Extreme Long Term Model S Battery Life

[westcoast]

I understand that the expected behavior of a purchaser is to replace the battery with something bigger and better 5 or so years down the line. By then, we all hope that there will be cheaper, higher density batteries. Makes sense. I'll probably be one of them.

However, there are some (hopefully unlikely) scenarios where we might be stuck with a battery indefinitely:

- Purchaser falls on hard times, and wants to use the existing battery until it's well and truly dead. If you're only using a model S for small local trips, then perhaps 20% battery is just fine.
- Tesla goes out of business.
- Zombie Armageddon. Society breaks down, brains get eaten, and the only way we can get around is by charging our S's off solar panels.
- Etc.

My questions are these:

1. What is the best guess as to what happens to an S battery in the very long term? How do you think it will perform after 10 years? 15 years? 20 years? Does the battery max charge decline linearly? Exponentially? Does it actually reach a steady state at some point where the maximum range STOPS dropping? In 20 years, does it still have 25% charge? Or does it reach a point before then where it simply up and dies?

2. Over time, does the cost per mile increase because of charging inefficiencies? Let's say the battery in 10 years is at 50% of its 2012 max charge. Is the cost to charge it to its max 50% of what it costs to charge it today, assuming prices are the same? Or, does the battery charge less efficiently over time, requiring more current to charge the same amount?

Lots of hypotheticals here, but I find the exercise interesting to think about...

 

[qwk]

There is a chart on Panasonics website that has the time vs capacity plotted for some cells.

 

[VolkerP]

Going from the cell to the battery pack is where problems start. After some years, you will have that "lowest brick" that is repeatedly the worst. I assume that further cycling the pack puts more stress on the most degraded bricks. From there on, things will quickly go south with that one. So I''d say that after 12 years there is a high probability that the total pack capacity collapses.

I don't know if Tesla designed some redundancy into the pack, that would allow to disconnect bricks that went bad and go on with the rest. In that case, the battery would be unusable (=not produce enough power) after some 90% of all bricks have died.

I'm on a thin limb here but I think it's any ones guess with the things we know today.

 

[W8MM]

I think Tesla has a pretty sophisticated way of preventing failing paralleled groups of cells from going into failure runaway. I haven't bothered to read all of Tesla's patents in detail, so all I can do is try to interpret what Tesla service people have told me and others.

There is apparently a function applied during charging that attempts to balance or equalize the voltage of each of the clusters connected in series that form the pack. In the Roadster battery, there are clusters of 69 cells in parallel and these are connected in series a total of 99 clusters high. The series connections surely must have voltage monitors at each joint, allowing the battery management system to monitor every voltage drop in the chain. I don't know how Tesla encourages each cluster to reach the same 3.5 - 4.1 volt value, but I'm given to understand that they do. The voltage measured across each cluster of 69 paralleled cells is a pretty good analog for state of charge and the pack thereby knows what's going on inside.

Does that help any?

 

[hcsharp]

Going from the cell to the battery pack is where problems start. After some years, you will have that "lowest brick" that is repeatedly the worst. I assume that further cycling the pack puts more stress on the most degraded bricks. From there on, things will quickly go south with that one. So I''d say that after 12 years there is a high probability that the total pack capacity collapses.

I don't know if Tesla designed some redundancy into the pack, that would allow to disconnect bricks that went bad and go on with the rest. In that case, the battery would be unusable (=not produce enough power) after some 90% of all bricks have died.

I'm on a thin limb here but I think it's any ones guess with the things we know today.

I think Tesla has a pretty sophisticated way of preventing failing paralleled groups of cells from going into failure runaway. I haven't bothered to read all of Tesla's patents in detail, so all I can do is try to interpret what Tesla service people have told me and others.

I've been trying since before I bought my Roadster (over a year) to get an answer to this question - Does the ESS have a way of bypassing a brick? TM techs usually don't know or give inconsistent answers. My best guess is that they can control it during and/or after charging, but not while discharging. I think VolkerP is correct that the battery will fail quickly if one brick starts to get a little worse than the others, creating a stress concentration that leads to a failure runaway.

Observing the logs we notice that the weakest brick never stays the same. If it changes, then it's likely that the electronics are rotating bricks in and out of the weakest position.

Tesla released a software update for the Roadster that allowed balancing after a std mode charge, something that was previously only performed after a range mode charge. They were able to implement this with only a software change, so the hardware is there to control the amount of current flowing through each brick, but we can't necessarily conclude that it happens during charging or discharging. Perhaps only right after charging?

To completely bypass a brick while discharging would mean re-routing 600 amps! (They're not doing that!) So failure runaway is inevitable at some point. That means the pack will degrade quicker near the end of its life, unlike your laptop. I've heard there are a lot of electronics inside the ESS. I'd love to have lunch with a Tesla battery design engineer.

There is apparently a function applied during charging that attempts to balance or equalize the voltage of each of the clusters connected in series that form the pack. In the Roadster battery, there are clusters of 69 cells in parallel and these are connected in series a total of 99 clusters high. The series connections surely must have voltage monitors at each joint, allowing the battery management system to monitor every voltage drop in the chain. I don't know how Tesla encourages each cluster to reach the same 3.5 - 4.1 volt value, but I'm given to understand that they do.

Are you sure this balancing to 4.1v per brick is performed during charging? Or after? During would be the best design for safety and longevity, and It seems like it would be easy to turn off power to a brick once it reached 4.1v. But many people have observed that an un-balanced pack only regains its capacity after several charge cycles, and that it basically stops charging when the strongest cells reach 4.1v without waiting for the weaker bricks. It appears that most balancing happens immediately after a charge.

 

[W8MM]

Are you sure this balancing to 4.1v per brick is performed during charging? Or after? During would be the best design for safety and longevity, and It seems like it would be easy to turn off power to a brick once it reached 4.1v. But many people have observed that an un-balanced pack only regains its capacity after several charge cycles, and that it basically stops charging when the strongest cells reach 4.1v without waiting for the weaker bricks. It appears that most balancing happens immediately after a charge.

No, I'm not sure at all. Perhaps there is a resistor divider chain that slowly "leaks" the cells into balance and would not compete with the high-current charging episodes?

 

[rdunniii]

Is there a chart anywhere with the expected charging efficiency over some range of charging? The spec says the max charging efficiency is 92% What's the lowest? 90% or 50% or... What are the circumstances that cause it to drop and are they manageable? For example I would guess, and it is a guess, it is less efficient during high ambient temperatures because more power has to be consumed doing thermal control during charging. PG&Es kilowatts get kinda expensive at tier 3 if it's closer to 50%

 

[hcsharp]

Is there a chart anywhere with the expected charging efficiency over some range of charging? The spec says the max charging efficiency is 92% What's the lowest? 90% or 50% or... What are the circumstances that cause it to drop and are they manageable? For example I would guess, and it is a guess, it is less efficient during high ambient temperatures because more power has to be consumed doing thermal control during charging. PG&Es kilowatts get kinda expensive at tier 3 if it's closer to 50%

The thermal management is generally less of an issue than charger overhead. The only reason charging efficiency would drop above 92% SOC is because it wouldn't continue to charge at 40 amps/charger. It has to lower the amperage for the last part of the charge to prevent over-voltage to the cells. When you lower the amps, charger overhead becomes more of a factor.

As for ambient temperatures, it's actually more efficient at higher temps up to a point. That point is where the A/C compressor would have to come on. Before that happens there is a fan but I don't think just the coolant fan without A/C consumes much power. At least not in the Roadster. Charging at 37 deg F is less efficient than at 90 deg F.

 

[mnx]

2. Over time, does the cost per mile increase because of charging inefficiencies? Let's say the battery in 10 years is at 50% of its 2012 max charge. Is the cost to charge it to its max 50% of what it costs to charge it today, assuming prices are the same? Or, does the battery charge less efficiently over time, requiring more current to charge the same amount?

I always wondered this as well...

 

[Norbert]

There were some recent remarks (forgot where, probably someone talking to engineers at one of the events), that the Model S design goal was for the batteries to be good for at least twice the warranty period/mileage. It was said in a way that suggested they achieved it, or even over-achieved it, but not necessarily under all circumstances, or perhaps only under favorable ones, or might not be sure about it. And that it wasn't guaranteed or even promised, just some unofficial remarks.

Of course, it depends on the usage and charging patterns etc.

 

[Norbert]

2. Over time, does the cost per mile increase because of charging inefficiencies? Let's say the battery in 10 years is at 50% of its 2012 max charge. Is the cost to charge it to its max 50% of what it costs to charge it today, assuming prices are the same? Or, does the battery charge less efficiently over time, requiring more current to charge the same amount?

I haven't heard that the charging efficiency would go down by a similar amount. I'd guess it also goes down, but by a quite small percentage max. Maybe JRP3 knows.

 

[AlexSV]

I understand that the expected behavior of a purchaser is to replace the battery with something bigger and better 5 or so years down the line. By then, we all hope that there will be cheaper, higher density batteries. Makes sense. I'll probably be one of them.

However, there are some (hopefully unlikely) scenarios where we might be stuck with a battery indefinitely:

- Purchaser falls on hard times, and wants to use the existing battery until it's well and truly dead. If you're only using a model S for small local trips, then perhaps 20% battery is just fine.
- Tesla goes out of business.
- Zombie Armageddon. Society breaks down, brains get eaten, and the only way we can get around is by charging our S's off solar panels.
- Etc.

My questions are these:

1. What is the best guess as to what happens to an S battery in the very long term? How do you think it will perform after 10 years? 15 years? 20 years? Does the battery max charge decline linearly? Exponentially? Does it actually reach a steady state at some point where the maximum range STOPS dropping? In 20 years, does it still have 25% charge? Or does it reach a point before then where it simply up and dies?

2. Over time, does the cost per mile increase because of charging inefficiencies? Let's say the battery in 10 years is at 50% of its 2012 max charge. Is the cost to charge it to its max 50% of what it costs to charge it today, assuming prices are the same? Or, does the battery charge less efficiently over time, requiring more current to charge the same amount?

Lots of hypotheticals here, but I find the exercise interesting to think about...

One day Timur was in the mood for teasing Nasreddin Hodja.

`Hodja,' he asked, `can you teach your donkey how to read?

`Yes I can Great Timur.'

`Come now Hodja! How can you teach a donkey to read?'

`If you give me ten years time and 3000 gold coins, then I can teach my donkey how to read.' the Hodja was adamant.

`But if in ten years time, your donkey cannot read, then I will punish you most severely for trying to mock the Great Timur.'

Nasreddin Hodja and Timur agreed on the terms, the Hodja took the 3000 gold coins and left Timur's luxurious tent. Hodja's friends who witnessed the deal were incredulous.

`Hodja Effendi, what did you do? You know you can't teach a donkey anything other than braying. Ten years later Timur will have your head chopped off!' However, Nasreddin Hodja was not worried.

`My dear fellows,' he said calmly, `before ten years are up, either I will die or Timur will die. Or, the donkey will die!'





Sent from my iPhone using Tapatalk

 

[stopcrazypp]

1. What is the best guess as to what happens to an S battery in the very long term? How do you think it will perform after 10 years? 15 years? 20 years? Does the battery max charge decline linearly? Exponentially? Does it actually reach a steady state at some point where the maximum range STOPS dropping? In 20 years, does it still have 25% charge? Or does it reach a point before then where it simply up and dies?

Battery capacity doesn't fall linearly (at least for the 85kWh version), it falls basically exponentially. Initially the drop is quick and then it levels out. See the datasheet for the battery cells for the 85kWh Model S:
http://www.teslamotorsclub.com/show...or-Model-S/page2?p=61317&viewfull=1#post61317
No battery cell will last forever. Even when degradation "levels out" there is still a small and steady decrease that will make it so the cell will have no capacity at all eventually. Basically the core reason for permanent loss in capacity is that the lithium ions responsible for moving charge between the cathode and anode gets stuck in the anode. As the battery ages, more and more will get stuck until eventually there are no ions left to move charge.

2. Over time, does the cost per mile increase because of charging inefficiencies? Let's say the battery in 10 years is at 50% of its 2012 max charge. Is the cost to charge it to its max 50% of what it costs to charge it today, assuming prices are the same? Or, does the battery charge less efficiently over time, requiring more current to charge the same amount?

Interesting question. Cell impedance (internal resistance) increases as it ages. So there is a definite drop in discharge efficiency. But I'm not sure if it's proportional to the drop in capacity and if the charging efficiency will drop proportionally also. I say it's not likely to be proportional because a significant portion of the charging losses is from the AC-DC and voltage conversion and also battery cooling.

 

[Robert.Boston]

There were some recent remarks (forgot where, probably someone talking to engineers at one of the events), that the Model S design goal was for the batteries to be good for at least twice the warranty period/mileage. It was said in a way that suggested they achieved it, or even over-achieved it, but not necessarily under all circumstances, or perhaps only under favorable ones, or might not be sure about it. And that it wasn't guaranteed or even promised, just some unofficial remarks.

Of course, it depends on the usage and charging patterns etc.

Perhaps you were thinking of this answer from Elon during the Q2 earnings call:

For the battery warranty, it essentially merges language that Nissan was using, and I think maybe the Volt is also using similar language. Which is to say that the battery is warrantied for eight years, and the amount of miles you can really put on it. Actually, depending on which version - it’s unlimited for the 300-mile range version. Unless you torture yourself, you won’t be able to put enough miles on that to matter. So for the 160-mile version, you’d really have to be doing a lot of traveling within the city to do 100,000 miles in eight years. In terms of where the pack energy level will be at that point, it’s going to vary depending on what sort environment the pack is experiencing and what sort of driver the person has been, in much the same way as an engine, after eight years, is going to depend on whether someone was driving it hard or not, and what sort of environment it saw. It will still be a very usable pack. And we’re expecting the packs actually have a useful life that’s somewhere around double the warranty level.
People will probably want to replace it sooner than that, but these really last for a long time. Even our first generation pack on the Roadster, one customer in Europe just passed the 200,000 km mark, and he’s not someone who drives the car slowly. In fact, he passed the 200,000 km mark in a race.

 

[Norbert]

Perhaps you were thinking of this answer from Elon during the Q2 earnings call:

Nice find. That sounds like the same message, however I think someone was reporting from a conversation with an engineer or so, with a bit more detail.

 

[Bardlebee]

It would be interesting for people who get their model S to track their mileage cap at the end of each week or month and share it with everyone else here. Just to see how different climates affect the car. Though I am sure Tesla has this already inside the car and they run statistics themselves.