Simulated battery cycle lifetime test. What is the lifetime of your model S battery?

Philip · Dec 21, 2014

Hi all,

I'm embarking on a real-world educational exercise in finding out how long a Tesla Model S traction battery might last over its lifetime.

I'm building an automated charger/discharger for around 16 x 18650 cells. The aim is to constantly charge then discharge the individual cells under varying conditions, and plot battery capacity as a function of charge/discharge cycles.

Extrapolating from that test, I hope to come up with a figure of how many miles a 60kwh pack may last. While I appreciate there are warranties, and battery replacement schemes, and other market (and marketing) influences, I would like to concentrate on the physical batteries themselves.

I understand that the Model S, 60KWH pack uses 7104 individual Panasonic NCR18650A cells, wired in a 74 parallel, 96 series 74P96S configuration. Please correct me if I'm wrong, and give me a reliable source. I used wikipedia (Tesla Model S - Wikipedia, the free encyclopedia)

I'd like to start with testing 3 different scenarios. All of the voltage and current figures below are per individual cell, not per battery pack.

1. General base line test:
- Charge at 0.85A to 4.1V
- Discharge at 1A to 3V
- Repeat the above till the battery dies

2. Performance/Lead-foot test:
- Charge at 4.69A to 4V (this symbolises a 120kw supercharger)
- Rest 5 minutes
- Discharge at 8.8A for 6 seconds (this symbolises accelerating to 97km/h at max motor power of 225kw)
- Discharge at 0.564A for 3 minutes (this symbolises cruising)
- repeat the last two heavy acceleration and cruising steps till batt = 3.3V then go back to step 1- charge via supercharger
- Every 20 cycles, charge at 0.85A to 4.1V, discharge to 3V and record capacity

3. General driving test:
- your feedback here!
As a start, I would guess many people would charge via a 15A 240V outlet? Is this a fair assumption? Are there any stats on how many people use what charge current?
Then I would discharge at between 0.5A and brief periods of 5A bursts symbolising city stop/start driving.
I would discharge to say 3.3V per cell (so roughly 30% battery capacity), then charge to 4V (about 80%), then repeat the whole process.

The Panasonic datasheet for the NCR18650A battery only goes up to 500 cycles. This graph may be OK for a laptop battery pack designer, but not for an electric vehicle designer.

View attachment 66717

The graph above shows what happens in a lifetime of a cell that is charged to 4.2V and discharged to 2.5V, at 3A. This is very harsh on the battery, and not what it would see in a vehicle application. That's why my proposed tests are more focused on what a normal EV cell would experience.

As you can see, I haven't quite worked out what I want to do for the "general driving" test. What do you think would be a realistic test?

Naturally this isn't designed to be a fully scientific test, just an indicative test. Think Mythbusters logic. It will give us some idea of what we might expect, not a definitive answer. Only two batteries per charge scheme will be tested. If there is a large discrepancy between the two cells, I'll have to throw out the data for that test. The test will take some months to do continuously, day and night, and is again therefore not totally indicative of what a normal battery pack would go through.

If the test takes 6 months, technology will move on in the mean time, and better batteries will no doubt be popping up sooner than this. Even now, Panasonic has released the NCR18650B, which has a greater capacity, though it's anyone's guess as to what cell Tesla will choose for their next battery pack design. If funds permit, I'll get tests on the newer NCR18650B in anticipation of Tesla adopting this battery.

I'll try to get the ongoing results published to a live server so that anyone interested can have a look at how it's going. The test might prove nothing, but we might also find something interesting in the data.

Let me know your thoughts!

Regards, Phil

tom66 · Dec 27, 2014

A few issues I can see but sounds like an interesting test.

Tesla does not use a standard NCR18650A battery but instead some kind of modified version which has a longer cycle life and higher energy density but no cell level protection as the pack provides this protection. So you would need to get hold of this cell perhaps from a wrecked vehicle. wk057 may be able to help here as he has torn down a Tesla pack, maybe he has a few spare cells?

Tesla only use the cell from about 3.35V to 4.15V for 100% range charge (or 4.08V in standard charge mode.) 3.35V is well into the "car is shutting down" territory, so I wouldn't go down to 3V/cell.

Remember to cool the cells when charging at 4.7A, or they will probably have a considerably shortened lifespan.

Saghost · Dec 27, 2014

A couple thoughts here: I'm pretty sure the 74P96S configuration is for the 85 kWh battery only - the 60's only have 14 modules from what I've read, not having the two stacked together at the front center.

Each module has 6 cell groups in series, so most likely the 60 kWh is 84S - and I believe fewer cells in parallel as well, though there doesn't seem to be a consensus on exactly how many (we've had pictures showing empty spots in modules, but not clear enough to be sure of a count.)

Also, your simulated Supercharger is much more aggressive than the actual Supercharger. A 60 kWh car never sees 120 kW to begin with (peaks at 105?) - and on all cars it tapers down by the time it reaches half charge.

Do you have a plan for temperature monitoring and active cooling of the test cells? They normally enjoy the benefit of a sophisticated TMS that keeps them from overheating...
Walter

William13 · Dec 27, 2014

I agree with many of the comments above. My suggestion is to use a heat sink to simulate the active cooling. Surround the cells with thick ziplock bags of water or another liquid. You need at least 10 times the mass of the cells to absorb heat. Consider a cooling fan for the bags or limit the heavy charge discharge cycles to avoid temps over 120F.

qwk · Dec 27, 2014

This is going to be a complete waste of time for you. We are now 2.5 years into the Model S launch, and there are a handful of cars over the 100k mile mark(independent of Tesla's test mule vehicles). An internet search will tell you more about battery life due to cycling than anything else. When it comes to battery loss vs. time(what nobody knows yet), it's all a guess. IMO, the capacity loss vs. time function is what is most important. This is because the cells Tesla uses seem to lose very little capacity from just cycling them(even if abused with frequent range charges, or hot weather climates).

Mnlevin · Dec 27, 2014

500 cycles??? I charged my car (the MS60) every night since I got it May 5 2013. I think it is beyond 500 cycles already. My degradation is about 5 miles at 90% and about 8 miles at 100% and the mileage is 41k miles. I don't believe you could come up with anything that would replicate the TESLA model.

jerry33 · Dec 27, 2014

Mnlevin said:
500 cycles??? I charged my car (the MS60) every night since I got it May 5 2013. I think it is beyond 500 cycles already. My degradation is about 5 miles at 90% and about 8 miles at 100% and the mileage is 41k miles. I don't believe you could come up with anything that would replicate the TESLA model.

A cycle means 100% full to empty, anything less than that (and you can't really do a full cycle because the software doesn't allow it) makes a big improvement. The smaller the partial cycle, the more partial cycles you can do--and it's not linear.

nwdiver · Dec 27, 2014

Only about ~20% of my miles come anywhere close to >60% discharge of the battery. Li-Ion cells like small short cycles. >50% of my driving is in the middle ~40% of the batteries SOC (State of Charge). In other words I usually don't charge to >75% and I don't discharge <35%. With that kind of cycling you should be able to pass ~500MWh through a 85kWh pack or ~1.5M miles. For an accurate representation of 'daily' driving try that... 35-75%... be willing to bet you can do that >10k times and your battery will still have >70% of its initial capacity.

The real killers of Li-Ion batteries are high-temp + high SOC, Low Temp + High charge-rate and deep discharge. The closer they are to 50% SOC the happier they are.

Robert.Boston · Dec 30, 2014

FWIW, I think the majority of Model S owners charge at 40A 240V, the normal current draw from a NEMA 14-50, which is the minimum install recommended by Tesla. Some folks buy the HPWC, which gives them the option of charging at 80A 240V.

It would be interesting to test @nwdiver's statement by comparing the effect of narrow cycling (between, say, 3.6V and 3.9V) and deep cycling (between, say, 3.4V and 4.1V). You'll need to standardize the number of cycles by the amount of energy drawn, so that you compare apples to apples.

rdrcrmatt · Jan 3, 2015

Here's a post I made a while ago.

rdrcrmatt said:
I'm still seeing 225-231 miles on 90% charge, 255-260 on 100% charge.. just over a year old, 41,000 miles.

I drive 160-240 miles per day when I have my long commute (3-4 days per week)

Mileage * Wh/Mi Avg / kilo / usable kwhr in battery
41000 * 330wh/mi / 1000 / 75 = 180.4

That's 180 "full" charge cycles so far. So at "2000 charge cycle life" I'm nearing 10% used in my battery. So, 400,000 miles for me to hit 2000 charge cycles or that 70% point.

I'm currently at 50,000 miles on the button. I get 221 to 226 on 90% charge now. It seemed to drop just after I made that post but IMO rated miles don't really mean all that much.

qwk · Jan 3, 2015

rdrcrmatt said:
Here's a post I made a while ago.

I'm currently at 50,000 miles on the button. I get 221 to 226 on 90% charge now. It seemed to drop just after I made that post but IMO rated miles don't really mean all that much.

What are your ideal miles for a full range charge?

scaesare · Jan 3, 2015

rdrcrmatt said:
Here's a post I made a while ago.

Similarly, here's some of my musing on the subject a bit back:

scaesare said:
It's been suggested (both here, and in various whitepapers and articles by Tesla and others), that the Roadster used earlier Li-ion technology (Li-Ni?) that would be good for ~500 full discharge cycles before having degraded to a specific point (70%).

The Model S has a newer chemistry (NCA?) that has rumored to be good for 2,000+ cycles to get to the same point.

As the rating is based on full discharge cycles, it's understood that two 50% discharge cycles is equivalent to a full cycle (more or less)... and is the same for four 25% cycles, etc... It's also understood that those shallower cycles in the "middle" of the charge range is better than extremes at the top or bottom of the range.

So, I'm almost exactly at 2,000 miles a month on my S85. That averages to 67 miles a day over 30 days. It just so happens my weekday commute is about 66 miles. Apparently my weekend driving keeps up this average pretty closely.

That's also just almost exactly 1/4 the range of a full charge. Which matches my typical charge at night I charge to 60% and typically plug in around 35% at night... or about 25% of charge I add nightly on average.

So... if the expected pack life is 2,000 full cycles, I could be looking at 8,000 of my average nightly charge to reach the same point. That's 22 years.

Now, I don't expect I wont see degradation before then... as there are ather factors such as just calendar age, etc... but as long as the thermal management of the car is doing what it should it will be interesting to see how well our packs retain capacity over time...

Philip · Jan 4, 2015

Thanks all,

Very interesting reading your replies. Some more thoughts:

Supercharging current: I have scaled back the assumed supercharger current to 3.91A per cell. This is mainly because I don't want to set fire to the battery, not knowing what it can actually take. 3.91A per cell is still 289A for the entire battery pack, and just over 100KW supercharging power. (101.2kw)

Active cooling: The datasheet shows a 0.5V drop between 0.2A and 2A draw, so at worst, the battery resistance is 0.25 ohms. This shows that each cell in the pack is dissipating 4 watts at 4 amps (P=I^2*R). Not much per cell, very manageable with just a small fan, but multiply by 7104 cells and you get over 28KW of heat that has to go somewhere, so you can definitely see why temperature management is very important in a dense pack.

General driving test: I like the thought of putting a 40% SOC cycle into the equation, attempting to maximise the life of the battery, but also subjecting it to periodic accelerating currents. Will think some more about this.

Cheers, Phil

- - - Updated - - -

tom66 said:
Tesla only use the cell from about 3.35V to 4.15V for 100% range charge (or 4.08V in standard charge mode.) 3.35V is well into the "car is shutting down" territory, so I wouldn't go down to 3V/cell.

Do you have a reference for this, or is this just a ballpark figure?

rdrcrmatt · Jan 7, 2015

qwk said:
What are your ideal miles for a full range charge?

Next time I do a range charge I will report back to let you know.

ljwobker · Jan 7, 2015

Aren't we actually interested in the amount of energy we can access in a "fully charged" battery? If so...

Is a more accurate way to identify "full" charge capacity the following:

Do a range charge + some amount of time to balance/stabilize.
Drive until rated miles == 0.
Read the consumed energy from the "time since last charge" meter...

thoughts?

rdrcrmatt · Jan 13, 2015

rdrcrmatt said:
Next time I do a range charge I will report back to let you know.

I got 289 on a range charge the other day, but I never got to charge completed. It was at 5 minute remaining for maybe 10-20 minutes and I had to go.

- - - Updated - - -

ljwobker said:
Aren't we actually interested in the amount of energy we can access in a "fully charged" battery? If so...

Is a more accurate way to identify "full" charge capacity the following:

Do a range charge + some amount of time to balance/stabilize.
Drive until rated miles == 0.
Read the consumed energy from the "time since last charge" meter...

thoughts?

Or, on the screen at Tesla service centers they can show the estimated available battery capacity.

David99 · Jan 13, 2015

Phil,

have a look at this video. It's what you are trying to do but on steroids
Why do Li-ion Batteries die ? and how to improve the situation? - YouTube

David

Philip · Jan 14, 2015

^^ Awesome link David99, thanks for that.

rdrcrmatt · Jan 16, 2015

qwk said:
What are your ideal miles for a full range charge?

291 this morning.

qwk · Jan 16, 2015

rdrcrmatt said:
291 this morning.

Interesting, thanks.

Simulated battery cycle lifetime test. What is the lifetime of your model S battery?

New Member

Member

Well-Known Member

Active Member

P130DL

Member

(S85-3/2/13 traded in) X LR: F2611##-3/27/20

Well-Known Member

Model S VIN P01536

Member

P130DL

Well-Known Member

New Member

Member

Geek.

Member

Well-Known Member

New Member

Member

P130DL

Similar threads