[starky]

From having worked a bit in the industry I'm a bit skeptical about this study, I've definitely seen studies and experiments that used different initial charging conditions that would have shown better fade performance if this was true.

Not to mention, how much does the increased SEI change the impedance of the cell (thus reducing the subsequent charge speed) and the capacity available.

[Joel_Mckay]

Agreed, the study summary needs better explanation to justify the contradiction with dozens of other lab tests. We have several boxes of 21700 cells from various manufacturers (Samsung/Sony/Panasonic) undergoing aging trials for over 2 years now.

All LiIon and LiPol chemistries have shown the following:

1. deep-cycle discharges below 60% full cuts usable charge cycle counts from 8000 to under 2000 uses.

2. high-current discharge or rapid-charging accelerates capacity losses by about 15% a year

3. Internal resistance goes up as dendrite defect shorts damage the cell. Additionally, the self-discharge rates increase as the cell is degraded.

Very surprising if the technique works for all cell chemistries. =3

[gamblor956]

There isn't a contradiction.

This study solely focuses on the very first charge. It doesn't claim that recharging at high currents benefits battery life, only that the first charge at high current forms a larger protective barrier than a first charge at a low current.

Other studies have shown that a larger protective barrier improves lifespan. (See other comments on this thread for more details on the science.)

[Joel_Mckay]

Some early Microchip LiIon chargers did not split the cycle into 3 to 4 stages (pre-prep, CC-prep, and CV rapid charge).

i.e. they would drop into a constant-voltage rapid charge mode assuming the cell was prepped already.

None of these systems showed any sort of increased capacity or longevity. Quite the contrary results, this is why the new study details are rather intriguing. =3

[gamblor956]

The problem was probably the pre prep stage.... that appears to count as the first charge as it is defined in this newer study.

When they say "very first" they mean it very literally.

[Joel_Mckay]

There is a lot of what-ifs in the poorly written press release.

In general, we track the internal resistance (+-0.02ohm for Samsung) of the inventory under both new, charged, discharging, and storage levels. While I don't doubt the Papers results, they need to be clear on the methodology so others can validate it is not BS (common in this area.)

Have a nice day =3

[starky]

Unless you actually work for a cell manufacturer you aren't getting completely fresh cells though. They are talking about the first charge after the jelly roll is sealed into the can. When I would build cells by hand the standard procedure was to do the first couple cycles at 0.01C, record the capacity, and then change them to the charge rate for the experiment.

[Joel_Mckay]

Perhaps, normal practice is already usually to first cycle the cells in the CV charging region a few times to condition them for best capacity. Shipping regulations means Li cells are no longer shipped fully charged anymore even when new from an OEM.

It is fascinating news, and I look forward to more details. =3

[jve]

> deep-cycle discharges below 60% full cuts usable charge cycle counts from 8000 to under 2000 uses.

That is, if you do it single time you are down from 8k to 2k? Or it decreases gradually and 2k is the worst case?

Where can I read about it? Not a paper, but something more down to earth for consumers? That is, for a consumer to know how to properly maintain various devices (phone/car) for longevity?

[magicalhippo]

Keep in mind that for a car, 2000 cycles is still a fair bit. My BEV has a range of 350 km fully charged (highway). At 2000 cycles that's 700000 km.

That said, I previously read studies suggesting to keep the batteries between 30-70% SOC for optimal longevity, though I imagine there's been a lot of research so it might be now outdated.

[Joel_Mckay]

Some older Tesla Models used the Panasonic 21700 cells.

In general, "the battery is the car" for EVs... The bigger the better in my opinion, as it will last longer due to reduced stress on the cells.

Even if people are stressing the vehicle pack, they should still get 5 to 12 years out of the car. Note, some companies hide the expected range loss by over-provisioning capacity.

Best regards, =3

[jajko]

> The bigger the better in my opinion

This is definitely true, recently faced issues with my motorbike battery and oh boy are they fragile and lose charge by themselves quickly compared to bigger car batteries.

Basically I left a few times older (10-15 years) diesel bmw standing whole winter without touching it, started always without any issue (I know not the best idea re fuel in the tank, but it worked).

I did that once to completely new motorbike (honda) with good brand battery but I didn't unplug it, and now battery is permanently damaged and loses full charge in less than 2 days to such levels that it can't start the engine even if those 2 days its completely unplugged.

[Joel_Mckay]

Motorbike Pb AGM batteries are much different in modern vehicles. The prismatic packs often increase the plate surface area to bump cranking amp ratings. Thus, the cells design are thinner and more fragile too. We used these in some equipment at one time, for the extended temperature range.

Tip: if a Pb pack is partially discharged, it is more vulnerable to cold-weather related standby failures. Most people that own boats/heavy-equipment get a plug-in trickle-charger for Pb batteries, as the adapter also helps keep the pack slightly warmed.

Best of luck, =3

[Joel_Mckay]

The problem is there are many different types of Li cells. Some tolerate a wider Safe Operating Area for power output and temperatures.

I don't want to get into the name-and-shame game with other manufacturers. The 3 brands mentioned are generally very good quality, and if you can source new cells without counterfeit/expired nonsense... they will perform as per their app notes.

The cycle limit is a function of whether your charger IC is smart, slow-charge/low-current-discharge, and if your firm uses capacity Boosting (stress costs cycle counts.)

>Or it decreases gradually and 2k is the worst case?

Anecdotally, sensitivity seems somewhat correlated with cell use/age. The more stress, and the faster the cell degrades.

Best of luck, =3

[algo_trader]

> Samsung/Sony/Panasonic

> 1. deep-cycle discharges below 60% full cuts usable charge cycle counts from 8000 to under 2000 uses.

Presumably these are NMC variant?

Major Chinese LFP brands come with 6000K/10K cycle guarantees (but with specific operational parameters). Are these cycle predictions unrealistic ?

catl/eve/etc

[Joel_Mckay]

Yes, China batteries are so good the claims seem impossible... There are good manufacturers like anyplace else, but they are rightly priced accordingly. =3

[m463]

Wasn't supercharging EVs a lot frowned upon at first, but later found out not to negatively affect battery life as much?

https://electrek.co/2023/08/29/tesla-battery-longevity-not-a...

[Roark66]

This may be so in Teslas that have quite robust heat management. It definitely doesn't apply to many other brands. Anyone who knows anything about lithium batteries and sees temperatures at which fast charging is done will not believe any of the longevity claims. It is up to 55C during charging and during subsequent driving on a motorway it can take half an hour for this temperature to drop below 40C (look up BYD Seal 1000 mile challenge on YouTube for an example - all above st 9C ambient).

[atwrk]

BYD started as a battery producer and still is one of the largest in the world - I can't imagine them not having considered proper thermal management for battery health. Especially since they have been producing electric buses since 2010, which, as utility vehicles, see way higher usage (=charges) than consumer cars.

[chrisandchris]

I can imagine, and I can also imagine the buyer of those batteries stripping down "functionality" because of costs. So EV cars on the lower end could have higher peak temperature, because it's cheaper (i guess).

[delusional]

Is it a case of them not thinking about it, or did they think about it and figure that consumers would not pay for the increase in quality?

Plenty of engineered things could be "better" if price wasn't a concern. Most of this board can probably think back to a time where they had to leave long term value on the table because of short term costs concerns. It doesn't seem impossible to me that the engineers would leave some battery longevity (something that's hard to gauge for the consumer) on the table in pursuit of faster charging speed at lower prices (headline marketing items).

[originalvichy]

Doesn’t using an 800V architecture solve some of the heat problem? I believe currently the Koreans (Hyundai/Kia) and Porsche are the only major manufacturers using it. No surprise both can push nearly double as fast charging compared to 400v competitors.

[cenamus]

The total pack voltage shouldn't really matter for internal heating, it's due to the cells internal resistance, higher voltage really only allow for thinner wiring than the crazy high current low voltage packs (especially important for chargers/contacts etc)

[idiotsecant]

Internal heating is proportional to current. Higher voltage means lower current. Charging at the same wattage implies lower current.

So: higher voltage implies lower heating.

[hnuser123456]

The cells need to charge whether they're in series or parallel. The efficiency they can absorb charge at high speeds without heating up is not primarily determined by how they're wired. You could wire 5,000 cells in series, charge them with 20kv at 4 amps, or wire them all in parallel, and charge with 4 volts at 20 ka. Each cell will produce the same amount of heat either way, they only charge with about 95% efficiency. Higher voltage doesn't really reduce the need for active cooling if you want to keep the cells under 40-50C.

Energy loss though resistance in the pack's internal wiring is likely a lot less than the loss due to the chemistry not being 100% efficient at absorbing (or delivering) charge without heating up. But it does allow for thinner wires to get max power out of the battery.

[mohaine]

But I'm pretty sure the voltage seen by each individual battery is always the same, regardless of the distribution system voltage.

There should be less heating for higher voltages but if most if the heating is in the battery vs the distribution system then the higher voltages will not help much.

Also, if they make all wires smaller to save money and weight then there might not be any change in heating.

[stetrain]

The current per cell is still the same. 800V charging just means that you put cells in series banks to achieve an ~800V module-level voltage. Current is reduced in the main charging cables, charge port, and pack fuse/contactor, but not in the individual cells.

[wffurr]

Lucid uses a 1000V architecture.

I think the higher voltage allows them to charge more of the battery pack in parallel at a time.

[jve]

And because of this, electric vehicle manufacturers should take note.

If only for a city only car that you mostly charge at home, don't do roadtrips with multiple fast charges in short period of time you may get away with passive cooling. And if you don't live in a hot climate. But those are too many IFs.

[sokoloff]

That describes our use case pretty well (for a 2-car household) and we’ve been quite happy with the 26K miles we put on our Nissan LEAF in MA over a coming up on 10 year period.

Charged at home >50% of the time and pre-pandemic on the 6.6kW chargers at work. I can only recall one attempt we made at a beyond single battery trip, using an EVGo DC charger at the mid-point. I can say it worked, but subsequent trips to that same location were in the ICE car, so take of that what you will.

The car is now 80+% charged at home and is a city/nearby suburb runabout (and used for more trips, albeit not more miles, than the other ICE/hybrid).

It still has about 85% of its original battery capacity, which means we charge it about once a week, which works just fine for us.

[jve]

That also works for me for the second car. I also am awaiting Leaf delivery with 27k km on odo. however I did not expect battery to loose 15% of its capacity over 26k miles (which is 42kkm)

It is healthy to know how to maintain car battery. I will probably charge the battery to ~80% except when I need more range.

[sokoloff]

It’s also 10 years, which is a factor in degradation as well, not just cycles or distance.

It’s down 1 bar (of 12) and that was 2 years ago, so the 85% is estimated, but is within -0% to +4%. I have a LEAF Spy but haven’t checked it a long time.

[vondur]

Wasn't that how the Nissan Leaf's used to be setup, with passive cooling? I know it greatly affected their range in warm climates. I think they now switched to active cooling.

[starky]

How long does supercharging take? Even at 30 minutes, that is only a rate of 2C which is not that extreme for some cell chemistries as long as temperature is controlled.

[jazzyjackson]

Watching the comparison of a cyber truck vs a ram 2500, supercharging at 250kw took an hour and a half

https://youtu.be/HIs8zudJFzg

[oska]

The analogy they use in the article is all sorts of dodgy too :

> Removing more lithium ions up front is a bit like scooping water out of a full bucket before carrying it, Cui said. The extra headspace in the bucket decreases the amount of water splashing out along the way. In similar fashion, deactivating more lithium ions during SEI formation frees up headspace in the positive electrode and allows the electrode to cycle in a more efficient way, improving subsequent performance.