Welcome to Tesla Motors Club
Discuss Tesla's Model S, Model 3, Model X, Model Y, Cybertruck, Roadster and More.
Register

How I Recovered Half of my Battery's Lost Capacity

This site may earn commission on affiliate links.
This is way off topic...so will be my last comment on this...but:
I don’t know that there is any evidence that charging at 250kW at 10% SoC is any more stressful than charging at ~170kW at 40% SoC. For all we know 250kW at 10% SoC may be less stressful. All depends on how fast those ions can diffuse I think?

So I’m not sure that following a non-optimal charging strategy on a road trip to attempt to preserve the battery is actually helpful for preserving the battery.

A better strategy for battery preservation might be to avoid V3 Superchargers I suppose. But they have the distinct advantages of being much faster, and also actually working, for now (many V2 Superchargers are broken in my experience - providing ~118kW rather than 150kW, for example - about 50% of the time I have to move stalls to get a higher rate). You could additionally search out partially broken V2 superchargers to further preserve your battery (it’s usually not hard to find one that will charge at 30-72kW, if you try them all). The overhead in searching them out doesn’t matter since time is obviously not a factor. Of course, going immediately to a shared charger with another vehicle is another winning strategy here (probably would annoy the other driver though).

I would have had a harder time driving from San Diego to Portland in a day without V3 Superchargers. Possible still (assuming the V2 were working properly), but the V3 sure are nice. 1076 miles at an average speed of 58mph! (~18.5 hours)

I was responding (in part) to the posting in this thread of texts from a service advisor. It included this:
If the battery has a large depth of discharge to a low state of charge (Red zone), allowing the battery to rest for a few hours before recharging will minimize the stress on the cells

So, to me, that suggests that running down to 20% then immediately supercharging is not so battery friendly.
If you were willing to park the car for a few hours before starting the supercharging maybe not so bad, but that is inconvenient.

Maybe it is my own superstition, but based on what I have read, I try to keep my battery above 40% as much as possible.

Also, back to what the original poster said about changing his charging routine to intentionally let it get down to 20% regularly to get more BMS readings... I think the service advisor suggestions would indicate that you should leave it at 20% for a few hours before you start charging, so hopefully that is on a charge timer that delays charging before starting. But, on the other-other hand, I have heard that leaving battery at low SoC for long periods is not good either, so don't delay too long. Again, I myself am not going to do that 20% thing to get better readings.

I also think Tesla doesn't want their average customer to feel that they have to be a battery chemistry expert. So, many service advisors are hesitant to say too much in this area. The "party line" is that the BMS and battery chemistry is so good that you can do whatever you want and it will take care of itself. Those of us trying to "second guess" best charging routines could get lumped into the "overly OCD" category.
 
Last edited:
I don’t know that there is any evidence that charging at 250kW at 10% SoC is any more stressful than charging at ~170kW at 40% SoC. For all we know 250kW at 10% SoC may be less stressful. All depends on how fast those ions can diffuse I think?

I think there are way to many variables to know what is more stressful. Or for that matter what charge rate you might be able to get with a particular f/w version - assuming a charge station is capable of delivering at its rated power.

Given by graphs like this (from TeslaLogger) showing max charge rates vs firmware version, Tesla might not be sure of the effects / benefits / stresses either.

Tesla Model 3 LR Charging.jpg
 
Last edited:
So, to me, that suggests that running down to 20% then immediately supercharging is not so battery friendly.

That's my understanding, especially if you were discharging at a highish rate. Not all of the cells give up energy / charge at exactly the same rate, so at the moment you stop driving there could be some energy imbalance. Leaving the battery to rest allows the charge to equalize before starting charging.Also, it is reasonable that the BMS also needs some open-circuit time to work out / learn / maintain view of what balancing is likely to be needed in future.

There are likely other factors though, as warm cells have lower internal resistance which is a good thing for charging.
 
Last edited:
For what it is worth, I attended a presentation by some Nissan battery engineers years ago and they were being grilled about the inaccuracy of their range gauge.
They showed charts of everything the computer was considering and it was extensive. Charge in, charge out, current levels, temperatures, voltages, etc.
But there was an amusing little box saying something like "spooky unpredictable battery chemistry effects" which they said they couldn't measure.
Even though Tesla's measurements are called into question sometimes, and some voodoo is suggested to help the computer learn better, Tesla still has done an admirable job of getting things "pretty darn close most of the time."
The whole industry struggles with this sort of problem.
 
Screenshot_20200724-2.jpg Screenshot_20200724-1.jpg

Great post and good discussion.

I have virtually the same charge profile over the last few months, recovering most of the battery's 'lost' capacity. It happened by accident but there are some similarities in the charging routines.

I started charging much less frequently at the end of May simply because of work-at home routine. I began letting the battery SoC get under 80 miles or so before plugging it in - often several days in between. I also dropped the charge level to ~ 80%.

The estimated full range rose steadily back almost to the fleet average and has stayed constant since then.

It was a total mystery until I saw this post, recognized the plot, read the discussion and noticed some similarities with my accidental changes to my charging routine. So thanks again, great thread.

2018 M3D extended range, FSD
 
"spooky unpredictable battery chemistry effects"

There are so many interdependencies that trying to identify them all at an 'ion' level as well as which are 'causes' and which 'effects' must be difficult to say the least, even if every cell was absolutely identical.

Tesla still has done an admirable job of getting things "pretty darn close most of the time."

I wouldn't argue with that.
 
Last edited:
OP: Thank you for this very detailed and informative post.

This tends to agree with what I've always supposed, that people saying they have lost range are in many cases misguided because what they are seeing is nor battery degradation but mis-reading. Does anyone ever actually check there range by running the car until the battery is at 0%? Of course not.
 
But on the other-other hand I have heard that leaving battery at low SoC for long periods is not good either, so don't delay too long. Again, I myself am not going to do that 20% thing to get better readings.
Fast charging a battery at low SOC is completely different than storing a battery at low SOC. The Tesla documentation (page 17) to the EPA states:

“To maintain service life, the battery pack should be stored at a state of charge (SOC) of 15% to 50%.”
 
Last edited:
I began letting the battery SoC get under 80 miles or so before plugging it in - often several days in between. I also dropped the charge level to ~ 80%.

The estimated full range rose steadily back almost to the fleet average and has stayed constant since then.

Were there other changes to your charging routine? eg: use of superchargers.

Looking at your post makes me wonder if the title makes sense. If this is showing 'battery degradation' then you just reversed degradation. If we are looking at 'calibration error' then you change in charging allowed the BMS to revise its calculations. If it is showing 'Inacessible capacity due to imbalance-related factors' then your new routine has addressed that.


Tesla M3 degradation report.jpg
 
First night of Deep sleep, I turned off Sentry and used TezLab to put it in deep sleep, probably 6+ hours or so, as you can see this is a slight bump. After waking the M3 and checking out these stats I put her back to sleep, have not charged at all, currently at 58% and on the battery health pictures looks like a slight bump up. Will wake her tomorrow morning and see if this increases.

2AD4A3B1-E47A-4AE3-9D87-E6BDF48ABC6F.jpeg


D864818E-FDE1-4165-B8E1-13F14698452C.jpeg



968B3C46-89E8-41FE-8C07-F7CC2030F63E.jpeg
 
I guess I don't understand. If there is really a suggested regime for charging to make the battery last as long as possible it should be in the user manual. If not, why not? Either Tesla stands behind the charging regime or they don't. We shouldn't have to guess at this. Here is an idea - put an option on the menu that forces the Tesla to do whatever it has to correctly estimate the true range of the current battery (hey Tesla could call this a new feature!).
 
So, to me, that suggests that running down to 20% then immediately supercharging is not so battery friendly.
If you were willing to park the car for a few hours before starting the supercharging maybe not so bad, but that is inconvenient.

Maybe it is my own superstition, but based on what I have read, I try to keep my battery above 40% as much as possible.

Also, back to what the original poster said about changing his charging routine to intentionally let it get down to 20% regularly to get more BMS readings... I think the service advisor suggestions would indicate that you should leave it at 20% for a few hours before you start charging, so hopefully that is on a charge timer that delays charging before starting. But, on the other-other hand, I have heard that leaving battery at low SoC for long periods is not good either, so don't delay too long. Again, I myself am not going to do that 20% thing to get better readings.

Just to clarify, I posted that text exchange because i found it interesting that the idea is now coming from two locations. I have never heard of it here or anywhere else, so I had my doubts.

As far as waiting a few hours if the battery is in the red zone - the recommendation wasn’t intended to be cut and dry. I asked for further clarification at the time (which I’m not going to post) and the response was “do it at your home/destination when the road trip is over”. Combining the OPs post with what my service advisor stated, it seems the intent is for the BMS to “have a better view of the bottom”. The official recommendation was do your best to avoid routine discharges below 20%.

Again, I just felt it was interesting that there was another instance where the same advice was given - I felt I should share.
 
  • Like
Reactions: Transformer
I guess I don't understand. If there is really a suggested regime for charging to make the battery last as long as possible it should be in the user manual. If not, why not? Either Tesla stands behind the charging regime or they don't. We shouldn't have to guess at this. Here is an idea - put an option on the menu that forces the Tesla to do whatever it has to correctly estimate the true range of the current battery (hey Tesla could call this a new feature!).

It would probably read like:

Only use the middle 20-80% SOC.
Don't accelerate or decelerate hard.
Don't charge when cold.
Don't supercharge.
Take a two month course to understand battery design, electro-chemical reactions and electronic control systems.
Install additional monitorring equipment.
Spend hours a week analysing your use and vehicle data.

Then maybe you might get some small benefit.

Or, you can opt to trust Tesla to do that stuff for you and for the most part just get on and drive the car.

Different approaches will doubtless suit different owners.
 
It would probably read like:

Only use the middle 20-80% SOC.
Don't accelerate or decelerate hard.
Don't charge when cold.
Don't supercharge.
Take a two month course to understand battery design, electro-chemical reactions and electronic control systems.
Install additional monitoring equipment.
Spend hours a week analysing your use and vehicle data.

Then maybe you might get some small benefit.

Or, you can opt to trust Tesla to do that stuff for you and for the most part just get on and drive the car.

Different approaches will doubtless suit different owners.
On a daily basis, I try to use your first suggestion but I charge between 80% to 90% but closer to 80 than to 90. On a performance model Y that works out to 241 miles on a daily charge meaning that I'm 50 miles less than the maximum of 100% which to me seems reasonable. It also means that within 2 hours I could charge it up to 100% if I needed/wanted to take a trip. I live in the northern CA coastal mountains just west of the fires and I always want to have sufficient charge to evacuate should it become necessary. I'm a new model Y owner so much of this is new to me. The 20-80% SOC suggestion is a good one but that leaves us with 60% of the battery. That would probably be hard to put into a manual for all to see. No wonder a 500 mile battery sounds like a good idea if 60% is really what you have to work with (300 miles of daily usage). My Covid 19 truth is that I rarely need even 50 miles. Nothing is open. I'm hoping that will change.
 
Not all of the cells give up energy / charge at exactly the same rate, so at the moment you stop driving there could be some energy imbalance.

Imbalance is caused by the initial energy capacity of a brick being lower (imagine a brick with 45 cells in parallel rather than 46). Roughly speaking, the bricks will discharge energy at the same rate (though not exactly the same rate, as current will be necessarily the same in each brick, but brick voltages will deviate over time during the discharge event). The brick imbalance (as I understand it) is really a voltage imbalance caused by initial energy content differences (and the imbalance I think can occur after both discharge and recharging events). There are a lot of possible scenarios. But in a simplistic case, I believe your battery capacity at 100% is approximately limited by the energy content of your minimum energy brick (since it will become empty first (since energy discharge rate is roughly the same for all bricks), and all bricks discharge in series so there is no way to extract the energy from the other bricks (caveat: that is the case without some downtime to somehow rebalance, but I think rebalancing is only a thing done at high SoC (above 90%), not low SoC - don’t think there is a provision to transfer energy from a healthy brick to a weak brick at low SoC - my understanding is rebalancing is done at high SoC by bleeding down (wasting energy) the (usually?) lowest capacity, highest voltage brick, so that charging of all other stronger (lower voltage at that instant) bricks can continue (resulting in a higher overall pack voltage after a subsequent top-up charge, which of course maximizes energy available)). But that weak brick will still be the limit during a discharge. But the “area under the curve” will be higher since you start with a higher pack voltage than you would have without rebalancing, after the top up. The CAC (calculated Amp-hour capacity) is a measure of the amp-hour capacity of that worst brick. But the energy available can be improved by other healthy bricks - since a healthy brick reduces its voltage slightly more slowly during a discharge than a weaker brick (for an identical current draw), the pack voltage will be higher for the discharge period after a rebalance, meaning there is more energy under the I*V Vs. time curve (power integrated over time is energy).

Does anyone ever actually check there range by running the car until the battery is at 0%? Of course not.

I’m not sure why we keep coming back to this. People do discharges to 5% all the time on road trips, and this is plenty to notice what is your range! Loss of capacity (which is real!) happens, and matters! How much? Not much. I’ve lost a true 10% capacity (287 rated miles, 70.3kWh, vs 78kWh (over 310 rated miles when using the same energy content per rated miles but rated miles are initially swollen to minimize customer consternation and to allow everyone to have 310 (pre-2020) rated miles when new, even if the energy contents of different brand-new packs differed by 1-2kWh. ) Do I notice loss of capacity? Sure. I can’t make certain stretches to Superchargers that I could have made before. Does it matter? Not too much at all, as long as Superchargers are spaced sufficiently closely. Slightly longer waits at Superchargers are needed to make the same legs now.

But loss of capacity is real; people measure it accurately all the time (either through a discharge monitor, or a charging event measurement). This is not something we need to continue to debate. It happens (to everyone!!!), it's normal, and it's fine (unless it's excessive).
 
Last edited:
Will wake her tomorrow morning and see if this increases.

Remember to ignore data points from Stats, except the ones reported with SoC greater than 90%. All the 50% extrapolated points and lower than 90% points can be ignored. They may indicate recovery, but they can and should be aggressively ignored until you have a good datapoint at 90% charge or higher (then you can decide whether they were reasonable datapoints post hoc). Obviously there is a lot of noise on the Stats results (it's good as an estimate of capacity +/- 5-7 rated miles).
 
The estimated full range rose steadily back almost to the fleet average and has stayed constant since then.

Remember to not draw conclusions about the cause, if you have done any software updates over the period. They're potentially confounding factors. Doesn't mean that what you're seeing isn't real (it likely is), but we just don't know what caused it.
 
  • Like
Reactions: ediflyer and Zoomit
I believe your battery capacity at 100% is approximately limited by the energy content of your minimum energy brick

My version (which I don't think contradicts yours but is still a theory) is that cells can weaken in (at least) two modes, one making them become internally leaky / shorty / lower internal resistance, and the other making them have higher internal resistance. Temperature and other conditions effect cell behavior too. Bricks exhibit behavior resulting from the mix of cell characteristics that make them up. A given brick with certain cell make up and resulting charateristics (but remember it's a characteristic resulting from the mix of cells in each brick) may increase in voltage quicker than the others. All the bricks have to be charged together (albeit at slightly different rates as controlled by balancing resistors). The better the picture BMS has of how the energy balance looks at the end of discharging, the better idea it has of which bricks had absorbed most energy / charge during previous charging and what needs to be done balance-wise to ensure that during the next charge cycle bricks take on evenly balanced charge. As bricks charge and their voltages rise, one brick will hit the maximum permitted cell voltage first at which point charging has to stop. The slower you charge, the longer the balancing system has to maintain charge balance. The better balanced the pack is, the less balancing is needed to get the most even energy distribution between cells as your 'weakest' brick reaches max voltage.

Same sort of thing happens at the fully discharged end, where the cell that had least energy reaches mimimum allowed voltage and require discharging to be stopped. At each end there will be unused energy or capacity in the cells that did not reach max / min voltage.
 
Last edited:
  • Like
Reactions: Phlier