80%, 90% charging?

[E-HP]

It's pretty common knowledge that avoiding charging to 100% will extend a battery's lifetime. Some people charge to 90% capacity, some even lower to 80%. I'm wondering if people use the capacity of the battery, or the voltage in determining 90%, or a percentage of the voltage range (e.g. 3.0V - 4.2V). Looking at the table in the linked article below, it looks like 90%, by capacity, would be around 4.15V/cell. 90% by voltage comes to 4.08V, I think.

4.20 100%
4.15 90–95%
4.10 85–90%
4.05 80–85%
4.00 70–75%
3.90 60–65%
3.80 35–40%
3.70 30% and less

https://batteryuniversity.com/learn/article/how_to_prolong_lithium_based_batteries

I just charge to 4.1V because it's easy to remember and to do mental math with :lol:

 

[spinningmagnets]

Nissan, Tesla, and Chevy-Volt use 4.05V as the cap under most circumstances, for max life.

A while back, there was a hurricane approaching the south-east coast of the US. Tesla has a built-in charging protocol to make the battery last as long as possible. They announced that any Tesla owner in that region would now automatically be able to charge their battery pack to 100%, in order to give the driver the absolute maximum amount of miles to help them escape the region, with no certainty that they would be able to find an empty charging station when their pack got low.

This led to some confusion among Tesla owners, with accusations that they payed full retail price for their EV, but until the hurricane, they had not been given the packs full range capability. In an associated side-note, owners were only allowed to access the "fast charge" capability a certain number of times each year, without pro-rating the warranty on the battery.

Both are for the same reason. If you charge a Tesla to 100% (hardening of the chemicals in the active material), and if you frequently "fast charge" (heat), the battery will not give you the number of warrantied miles before it loses capacity to the promised 10-year rated 80% capacity

 

[eMark]

If your Charger is adjusted to 4.1V/cell how often and how long do you leave the charger plugged in after the green light comes on so your BMS can correct any imbalance or isn’t that practice ever necessary?

 

[john61ct]

Some people charge to 90% capacity, some even lower to 80%. I'm wondering if people use the capacity of the battery, or the voltage in determining 90%, or a percentage of the voltage range (e.g. 3.0V - 4.2V). Looking at the table in the linked article below, it looks like 90%, by capacity, would be around 4.15V/cell. 90% by voltage comes to 4.08V, I think.

No that's not how it works, you are mixing up units.

SoC% is in Ah, and is pretty hard to measure accurately without specialized knowledge and lab-style equipment.

So forget about 80% 90% malarkey, really just an abstract idea very hard to implement, pretty useless in practice, owners only think they know what the SoC outcome of their charger's algorithm is giving when it cuts out.

And do not rely on that BU site for any IRL details, only good for understanding fundamentals, too much generalizations.

Voltage is not related to precise SoC in any straightforward way.

Best to just talk about those voltages, with full details in order to be precise

of course varies with chemistry too.


> charge to 4.1V

That is a fine generalization, but without the details, does not have a concrete meaning

You have voltage at which charging is halted, varies **a lot** based on the C-rate

also how long CV is held, if there is an Absorb cycle.

Then there is voltage after a load (even only a tint current) is applied.

Then you have fully isolated cells, resting voltage, long after the surface charge has been eliminated via the above.

If you do want to (try to) correlate a voltage to SoC%, then that is the way to start with an objective and meaningful standard of accurate measurement.

 

[john61ct]

If your Charger is adjusted to 4.1V/cell how often and how long do you leave the charger plugged in after the green light comes on so your BMS can correct any imbalance or isn’t that practice ever necessary?

Buy gear that lets **you** decide when balancing is required

should ideally be only very occasionally.

The method and hardware you use for balancing when it **is** necessary,

should not dictate how you do your charge cycling for normal usage.

And besides being fast (high balance current) it should let **you** decide at what voltage you want the cells to get balanced

including at the bottom of your SoC% curve, or the midpoint, whatever.

Normal cheap BMS rarely are a good choice for packs that have a need for robust or frequent balancing, best to just use them for their protective features.

 

[ZeroEm]

80%, 90% charging?
by E-HP » Nov 01 2020 2:32pm

It's pretty common knowledge that avoiding charging to 100% will extend a battery's lifetime. Some people charge to 90% capacity, some even lower to 80%. I'm wondering if people use the capacity of the battery, or the voltage in determining 90%, or a percentage of the voltage range (e.g. 3.0V - 4.2V). Looking at the table in the linked article below, it looks like 90%, by capacity, would be around 4.15V/cell. 90% by voltage comes to 4.08V, I think.

I would like to use capacity but have noticed that my capacity varies some with how I used the battery. If I do mostly trail/greenway riding, below 24 kph with assist <=150w vs road riding above 32 kph using assist >=300w and using throttle so pull hills. My watts don't seem to match up with pack Voltage but still I use the Voltage to judge my SOC.

 

[Comrade]

I would like to use capacity but have noticed that my capacity varies some with how I used the battery.

Not surprising given that capacity is inversely related to the discharge rate.

 

[LeftieBiker]

Fist, let's be clear that this applies to lithium batteries only - NOT to lead-acid, which wants a 100% charge like a flower wants the rain. As for EV automotive lithium batteries, it's always a good idea to keep the car at less than 100% (indicated) charge unless you need the full charge shortly, even though the real SOC is almost always lower than indicated. Lithium battery chemistry has improved since the Leaf was introduced in 2010-2011, but it's still better to keep your Leaf, Bolt or Tesla at 80% than it is to keep it fully charged. I also try to keep my e-bikes in the 60-80% range, especially when I'm not using them.

 

[john61ct]

Yes lead needs to spend most of its life at 100% for longevity, and takes a long time to get there no matter how many charging amps are made available.

However just as with LI, the actual SoC% can't be easily determined just from observing voltage, as outlined above.

However, given the long history, battery monitors (SoC meters) are generally more reliably accurate with less fussing, e.g. Merlin SmartGauge

Also the inverse relationship between life cycles and avg DoD% holds for both LI and Pb families,

but whether people pay attention to that or not varies a lot by use case, specific chemistries involved, many factors.

 

[John in CR]

I use a 4.05V/cell top of charge voltage, because that's what the car makers do, and I've noted far slower diminished capacity with use, and almost never need balancing (I'm talking years now that I use automotive grade cells, but even my son's RC Lipo goes many months before it gets more than +/- .03V/cell which poses no danger charging to 4.05V. That said, I also have extra chargers that will charge the packs to 4.15V+ that I use for the rare times I'm going for an extra long ride.

I'm definitely a firm believer in big packs and conservative charging and discharge. It's so nice to never have range anxiety (and I only charge 2-3 times in a typical week of daily use). Plu it's also nice to be my own BMS, so no Battery Murdering System can ever kill my packs.

Also, if you charge to 4.2V capacity diminishes so quickly that if you had used 4.05 all along, it's now more like 90% or more of what you'd have in less than a year of 4.2V charging. Also, charging to full makes tiny out of balance conditions significant and make balancing more necessary.

 

[ZeroEm]

My Battery is not quite two yrs old, would like 5 yrs out of it. It's 1800w, my rides use from 0w to 1300w of what I remember.
Most of my rides are 300w but did not want to charge every time my trike was used. 90% of my charges are <=4.0V, when I ride often will charge to 4.05V - 4.1V and every 3-6 months will charge to full, what I don't like is the charger will charge to 4.225V or 84.5V.
I average charging every three weeks. I know most can not carry a big battery like 15lbs, 140 cells.

 

[John in CR]

....what I don't like is the charger will charge to 4.225V or 84.5V.

Adjust it.

I know most can not carry a big battery like 15lbs, 140 cells.

LOL!!! That's still a small battery.

 

[Comrade]

However just as with LI, the actual SoC% can't be easily determined just from observing voltage, as outlined above.

Are you saying can't be easily determined in real world conditions, like while riding a bike, or can't be determined after a battery voltage settled for days?

 

[john61ct]

Are you saying can't be easily determined in real world conditions, like while riding a bike, or can't be determined after a battery voltage settled for days?

Resting voltage does eliminate one source of gross inaccuracy.

As stated:

If you do want to (try to) correlate a voltage to SoC%, then that is the way to start with an objective and meaningful standard of accurate measurement.

You will still need to painstakingly build a "V vs SoC%" table using an accurate coulometer for the latter.

So why not just use the coulometer as needed? Skip the voltage part.

 

[Comrade]

You will still need to painstakingly build a "V vs SoC%" table using an accurate coulometer for the latter.

But in the case of say an electric vehicle, which will have a life of over 10 years, it clearly relies on coloumb counting given the fast charging and high discharge rates that they are designed for.

How does it keep track of actual state of charge given constant aging of the battery?

 

[E-HP]

The table I referenced in the article has three columns, voltage, charge cycle, and stored energy. The latter two provide a fairly broad range for each charge voltage, since it would appear that several factors could influence those parameters. My question is related to what people use when they say they are charging to a level of less than 100%, not necessarily what the state of charge is, although when people say they charge to 80%, for instance, that could imply energy. I don't really care about state of charge, except if it equates to something I can use to stop the charger at the right level to improve cycle life.

I wanted to stay away from the SOC conversation because it's never fruitful or anything normal ebike folks have the equipment to even estimate. I actually deal with this question on a larger scale, and it's headache, since battery storage doesn't align well traditional methods for valuing energy and capacity.

 

[john61ct]

But in the case of say an electric vehicle, which will have a life of over 10 years, it clearly relies on coloumb counting given the fast charging and high discharge rates that they are designed for.

How does it keep track of actual state of charge given constant aging of the battery?

The EV companies can afford to spend millions on engineers and programmers working for years so the onboard computers can try to take the myriad factors into account.

I'm sure they are constantly iterating improvements over the years based on ongoing feedback.

But the result will still just be a **more** accurate guesstimation.

And when the battery type, chemistry etc changes, all needs to be recalibrated, learning curve in effect starts over.

 

[john61ct]

what people use when they say they are charging to a level of less than 100%, not necessarily what the state of charge is, although when people say they charge to 80%, for instance, that could imply energy.

That's why I keep saying, do not even use SoC% numbers for such statements, just specify the charging profile directly.

There are at least three versions of "100% capacity"

A. cell / pack nameplate rating

B. "theoretical 100%" as per maker spec sheet, e.g. 4.20V and hold CV until current tapers to 0A absorption 

C. "functional usage 100%, for BM reset / calibration" this is up to the user, basically arbitrary

e.g. 4.05V, taper to .03C and Stop

Also could have

D. "usage cycling, CC-only stop-charge"

e.g. charge **to** 4.10V and Stop, no CV

however there is no accuracy with that, the actual capacity point for that version of 100% will vary widely depending on the C-rate offered by the charge source; the higher the rate, the lower the energy stored.

Another factor might be if you use a BMS to top balance, so you might occasionally need to hold the required CV point until that process is complete

hopefully not too high, not too long, not too often.

Anyone can of course do as they like.

But for me, as long as the B. "theoretical 100% Full" point is defined precisely and standardized, and is higher than the C. "functional 100%" to set the BM to

then your **working estimate** of SoC% is the coulometer counting down from that latter point.

To the extent you use your D profile and that is setting a lower capacity, then the BM estimate will get less and less accurate each cycle, until you reset (calibrate) it at your more accurate C. definition of "100%"


The problems come from everyone tossing SoC% around without saying what standard they used to define it.

 

[E-HP]

The problems come from everyone tossing SoC% around without saying what standard they used to define it.

That's why I don't care about the state of charge, but want to know how people charge for better cycle life, and like you're saying, probably using different yardsticks. In the end, it will inform me about whether to set my charger to result in 4.1V/cell (resting), 4.15V/cell or something else that I can actually measure.

 

[john61ct]

whether to set my charger to result in 4.1V/cell (resting), 4.15V/cell or something else that I can actually measure.

Either would be fine, really up to you

If giving up a bit more range is OK, then 4.05V is really prioritizing longevity.

For any voltage setpoint, holding CV for long is equivalent to CC-only at a bit higher voltage.

So even "just stop at" 4.2V is better than what most non-adjustable chargers do.

IMO "just stop at" 4.15 at a decent C-rate, 4.10V at a low current, and 4.05V held CV for a while are all good balanced choices.

If you aren't sitting there watching V&A and terminating charging manually

or setting up an independent HVC to do the job (CC only)

then a lot will depend on the stop algorithm used by the charge source.

Don't get too hung up on "perfection" no such thing, everyone will have a slightly different opinion anyway.

 

[E-HP]

then a lot will depend on the stop algorithm used by the charge source.

This is something that I didn't anticipate when adjusting my charger to a pack level voltage to equate to 4.1V/cell. Since the charger simply stops when it hits that voltage, the resting voltage comes down to something more like 4.04V/cell. I'm charging at 5A, or around 0.18C for my pack. I'm going to play around with the charger voltage to see what results in something closer to 4.1V/cell resting.

 

[john61ct]

Exactly!

Now personally my experience is more with LFP, where the resting target without excess "surface charge" is 3.33 to 3.35Vpc

and to get there in reasonable time charge voltage needs to be about 0.1Vpc higher, at low C rates like that 3.45V is plenty.

Most people and chargers follow the stressful vendor max and (over) charge at 3.6V.

I'm told with these 3.6-3.7 nominal LI chemistries like NMC NCA LMO, the drop between charger setpoint and resting Full voltage is not as great

so others will need to advise you as to the specific numbers for your testing with your specific setup.

But in general, closer to 4.05V is going to be a better default than 4.15V, to the extent you prioritize longevity over range / capacity

 

[ZeroEm]

by E-HP » Nov 02 2020 2:22pm
That's why I don't care about the state of charge, but want to know how people charge for better cycle life, and like you're saying, probably using different yardsticks. In the end, it will inform me about whether to set my charger to result in 4.1V/cell (resting), 4.15V/cell or something else that I can actually measure.

If i'm shooting for 80% or 4V or 4.05 and watching my voltage I go over a Volt because my resting Voltage will be .5 to a volt lower. I charge at 4 amps.

 

[john61ct]

I charge at 4 amps.

Per cell or at the full pack-level?

What c-rate does that translate to?

 

[John in CR]

I charge at 4 amps.

Per cell or at the full pack-level?

WTF? You didn't really ask that did you?