulate (which presumably would not happen when the battery is warmer).

Maybe for high SoC, rate of maximum charging is affected by another factor other than diffusion rate into the anode, though?

Somehow have to explain how a certain rate could be allowed for a while, but not sustained...which seems to happen at only low temps...but at higher temps, another higher maximum limit applies (hit during supercharging), which is limited perhaps by a different physical process that is present even at high temps.

No idea.

 

[mechrock]

I don’t understand why long duration regen is limited when that regen is 10’s of kW (10-50+), whereas Supercharging is much higher. @EV-Tech Exp What allows the continuous Supercharging?

I have the same question.
I dropped 3000ft of elevation and saw some dots show up, they went away on some small up hill sections during the decent. Because I controlled the regen, never got close to the limit, but it still limited it. Maybe it just does it as precaution when it detects you are regening a lot?

 

[AlanSubie4Life]

I dropped 3000ft of elevation and saw some dots show up, they went away on some small up hill sections during the decent. Because I controlled the regen, never got close to the limit, but it still limited it. Maybe it just does it as precaution when it detects you are regening a lot?

Was the car fully warmed up when this happened?

What was the SoC you got up to on the descent?

It seems a little tricky...because you would think this would be an issue for Supercharging too. If you had been driving for an hour or so before this happened, I would be surprised, because then things would be warm. If it was early in the journey, see above for speculation.

It sounds like ions need to diffuse into the anode, and that process is limited by temperature. I wonder whether the actual bonding to the anode is a separate process, which is limited by the number of sites available.

Pure, 100% speculation:
So maybe in the case of a cold battery with regen, it's like trying to Supercharge when cold - the ions can't diffuse into the anode due to the temperature, so they are limited to bonding to the anode on the outside surface, and quickly all spaces in the small available volume get occupied and the allowed rate drops to prevent lithium plating. Whereas when the battery is warm, ions diffuse readily into the anode so they have the entire volume of the anode to bond to. So in that case the charge rate is just limited by the state of charge and the number of spots available to bond to?

Again, pure, 100% speculation, for a physical model of this. I know basically nothing about batteries. Seems to me that there must be two separate physical limiting factors to explain the behavior. Just one factor can't really explain it.

I think what isn't speculation is that whatever is happening, it is in order to prevent lithium plating, which is irreversible apparently, and also dangerous.

 

[darth_vad3r]

Does anyone know what the regen power looks like in terms of current and voltage? This could answer some questions.

 

[AlanSubie4Life]

Does anyone know what the regen power looks like in terms of current and voltage? This could answer some questions.

4200lbs * 9.81m/s^2 * sin(arctan(10/100)) * 40mph - Wolfram|Alpha


You can plug in values as you see fit. This is for a 10% grade. Anyway, for the situations where I have seen significant regen limitations, the regen power for that given SoC would be a lot lower than a Supercharger would give at that same SoC (assuming a fully warm battery of course - which is presumably the difference). (For example, the above was not possible for a 70% SoC.)

 

[darth_vad3r]

4200lbs * 9.81m/s^2 * sin(arctan(10/100)) * 40mph - Wolfram|Alpha


You can plug in values as you see fit. This is for a 10% grade. Anyway, for the situations where I have seen significant regen limitations, the regen power for that given SoC would be a lot lower than a Supercharger would give at that same SoC (assuming a fully warm battery of course - which is presumably the difference). (For example, the above was not possible for a 70% SoC.)

Perhaps I phrased my question poorly. Does anyone know how many Amps and Volts regen is generating for various levels. The system has various underlying voltage and current constraints.

e.g.
50 kW of power delivered via 500 V x 100 A might be voltage-limited by the system (if there’s a max 400 V limit)
50 kW of power delivered via 250 V x 200 A might be current-limited by the system (if there’s a max 125 A limit)

*Totally made up limits, just to illustrate what I’m asking and why. Power isn’t just “kW”, it’s about voltage and amperage.

 

[AlanSubie4Life]

Perhaps I phrased my question poorly. Does anyone know how many Amps and Volts regen is generating for various levels. The system has various underlying voltage and current constraints.

e.g.
50 kW of power delivered via 500 V x 100 A might be voltage limited by the system.
50 kW of power delivered via 250 V x 200 A might be current limited by the system.

Use the power above (reduced by the efficiency of the conversion - see below). The current back into the battery will be set by Ibat = P_regen/Vbat.

Vbat will be dependent on the current SoC of the battery (and slightly impacted by battery internal resistance of course). But that's not going to be significantly different than what happens during Supercharging.

It is possible that the AC-DC conversion electronics within the motors do get too warm, which may be what you're getting at - those would not be in use during Supercharging. But I very much doubt it given the powers in question and the ability of the motors to produce significantly more power on a continuous basis for quite some time, and various other reason. Obviously the power flow is in reverse so it will be different than producing wheel HP, but the heat sinking ability of the motor & regen electronics is probably pretty similar for both power flow directions (my understanding is it uses a lot of common circuitry within the inverter/converter).

We're probably talking about something like 10-15% loss (that may be too high, not sure) on the AC-DC conversion within the motors. So 4kW of constant power dissipation in the case above. Doesn't seem like it should be limiting. The battery seems like a more likely factor, especially since this issue seems to be the biggest issue when the battery & car is cold (power dissipation issues wouldn't behave that way). It also takes several minutes for the regen power to get pulled even further - longer than I'd probably expect for the thermal time constants involved. Anyway, lots of reasons to think this is a battery limitation.

 

[darth_vad3r]

Use the power above (reduced by the efficiency of the conversion - see below). The current back into the battery will be set by Ibat = P_regen/Vbat.

Vbat will be dependent on the current SoC of the battery (and slightly impacted by battery internal resistance of course). But that's not going to be significantly different than what happens during Supercharging.

It is possible that the conversion electronics within the motors do get too warm, which may be what you're getting at - those would not be in use during Supercharging. But I very much doubt it given the powers in question and the ability of the motors to produce significantly more power on a continuous basis for quite some time, and various other reason. Obviously the power flow is in reverse so it will be different than producing wheel HP, but the heat sinking ability of the motor & regen electronics is probably pretty similar for both power flow directions (my understanding is it uses a lot of common circuitry within the inverter/converter).

We're probably talking about something like 10-15% loss on the AC-DC conversion within the motors. So 4kW of constant power dissipation in the case above. Doesn't seem like it should be limiting. The battery seems like a more likely factor, especially since this issue seems to be the biggest issue when the battery & car is cold (power dissipation issues wouldn't behave that way).

No, I know how to do I = P/V, LOL.

The motor will generate AC naturally at some power level with a tendency to what AC voltage and current? As regen increases, does the current increase, the voltage increase, or both?

As the AC current or voltage increases during regen, do either of these encounter a limit in the system along their path from AC to DC to the battery? e.g. if the wiring/busbars are a certain gauge that can handle 200 A, but regen wants to give you 250 A, what happens? Likewise for voltage.

Do we know the system limitations? It doesn’t just go “backwards” through the DC-to-AC inverter, right? It has to go through a AC-to-DC converter possibly with a transformer behind that??

Or how does it work? I don’t know, but what I’m asking is how many AC volts and amps will the motor generate at given regen power outputs, and is that a fixed profile based on physical properties of the system, or tunable?
e.g. Will 33 kW of AC power from regen always be 330 V x 100 A? Or can they tune it to output varying amounts like 400 V x 82.5 A? How do these voltage and current levels vary, if at all, for different power levels (50 kW, 60 kW, ...)?
What does the AC go through to get to DC, and does that have input/output limits that force a constraint on regen?
Same question along every point from motor -> AC wiring -> converter -> DC wiring -> battery.
There are at least 5 areas here that may have constraints.

[This is a general question, and not related particularly to your apparent-cold-battery limitation]

I have seen limited regen on flat driving on the highway when the battery has been sitting at 25C all day and the SoC is in the 60’s. When this happens there are no regen dots while driving under power, but they sometimes appear while regen. The dots then go away as soon as power is applied again.

 

[AlanSubie4Life]

The motor will generate AC naturally at some power level with a tendency to what AC voltage and current? As regen increases, does the current increase, the voltage increase, or both?

The AC waveforms will vary in frequency & voltage depending on vehicle speed, etc. But this is all converted within the AC-DC converter, which preserves (with loss) power. It generates a certain amount of AC power which is converted to DC power.

As the AC current or voltage increases during regen, do either of these encounter a limit in the system along their path from AC to DC to the battery? e.g. if the wiring/busbars are a certain gauge that can handle 200 A, but regen wants to give you 250 A, what happens? Likewise for voltage.

I very much doubt the wiring is a limitation for regen, as the system can support 350kW of DC power, and massive amounts of AC power as well. All of the bus wiring from the inverter/converter is common AFAIK.

Do we know the system limitations? It doesn’t just go “backwards” through the DC-to-AC inverter, right? It has to go through a AC-to-DC converter possibly with a transformer behind that??

There's only a DC connection to the motor unit AFAIK. The inverter/converter is an integrated unit within the drive unit assembly, which probably uses as much common circuitry as possible. I definitely don't know exactly how either direction of power flow works, but I'm fairly sure the same power MOSFETs will be used in both directions - and that's where all the significant power dissipation is going to be and they are critical to overall system efficiency. The battery establishes the voltage on one side of those FETs, and then the motor is producing AC voltage on the other side that is switched in such a way as to end up sending DC current flow into the battery (which results in AC current flow through the stator windings on the motor side, which resists rotation of the rotor...). Beyond that is a lot of complexity.

what I’m asking is how many AC volts and amps will the motor generate at given regen power outputs, and is that a fixed profile based on physical properties of the system, or tunable?

The details of modern AC motor control are complicated. The AC voltage generated is proportional to the rate of change in the magnetic field through the motor windings. And the magnetic field is dependent on the current flow through the windings...and current is induced in the rotor windings...which generates a magnetic field...and how this field aligns with the stator field is important... I'm fairly sure there are various waveforms that could be generated depending on how much regen you were deciding to produce, etc. The amount of energy (and resistance to rotation) that can be extracted is controllable by the converter (the more current it pumps into the battery the more the motor resists rotation).

There are presumably maximum AC voltage limits allowed so as not to zap the MOSFETs, but I think that can be managed by the motor controller, and I don't see how that would be relevant to this particular phenomenon.

To be clear, I don't know the answers to all (any?) of your questions, but my guess is that the pulling of regen we see at increasing levels, the more you do, is not related to the inverter/converter unit and its power handling capabilities.

To the topic of this thread, I am fairly sure it is to prevent battery degradation. But I could definitely be wrong!

 

[darth_vad3r]

To be clear, I don't know the answers to all (any?) of your questions, but my guess is that the pulling of regen we see at increasing levels, the more you do, is not related to the inverter/converter unit and its power handling capabilities.

To the topic of this thread, I am fairly sure it is to prevent battery degradation. But I could definitely be wrong!

Not sure if you quoted and starting writing your post before I added the last two paragraphs in my post saying I’m not just considering “your” cold battery scenario, but I’ve seen limited regen on flat ground at low SoC at comfortable temperatures as well:

[This is a general question, and not related particularly to your apparent-cold-battery limitation]

I have seen limited regen on flat driving on the highway when the battery has been sitting at 25C all day and the SoC is in the 60’s. When this happens there are no regen dots while driving under power, but they sometimes appear while regen. The dots then go away as soon as power is applied again.

And yes, I believe we are definitely on-topic because limiting regen is done mainly to prevent battery degradation (other than any other system constraints that may limit it).

 

[AlanSubie4Life]

but I’ve seen limited regen in flat ground at low SoC at comfortable temperatures as well:

Those dots that showed up - was it at the beginning of your drive? I definitely see plenty of evidence that 25C is not warm as far as the battery is concerned. I've never seen that sort of behavior (on the flat, at 60% SoC) unless it's really cold, but you do have a smaller battery...though with less total regen capacity due to having only one motor and lighter weight, I suppose.

I feel like they may be super-conservative about sustained regen or even not-so-sustained regen when the battery is not up to operating temperature. But I wouldn't have expected that behavior at 60% SoC once you had been driving for a bit.

 

[mechrock]

Was the car fully warmed up when this happened?
Hard to say, but I ascended that same mountain a few hours earlier, went hiking and then drove back down. In the Mid 70s when I parked.

What was the SoC you got up to on the descent?
~52%

It seems a little tricky...because you would think this would be an issue for Supercharging too. If you had been driving for an hour or so before this happened, I would be surprised, because then things would be warm. If it was early in the journey, see above for speculation.

I've had that several times on trips going down I40 from Asheville, but I was probably only 30 minutes driving at that point having left at 90%.

 

[AlanSubie4Life]

Hard to say, but I ascended that same mountain a few hours earlier, went hiking and then drove back down. In the Mid 70s when I parked.

Yeah, not surprised. I think this is kind of the same thing I've seen. If you descended a few thousand feet with a battery cold-soaked to 70 degrees or so, it seems like it happens. Does not have to be very cold at all. (Also not sure what type of vehicle you have, etc.)

The interesting question is whether this can happen on a large descent after driving for a while, while at a fairly low SoC.

The same thing happening at high SoC after 30 minutes is of course not surprising - that could be the same thing but would be very sensitive at a high SoC.

 

[Dan203]

Since I'm leasing with no option to buy at the end should I just charge to 100% every night and not worry about it? Or is there a short term reason to not do that? I'd rather have the full range available every time I leave the house if possible, especially since it seems that Performance M3s get a lot less real world miles than the rated miles.

 

[AlanSubie4Life]

Since I'm leasing with no option to buy at the end should I just charge to 100% every night and not worry about it? Or is there a short term reason to not do that?

Ha. Definitely bad for the battery, but I guess as long as it lasts three years, all is well?

A reason not to pick up any lease vehicle from Tesla as CPO! One reason not to charge to 100%, I guess, is that they may very well allow you to buy the vehicle at the end of the term, regardless of the current position on this. Maybe you won't want one that's been abused.

A couple other reasons not to:

1) You'll lose a LOT of regen braking, especially when it cools down. This will hurt your efficiency. And you thought it was bad before!!!!
2) It'll be slightly less efficient charging, since the last few % are slow but still have the high ~200W overhead.
3) Tesla might call you to complain about just WTF is wrong with you. Unlikely though.
4) It probably increases the chance of a catastrophic battery failure (but I have no idea). This could be inconvenient...if it doesn't cause a fire, that's good. But if it causes your battery to lose a bunch of range, you might have your car out of commission for a bit while Tesla replaces it. Both very unlikely regardless, but probably more likely with that sort of charging pattern.

 

[Dan203]

So what's the recommended max you can charge and still get regen?

 

[AlanSubie4Life]

So what's the recommended max you can charge and still get regen?

We are getting off topic here. Anyway, in Carson City (?), depends on whether you store your car inside. 80% should be pretty safe but if it is cold you will see limited regen for a bit. 90% will nearly always give you limited regen, and in winter, even if you store your car inside a garage, you’ll probably see it, possibly significantly limited. It’s really about degrees of regen - a little limiting is no big deal. For you, most of the time 90% should be fine. If you need more range, charge up to a higher %. It is not a big deal.

 

[Dan203]

We are getting off topic here. Anyway, in Carson City (?), depends on whether you store your car inside. 80% should be pretty safe but if it is cold you will see limited regen for a bit. 90% will nearly always give you limited regen, and in winter, even if you store your car inside a garage, you’ll probably see it, possibly significantly limited. It’s really about degrees of regen - a little limiting is no big deal. For you, most of the time 90% should be fine. If you need more range, charge up to a higher %. It is not a big deal.

I don’t really want to futz with it day to day. For most days I wont drive enough for it to even matter, but if I drive up to Reno and end up driving around all day and then back home I don’t want to have to worry about it either.

So assume you don’t care about the longevity of the battery, you have a Performance, and you drive like an idiot. (i.e. hard pulls from every light, 15 over on the freeway, etc...) What would you charge it to?

I read somewhere that Tesla recommends 90%. Is that right? How much practical range will that get you in a performance?