Battery pack/module replacement (out of main)

[KarenRei]

BTW, I was thinking about the modular design of the Model 3 battery, and failure modes... and something occurred to me.

The plan for Tesla is that if a battery dies, to only replace the defective module, not the whole pack. On the downside, you're not getting a whole new pack. On the upside, a module probably only currently costs Tesla a bit over $2k (vs. ~$10k for a full pack), so add labour and any service margin to that.

But let's think about what this means for an older Model 3 - say 10 years old, 120k miles (average US driver = 12kmi/yr). This car, if degradation is like a typical S pack, has about 6% degradation by then. Now a module dies. What's the replacement like? Well, its not just new - it's also better and cheaper. Let's use the Roadster pack upgrade as an example - 8 years later, and the price went down from ($54k?) to $29k, and energy up 40%. So for a bit-over-$2k Model 3 module, it's now around $1,1k (plus labour and margin). The dead module's degradation is gone, and the new module has 40% more energy than the one removed. Now the pack as a whole actually has 5,5% more energy than it had when it was new, for a rather marginal cost!

Would that actually work? Honestly, I'm not sure. The added energy density is highly uneven; the voltage will drop a lot more slowly from the new module than the old ones. You've also got one higher-energy 2-module circuit alongside one lower-energy 2-module circuit. I'd think that the system would have to constantly rebalance to compensate, and I'm not sure of the net effect of that (I'm not sure how the 3's BMS handles charge balancing between modules, if at all). But then again, since the pack is designed for module replacements, even if the exact same type of module is inserted vs. the one that was removed, there's always going to be some imbalance between the new module and the old ones (the old ones being degraded), so the system has to have some way of dealing with this.

Regardless... I think it might be surprisingly cheap to keep old Model 3s running due to their modular design. And a failure might actually leave you with a pack that has more range than the car did when new. (And even if they couldn't deal with such an imbalance? Well, they could just make the replacement modules with fewer cells per brick, making them even cheaper )

 

[jerry33]

Would that actually work? Honestly, I'm not sure. The added energy density is highly uneven; the voltage will drop a lot more slowly from the new module than the old ones. You've also got one higher-energy 2-module circuit alongside one lower-energy 2-module circuit. I'd think that the system would have to constantly rebalance to compensate, and I'm not sure of the net effect of that (I'm not sure how the 3's BMS handles charge balancing between modules, if at all). But then again, since the pack is designed for module replacements, even if the exact same type of module is inserted vs. the one that was removed, there's always going to be some imbalance between the new module and the old ones (the old ones being degraded), so the system has to have some way of dealing with this.

It would if there were separate channels for each module, so that each module could charge and power at it's peak level. If they are all together, then the power and charging rate and C limit would be the max of the weakest module. In other words, there would need to be some kind of a splitter box to direct charging and power to each module.

 

[mongo]

So for a bit-over-$2k Model 3 module, it's now around $1,1k (plus labour and margin). The dead module's degradation is gone, and the new module has 40% more energy than the one removed. Now the pack as a whole actually has 5,5% more energy than it had when it was new, for a rather marginal cost!

Would that actually work? Honestly, I'm not sure. The added energy density is highly uneven; the voltage will drop a lot more slowly from the new module than the old ones. You've also got one higher-energy 2-module circuit alongside one lower-energy 2-module circuit. I'd think that the system would have to constantly rebalance to compensate, and I'm not sure of the net effect of that (I'm not sure how the 3's BMS handles charge balancing between modules, if at all). But then again, since the pack is designed for module replacements, even if the exact same type of module is inserted vs. the one that was removed, there's always going to be some imbalance between the new module and the old ones (the old ones being degraded), so the system has to have some way of dealing with this.

It would if there were separate channels for each module, so that each module could charge and power at it's peak level. If they are all together, then the power and charging rate and C limit would be the max of the weakest module. In other words, there would need to be some kind of a splitter box to direct charging and power to each module.

Like Jerry wrote, if the modules are in series it won't help capacity. If they are in parallel it will. No extra electronics needed other than monitoring (dis)charge rate on a per module level and clipping to the most utilized module.

 

[JRP3]

Would that actually work?

No way, that's a BMS nightmare. Plus future costs will be lower, no reason to mess with modules, just replace the whole pack.

 

[mongo]

No way, that's a BMS nightmare. Plus future costs will be lower, no reason to mess with modules, just replace the whole pack.

BMS isn't an issue as long as it only tries to balance on the top end. Get all the groups to 100% SOC at the same time and ignore that they are 20% different at the low end.

For degredation, yeah replace all the modules($5k-$7k per Elon). Single module replacement would be calked for if there were a cell that turned into a resistive short, or you lost 3 cells in the same parallel group (10% pack loss).

 

[JRP3]

BMS isn't an issue as long as it only tries to balance on the top end. Get all the groups to 100% SOC at the same time and ignore that they are 20% different at the low end.

The different capacities and different effective resistance between the old and new modules means they will likely get out of balance each cycle. Might overwhelm the bleed resistors. Remember the BMS is designed for closely matched cells, so very small differences in capacities, and little need for heavy balance resistors.

 

[Artful Dodger]

Might overwhelm the bleed resistors.

I don't think we know what TOC balancing tech Tesla is using. There are more ways to balance parallel groups than simple top bleeding. For example, "charge pumps" are designed to balance groups of cells continuously throughtout their complete range of SOC.

Active Cell Balancing using Microcontroller - IEEE Conference Publication | 2018

Abstract— The proposed idea uses a microcontroller as
brain to regulate the switch connected to inductor, which is a
passive element and is used for transfer of charge from one cell to
another cell. This idea uses deviation of each cell’s voltage level
from the mean voltage calculated by the microcontroller. The
proposed balancing algorithm can reduce the micro cycles taken
to transfer charge from one individual cell to another.

Keywords— Active Cell balancing, Battery management system,
charge equalization, State-of-Charge​

Fig 3 Discharging of Inductor and Charging of C2

Using this techique, 100% of the packs capacity can be accessed, even when individual modules differ in spec for capacity and internal resistance.

I have yet to see a technical description of the balancing technique used by Tesla. If you have a pointer that'd be v. informative.

Cheers!

 

[mongo]

The different capacities and different effective resistance between the old and new modules means they will likely get out of balance each cycle. Might overwhelm the bleed resistors. Remember the BMS is designed for closely matched cells, so very small differences in capacities, and little need for heavy balance resistors.

Not so. Each parallel cell group is impacted by current in/put and internal parasitic resistance (leakage). Capacity difference doesn't matter if the BMS is programmed to balance to a specific SOC. It only needs to compensate for the difference between the best and worst parasitic resistances.

 

[JRP3]

I have yet to see a technical description of the balancing technique used by Tesla. If you have a pointer that'd be v. informative.

Not off the top of my head but I remember wk057 saying only very small currents could be handled by the BMS.

 

[JRP3]

It only needs to compensate for the difference between the best and worst parasitic resistances.

Which will be larger between new and old modules.

 

[mongo]

Not off the top of my head but I remember wk057 saying only very small currents could be handled by the BMS.

S looks like it can bleed 100mA (4V across (158/4)ohm) if I am reading the resistors and interpreting the design properly.
http://evtv.me/evtv-word-press/wp-content/uploads/2017/01/Tesla-BMS-ModuleAnnotated-Front-2.jpg
Model 3 is completely different and may be based on LT chips. I don't see any active balancing components in this picture:

Tesla Model 3 - NextGen Battery - EVTV Motor Verks

Which will be larger between new and old modules.

Larger sure, but does leakage get that much worse with age?

 

[KarenRei]

Like Jerry wrote, if the modules are in series it won't help capacity. If they are in parallel it will. No extra electronics needed other than monitoring (dis)charge rate on a per module level and clipping to the most utilized module.

They're neither in simple series nor simply parallel. There's two parallel circuits, each comprised of two modules in series (4 modules total).

No way, that's a BMS nightmare. Plus future costs will be lower, no reason to mess with modules, just replace the whole pack.

Few people will replace the entire pack on a 10-year-old Model 3, even with the assumed price 29/54ths price drop implied by the original Roadster pack upgrade (which would put in the ballpark of $5,4k, plus labour and margin).

Tesla specifically designed the pack to be modular so that one can just replace individual modules.

The different capacities and different effective resistance between the old and new modules means they will likely get out of balance each cycle. Might overwhelm the bleed resistors. Remember the BMS is designed for closely matched cells, so very small differences in capacities, and little need for heavy balance resistors.

It has to be able to handle some meaningful difference, however. If a module dies on a degraded pack, and is replaced by a new one, there could be a meaningful difference in capacity between the old ones and the new one. The question is "how much is too much".

There's also always the option of also replacing the matching module on the opposite side of the pack so that each of the two series paths contains one old, degraded module and one new, high capacity module. That would, of course, double your parts costs (~$2,2k) and (somewhat) increase your labour costs.

But as noted elsewhere, it might just make more sense to simply include fewer cells in the "future replacement modules", making them even cheaper, rather than using just as many of the "future-higher-capacity cells". Module costs would drop to under $1k in such a scenario. Even after labour and margin, that's an easy-to-justify fix, even on an old car.

Who is Jack Richard and what was the "demand hypothesis" after Q1?

Typo, of course - Jack Rickard.

The "demand hypothesis" was the notion that demand for Tesla products was collapsing. That US sales had gone into some sort of permanent low-purchase-rate state, that Tesla had exhausted all of its carrots and couldn't bring them back up (and they'd only decline further in Q3 '19 and Q1 '20 as tax credits expired), that Tesla had exhausted European demand in Q1, that there were no new markets of significance left to move into, and basically, Tesla was screwed.

Just your general pointless Q1 freakout. Gali fell for it, for a period (he also started hyping up NIO at the time). Then a couple weeks later he was calmed down and buying Tesla stock again.

 

[Sean Wagner]

Regarding replacement of a Model 3 battery module, there will be tons of packs available to serve as parts donors.

The Teslas presently rolling off the line will be near classics by then - then being this side of the non-roaring '20s.

For countries and corporations and investors and UAW workers, it pays to think about the changes the next ten years will bring.

 

[hacer]

... The dead module's degradation is gone, and the new module has 40% more energy than the one removed. Now the pack as a whole actually has 5,5% more energy than it had when it was new, for a rather marginal cost!

Would that actually work? Honestly, I'm not sure. The added energy density is highly uneven; the voltage will drop a lot more slowly from the new module than the old ones. You've also got one higher-energy 2-module circuit alongside one lower-energy 2-module circuit. I'd think that the system would have to constantly rebalance to compensate, and I'm not sure of the net effect of that (I'm not sure how the 3's BMS handles charge balancing between modules, if at all). But then again, since the pack is designed for module replacements, even if the exact same type of module is inserted vs. the one that was removed, there's always going to be some imbalance between the new module and the old ones (the old ones being degraded), so the system has to have some way of dealing with this.
...

No, it won't "work" in the sense of being able to use any of that extra energy. The modules are in series and thus the pack energy available is never greater than N (the number of modules) times the energy of the weakest module. Re-balancing is done with resistors bleeding down individual modules. This allows the pack to charge to the most capacity that it can with no module going over-voltage, but the pack will still be depleted when the weakest module hits it low voltage limit.

 

[hacer]

I don't think we know what TOC balancing tech Tesla is using. There are more ways to balance parallel groups than simple top bleeding. For example, "charge pumps" are designed to balance groups of cells continuously throughtout their complete range of SOC.

Except engineers have torn down Tesla's battery packs and that is not what they're using.

 

[Artful Dodger]

Except engineers have torn down Tesla's battery packs and that is not what they're using.

Jack Richard wrote about the BMS on May 27, 2018: (see the whole blog for context)

"And in a way, I am shocked. It’s as if everything we thought we knew about the Model 3 is totally wrong. And the story of this remarkable device is simply a not yet revealed secret. And with all the hoopla about the Model 3, no one has ventured anything on this at all. What are they doing over at TeslaMotorsClub forums? I was aware they were kind of simple minded, but they do go on and on and on and I would think SOMEBODY would have actually looked under the hood or something."​

Since this is likely too far in the weeds for the General thread, I'm going to depart this topic with this final comment: "Show me a bleed resistor on a Tesla Model 3." TIA.

Cheers!

 

[hacer]

They're neither in simple series nor simply parallel. There's two parallel circuits, each comprised of two modules in series (4 modules total).

That does NOT agree with Jack's model 3 LR battery tear-down which show all 4 modules in series. Think about it: a 23S module has about 100 Volts, and a 25S module about 108V In series this is only 208V i.e. half of the full pack, so that alone debunks the idea of paralleled modules.

 

[hacer]

Jack Richard wrote about the BMS on May 27, 2018: (see the whole blog for context)

"And in a way, I am shocked. It’s as if everything we thought we knew about the Model 3 is totally wrong. And the story of this remarkable device is simply a not yet revealed secret. And with all the hoopla about the Model 3, no one has ventured anything on this at all. What are they doing over at TeslaMotorsClub forums? I was aware they were kind of simple minded, but they do go on and on and on and I would think SOMEBODY would have actually looked under the hood or something."​

Since this is likely too far in the weeds for the General thread, I'm going to depart this topic with this final comment: "Show me a bleed resistor on a Tesla Model 3." TIA.

Cheers!

From the same blog page see Tesla Model 3 - NextGen Battery - EVTV Motor Verks which means it doesn't really matter what's inside that chip (but it's almost certainly FETs for bleeding) because it connects to each cell through a resistor far too large to allow any switch-mode power transfer. I'd point out the resistors for you, but they are on the back side of the board (per Jack's comment) and he didn't post a back-side photo.

 

[UncaNed]

BTW, I was thinking about the modular design of the Model 3 battery, and failure modes... and something occurred to me.

The plan for Tesla is that if a battery dies, to only replace the defective module, not the whole pack. On the downside, you're not getting a whole new pack. On the upside, a module probably only currently costs Tesla a bit over $2k (vs. ~$10k for a full pack), so add labour and any service margin to that.

But let's think about what this means for an older Model 3 - say 10 years old, 120k miles (average US driver = 12kmi/yr). This car, if degradation is like a typical S pack, has about 6% degradation by then. Now a module dies. What's the replacement like? Well, its not just new - it's also better and cheaper. Let's use the Roadster pack upgrade as an example - 8 years later, and the price went down from ($54k?) to $29k, and energy up 40%. So for a bit-over-$2k Model 3 module, it's now around $1,1k (plus labour and margin). The dead module's degradation is gone, and the new module has 40% more energy than the one removed. Now the pack as a whole actually has 5,5% more energy than it had when it was new, for a rather marginal cost!

Would that actually work? Honestly, I'm not sure. The added energy density is highly uneven; the voltage will drop a lot more slowly from the new module than the old ones. You've also got one higher-energy 2-module circuit alongside one lower-energy 2-module circuit. I'd think that the system would have to constantly rebalance to compensate, and I'm not sure of the net effect of that (I'm not sure how the 3's BMS handles charge balancing between modules, if at all). But then again, since the pack is designed for module replacements, even if the exact same type of module is inserted vs. the one that was removed, there's always going to be some imbalance between the new module and the old ones (the old ones being degraded), so the system has to have some way of dealing with this.

Regardless... I think it might be surprisingly cheap to keep old Model 3s running due to their modular design. And a failure might actually leave you with a pack that has more range than the car did when new. (And even if they couldn't deal with such an imbalance? Well, they could just make the replacement modules with fewer cells per brick, making them even cheaper )

That's a great idea. It would be WONDERFUL if they make it work like you describe.

What a slap in the face to naysayers and FUDspewers if Tesla turns a bit of battery trouble into a huge, low-cost, range-improving benefit instead! I hope you can persuasively suggest this to someone who will listen at Tesla so they will make it happen.

Also...

I'd hoped I was the first one to publish the word "fudspewer", but it Google, sadly, shows otherwise.

 

[MP3Mike]

They're neither in simple series nor simply parallel. There's two parallel circuits, each comprised of two modules in series (4 modules total).

Who are you and what have you done with our Karen?

Our Karen would know that all 4 modules are in series, the only parallel wiring is internal to the modules for the individual bricks that are inside the module.

As far as replacing an individual model @wk057 says that it is almost impossible, at least with the S&X packs, because they would have to have almost the identical capacity as the other modules or things get out of balance too quickly and cause problems.