Are there any benefits to fully cycling the battery periodically (full to turtle)?

[LMF5000]

On many electronics, it is commonly believed that draining the battery until the device dies helps to "recalibrate" the fuel gauge.

I've heard mention of this with the leafs as well. It makes a bit of sense - we all know the BMS is just trying to guess true SOC and the only way for it to know for sure how much is in the tank is to discharge it fully. My question is, does that actually happen?

One of my leafs is a 2016 24kWh with 82% SOH and 75% Hx. I notice there's a lot of hidden reserve at the bottom end. The car discharges a little less than 1% per km from 100% down to about 15% indicated charge on the dashboard (stop and go traffic here)... but under 15% it just drives and drives, with the SOC only decreasing very slowly.

Yesterday I purposely drove it all the way down to the last bubble of turtle (and eventually until it completely died), and that took 146km. However my initial prediction by direct proportion was only around 110km, since the battery went from 100% to 15% after 90km.

The whole time I was monitoring leafspy, and at the end it got stuck at 2.0kWh remaining with about 4% GIDS - and that persisted for the last 10km or so. So seems even the car itself didn't know how much was left in the battery at the end - it wasn't just leaving the deliberate buffer at the bottom like we know the leaf usually does.

So, the question is - will the BMS now actually learn how much charge is there, and recalibrate itself somewhat to give me a more linear percentage gauge going forward? Or did I discharge it all the way for nothing?

Does occasional full/empty cycling improve the accuracy of the gauge, or have any other benefits? Does 100% charge help the cells balance for example, or does balancing occur at any SoC?

P.S. For anyone intending to do the same, be advised that if you totally discharge the traction battery until it turns off, the car won't "start" until you charge it back up to 10% indicated on the dash. I thought I'd be able to put in a little charge and drive it the last few meters into my garage, but it turned out to be a 2-hour ordeal with a 16A cable to get the car moving under its own power again.

 

[LeftieBiker]

My advice is to avoid doing that. Even if it does recalibrate the BMS's 'understanding' of capacity and state of charge, it isn't good for the battery. That should take priority, as for that matter it should with your cell phone. For many years, and for all I know even now, a lot of devices with lithium batteries still come with outdated instructions, left over from previous battery chemistries, to fully cycle the battery. There really is no good reason to do that.

 

[Stanton]

The only time I would drive a battery to "turtle" is if you are trying to balance the pack after installing new cell(s)/module(s). Other than that, I agree with Lefty: no real reason to stress the pack that much. FWIW, I don't even do that with my CE devices (laptops/phones/drones/etc); in fact, I try to stay above 20% SoC on all Lithium-based battery products.

 

[knightmb]

On many electronics, it is commonly believed that draining the battery until the device dies helps to "recalibrate" the fuel gauge.

I've heard mention of this with the leafs as well. It makes a bit of sense - we all know the BMS is just trying to guess true SOC and the only way for it to know for sure how much is in the tank is to discharge it fully. My question is, does that actually happen?

I do this rarely a one or two times a year, but that leaves out of lot of "how" I am doing it. I take precautions to avoid stressing the pack out and it takes hours to hit the bottom carefully. I would not drive my Leaf past 0% (unless it was an emergency) to just randomly hit the pack shutdown voltage. So yeah, you can get the BMS to better learn the battery capacity and even balance out the pack better for the future. But, as others have already said, this is "stressing" the pack and you have to take precautions ahead of time. If you suspect you have some cell balancing issues or the BMS is being way to conservative with your range, this is one of the only ways to test the battery without actually taking it apart or needing diagnostic equipment, but it won't be definitive and the data you collect will take a little analysis.

One of my leafs is a 2016 24kWh with 82% SOH and 75% Hx. I notice there's a lot of hidden reserve at the bottom end. The car discharges a little less than 1% per km from 100% down to about 15% indicated charge on the dashboard (stop and go traffic here)... but under 15% it just drives and drives, with the SOC only decreasing very slowly.

Lithium batteries don't have a linear discharge curve like an AGM for example. So you can't really look at the discharge rate per distance and get an exact picture of how much capacity you have left. That is why the Leaf BMS is actually counting energy in and energy out to make predictions on the battery state.

Yesterday I purposely drove it all the way down to the last bubble of turtle (and eventually until it completely died), and that took 146km. However my initial prediction by direct proportion was only around 110km, since the battery went from 100% to 15% after 90km.

The whole time I was monitoring leafspy, and at the end it got stuck at 2.0kWh remaining with about 4% GIDS - and that persisted for the last 10km or so. So seems even the car itself didn't know how much was left in the battery at the end - it wasn't just leaving the deliberate buffer at the bottom like we know the leaf usually does.

You are correct, it is predicting based on previous measurements. This would indicate there was more capacity than it predicted so as long as the battery voltage stays above turtle mode (and the final shutdown), it will let you keep going.

So, the question is - will the BMS now actually learn how much charge is there, and recalibrate itself somewhat to give me a more linear percentage gauge going forward? Or did I discharge it all the way for nothing?

The BMS did learn from the experience and will use that data to better predict the mileage in the future, so it wasn't for nothing. But...

Does occasional full/empty cycling improve the accuracy of the gauge, or have any other benefits? Does 100% charge help the cells balance for example, or does balancing occur at any SoC?

P.S. For anyone intending to do the same, be advised that if you totally discharge the traction battery until it turns off, the car won't "start" until you charge it back up to 10% indicated on the dash. I thought I'd be able to put in a little charge and drive it the last few meters into my garage, but it turned out to be a 2-hour ordeal with a 16A cable to get the car moving under its own power again.

It will and yes, it can help with balancing out cells if they stray too much but are still as healthy as the other cells. But... the way you went about fully discharging the pack is probably not the best way to do it. In an emergency situation where you need to squeeze out every last watt of power to reach somewhere, I won't argue with that. But if you are doing it on purpose, the drive method is the most stressful way to do this. The least stressful way is to have both a pack that is below 82F and an energy draw low enough along with some LeafSpy monitoring to watch the cell performance. This means that when draining the pack beyond 0%, the Leaf is usually stationary with a small load running for the next few hours to slowly and evenly drain all the cells. If could be running the AC or Heat in a low mode to slowly siphon away the power until you get it as close to shut down as possible or if you hit shutdown by accident, be prepared to plug in a L1 or L2 to begin charging the battery immediately once those are reached so you are not sitting in a such a low SOC for so long as to cause long term damage. Charge all the way to 100% so the BMS can measure the entire session and then figure out how much capacity the battery really has.

As others have said, not recommended, but if you are going to do this, do it the less stressful way possible.

 

[LMF5000]

I understand how it's preferable to discharge it gently- I've done that before but this time round I wanted to measure the actual mileage the car had in it below the "---%" battery indication (to see how far I can push it in a real emergency), and the only way to do that (short of having a sophisticated rolling road) was to drive the car. I drove until the very first bubble of power disappeared, indicating one cell was dangerously close to dipping below 2.75V. At that point I rerouted to get to the garage via the shortest, slowest road possible, a distance of about 0.5km at 20km/h. Over that time it slowly lost bubbles, then once in the garage with the A/C on it was completely depleted in a few minutes.

I then gave it a full charge to 100% and left it plugged in for a further two hours. Today I've put almost 40km on it and the battery is down to 66%, so already the gauge is behaving more linearly (it used to decrease very quickly when the SoC was over 80%).

Crucially, the Hx value is still the same, at 75%, and SoH has risen slightly, from 81.9% to 82.1%. I suspect the BMS is realizing that there was a lot of battery capacity at the low end that it thought had gone, but now that it knows it's there and usable it's revising its health estimate.

Another possibility is the recovery effect of lithium batteries at low charge. The only place I've seen this documented is in one NASA publication which I will link below, but in a nutshell they discovered that some of the capacity losses induced on lithium ion cells that were charged at 4.1V could be reversed by letting the cells dwell at a very low voltage (below 3.6V), for up to a year. Apparently that causes the redistribution of lithium ions and some of the ones that were previously inactivated become available again, reversing some of the cell degradation.

That's my interpretation anyway... here's the study for anyone who wants to form their own interpretation: https://www.nasa.gov/sites/default/files/atoms/files/cap_rec_li-ion_cells_azimmerman.pdf

On the 24kWh leaf, to take cell voltages below 3.6V/cell requires a very deep discharge, corresponding to less than 2-3kWh remaining in the pack (about 3-5 GIDs or "---%" on the dashboard display). I thought that letting the battery dwell down there for an hour might restore some capacity like it did in the NASA study. One day I need to try this on some severely degraded R/C model lipo batteries in my collection to test the hypothesis more scientifically.

 

[LeftieBiker]

I'm replying to this now in order to provide a counterpoint for anyone reading the above in the future. It seems to me that the risk of damaging battery cells outweighs the possible benefit of slightly improving the SOH. Things don't always happen the same way in the Real World as they do in laboratory conditions...

 

[Astral]

since the battery went from 100% to 15% after 90km.

So, the question is - will the BMS now actually learn how much charge is there, and recalibrate itself somewhat to give me a more linear percentage gauge going forward?

SO, what is your conclusion? How much kilometers do you do after the calibration from 100 to 15%?

Probably you are in the best position to answer your question by yourself.

 

[Dooglas]

Crucially, the Hx value is still the same, at 75%, and SoH has risen slightly, from 81.9% to 82.1%. I suspect the BMS is realizing that there was a lot of battery capacity at the low end that it thought had gone, but now that it knows it's there and usable it's revising its health estimate.

Your numbers don't appear to support the idea that the BMS "discovered" a lot of unused battery capacity. I can't help but wonder just how much variation there is in the measurement of SoH. I suspect those numbers may indicate no significant difference in measured SoH before and after your experiment.

 

[Akuma]

On many electronics, it is commonly believed that draining the battery until the device dies helps to "recalibrate" the fuel gauge.

I've heard mention of this with the leafs as well. It makes a bit of sense - we all know the BMS is just trying to guess true SOC and the only way for it to know for sure how much is in the tank is to discharge it fully. My question is, does that actually happen?

One of my leafs is a 2016 24kWh with 82% SOH and 75% Hx. I notice there's a lot of hidden reserve at the bottom end. The car discharges a little less than 1% per km from 100% down to about 15% indicated charge on the dashboard (stop and go traffic here)... but under 15% it just drives and drives, with the SOC only decreasing very slowly.

Yesterday I purposely drove it all the way down to the last bubble of turtle (and eventually until it completely died), and that took 146km. However my initial prediction by direct proportion was only around 110km, since the battery went from 100% to 15% after 90km.

The whole time I was monitoring leafspy, and at the end it got stuck at 2.0kWh remaining with about 4% GIDS - and that persisted for the last 10km or so. So seems even the car itself didn't know how much was left in the battery at the end - it wasn't just leaving the deliberate buffer at the bottom like we know the leaf usually does.

So, the question is - will the BMS now actually learn how much charge is there, and recalibrate itself somewhat to give me a more linear percentage gauge going forward? Or did I discharge it all the way for nothing?

Does occasional full/empty cycling improve the accuracy of the gauge, or have any other benefits? Does 100% charge help the cells balance for example, or does balancing occur at any SoC?

P.S. For anyone intending to do the same, be advised that if you totally discharge the traction battery until it turns off, the car won't "start" until you charge it back up to 10% indicated on the dash. I thought I'd be able to put in a little charge and drive it the last few meters into my garage, but it turned out to be a 2-hour ordeal with a 16A cable to get the car moving under its own power again.

Let me expose some facts regarding this myth about the Nissan Leaf not calibrating the battery cells.
That's simply not true, the Nissan Leaf always calibrates the battery cells at the end of the charge. Hadn't you notice, that when you charge the car and you check in the application that it estimate for example 2 hours to finish charging completely, then when you open again the aplication and check the charge bar level is complete it says that the estimate time to finish is for example 40 minutes. BUT then, when you get again in the application after 40 minutes, it still say that there's another 40 or 30 minutes left to finish charging?
Maybe you thought that there was a bug or error in the charge app or it's inaccurate meassuring it, but that's not the case. The car in fact has finished charging the battery, and what it's doing till completion is calibrating the cells of the battery.
You can see this using Leaf Spy, that you'll se that there's no more Kwh to add to the current SoC of the battery but the led light still blinks. It's simply the calibration phase and it's always done at the end of the charging process.
That's why from time to time is good for the battery to charge it to 100%, to let it calibrate the cells. If you always charge it only to the 80%, it won't calibrate anything, just finish charging at 79%-80%.

This method of calibrating the battery is done by every systems I'm aware of, even those batteries used for solar panel energy storage.

Remember that calibrating the battery cells won't give you more energy or revitalize dead cells, it simply adecuates the charge level for every cells so you can have a better percent usage measure of the battery in each moment.

Cheers.

 

[woeful]

Charging to "100%" balances the state of charge of each cell pair, as well as providing the BMS with a check as to how much charge the battery really holds. Cell balancing is needed for long life of any battery pack connected in series, though it happens chemically rather than electronically for a lead-acid battery.

For lithium ion cells, all the battery management electronics can measure is the temperature, the voltage across the cell and the current running through it. The chemistry of the cell is such that the voltage hardly changes between about 20% and 80% charged, so the state of charge in this range is estimated by how much current has gone in or out. It's probably no coincidence that this is also the range where the battery cell deterioration is least: the reactions going on are different from the under-20 and over-80 states.

The effect of that is that gradually the BMS loses track of the battery capacity, and each cell can drift towards a higher or lower state of charge than the average. Charging above 80% or running down below 20% gives the BMS an update on the true state of charge, based on the voltage. It also gives a chance to boost the less-charged or bleed alittle charge out of the most-charged, to "balance" the battery pack. That genuinely increases its capacity, because that's limited by the maximum charge in any cell (at the top end) and the minimum charge in any cell (at the bottom end). If every cell reaches maximum charge more or less together, the capacity limit is minimised.

There's a set of low-power resistors which can alter the charge current slightly for each individual cell. When the battery is almost fully charged, and the charge current is very small, the weakest cells can be boosted (by reducing the curent though the others). That's what happens for than last 30 to 60 minutes of a 100% charge. It's good for the battery, provided not done too often, and provided the battery is run down a bit within a day or so afterwards.

Hence the advice - charge overnight to 100% about once a month, when you confidently expect to be driving at least 10 miles next day.

 

[LMF5000]

Sorry for missing all your replies, my notifications must have been switched off.

To answer the question, fully discharging the battery all the way down to turtle did indeed make a massive difference. The BMS was vastly underestimating the amount of charge available in the battery.

Before doing the exercise, the charge level would decline by 1% for every 1km driven. So I'd start at 100% and after 80km I'd be at around 20%. When I did the test, the percentage gauge first got stuck at 15%. That means, SoC reading remained saying 15% while I put on about 5km of driving. Then it got stuck again at 9% SoC. Then the display turned to "---" and I had to use LeafSpy.

THEN LeafSpy itself got stuck, at around 1.2kWh remaining. By this time I was slowly orbiting my garage, waiting for the first sign of turtle to divert straight back to the garage and my waiting granny charger.

Guess how long it took in all to actually drain the battery? 50km. That's right, my battery that was losing 1% of charge for every 1km driven managed to do a whole 50km after the car thought it was empty, for a total range of 158km on a single charge.

I also learned two important lessons:
1. As soon as the first bubble of turtle appears, the rest aren't proportional at all. I went from one missing bubble to almost all bubbles missing within the span of about 100 meters.

2. If you let the car die completely because you pause for a few minutes to make space for it in the garage, the car doesn't re-start immediately when you plug it in. You actually have to charge it to at least 10% SoC for it to "start" and give you motor power. Found that out the hard way, had to wait an hour charging it outside on the granny cable because I lacked the courage/strength to push it single-handedly into the slight uphill of my garage and executing a 90-degree turn and avoiding the garage walls while pushing and steering the car through the open driver's side window.

After that exercise and fully charging the car, the displays became much more accurate. More indicated range, more linear decline of SoC against kilometers driven. It lasted about a year, now the car is back to its old ways where SoC declines rapidly just after a charge, is linear in the middle, then declines slowly at the low end. Maybe I will redo the calibration some day. But I'm planning a gearbox oil change, full service, new tyres, possibly wheel alignment in the near future so any calibration will be done after all this to see what effect (if any) they have on ultimate range.

 

[Joaozinho]

Do you think one could get the same effect by running the heaters to discharge the battery?

 

[LeftieBiker]

Do you think one could get the same effect by running the heaters to discharge the battery?

I don't see why not. The heat pump doesn't use a lot of power, but the resistance heater certainly does. More than driving gently on level terrain.

 

[DaveinOlyWA]

On many electronics, it is commonly believed that draining the battery until the device dies helps to "recalibrate" the fuel gauge.

I've heard mention of this with the leafs as well. It makes a bit of sense - we all know the BMS is just trying to guess true SOC and the only way for it to know for sure how much is in the tank is to discharge it fully. My question is, does that actually happen?

One of my leafs is a 2016 24kWh with 82% SOH and 75% Hx. I notice there's a lot of hidden reserve at the bottom end. The car discharges a little less than 1% per km from 100% down to about 15% indicated charge on the dashboard (stop and go traffic here)... but under 15% it just drives and drives, with the SOC only decreasing very slowly.

Yesterday I purposely drove it all the way down to the last bubble of turtle (and eventually until it completely died), and that took 146km. However my initial prediction by direct proportion was only around 110km, since the battery went from 100% to 15% after 90km.

The whole time I was monitoring leafspy, and at the end it got stuck at 2.0kWh remaining with about 4% GIDS - and that persisted for the last 10km or so. So seems even the car itself didn't know how much was left in the battery at the end - it wasn't just leaving the deliberate buffer at the bottom like we know the leaf usually does.

So, the question is - will the BMS now actually learn how much charge is there, and recalibrate itself somewhat to give me a more linear percentage gauge going forward? Or did I discharge it all the way for nothing?

Does occasional full/empty cycling improve the accuracy of the gauge, or have any other benefits? Does 100% charge help the cells balance for example, or does balancing occur at any SoC?

P.S. For anyone intending to do the same, be advised that if you totally discharge the traction battery until it turns off, the car won't "start" until you charge it back up to 10% indicated on the dash. I thought I'd be able to put in a little charge and drive it the last few meters into my garage, but it turned out to be a 2-hour ordeal with a 16A cable to get the car moving under its own power again.

Not only ZERO benefit but very detrimental.

Center your SOC as much as possible. Be it 45-55% or 20-80%, Lithium lasts longer when extremely low/high SOC is avoided. Realize the LEAF dash gauge is shifted towards zero (means 40% on the dash is actually closer to 50%) at lower SOC readings.

You may have gotten this impression from Tesla recommendations to charge to full once a month. That has nothing to do with battery health. It is a calibration for the range calculator. DO NOT DO IT.

 

[benjwhite9]

I don't have any inside information on this subject, but my experience with my own 2019 Nissan Leaf (the 40 kWh pack) is that I rarely let the SOC go above 85% or below 15%. And I've seen nothing funny with the SOC indicator on the dash, it seems to be very consistent. In other words, I've never seen the problem described above, where the indicated driving range is vastly less (or vastly greater) than actual.

 

[Akuma]

I don't see why not. The heat pump doesn't use a lot of power, but the resistance heater certainly does. More than driving gently on level terrain.

This won't work. At turtle level, the resistance heater and air pump are switched off and can't be switched on again to save energy for the electric engine.
Even when the Nissan Leaf traction battery hasn't enough energy to move the wheels, the traction battery isn't completely empty.
With less than 5 GIDs and 2,1% of available SoC (0,4 kwh) the car set the N position till you or the car itself stops moving. Then after that you can try to power on but it'll refuse to move by not letting you to put on D, B or R positions.

As I said before, I don't see the point of calibrating the cells this way instead of doing it charging to 100% from time to time. It's much more harmful to discharge this kind of battery to this level than to charge it to 100%.
If you have to do it (charge or discharge), the trick to not damage the battery is to charge it after got discharged or to use the car after gets fully 100% charged. You don't need to do this inmediatly, but don't leave the car in that state for too long.

Cheers.

 

[LeftieBiker]

Thanks - I'd forgotten that Turtle mode disables the heat. I've never experienced it. I agree completely that this isn't the way to go - especially if you want to keep that battery for many years.

 

[pmbdk]

Center your SOC as much as possible. Be it 45-55% or 20-80%, Lithium lasts longer when extremely low/high SOC is avoided. Realize the LEAF dash gauge is shifted towards zero (means 40% on the dash is actually closer to 50%) at lower SOC readings.

Very good advice!

45-55% is better than 20-80%.

I also recently saw a NMC Li-ion battery study where they claimed that in cold weather, it was actually better to be quite a bit on the low side (20-40%) regarding longevity. For the Leaf this is even more important imo as (as you say) 50% on the dash is actually more than 50%. However, when you look into real-life EV batteries, I think it is important also to take into account the differences in cell voltages, where some of the cells in a 30% pack SoC may be at 20% while other in the same pack may be at 40%. Looking at it that way, just staying at around 50% is probably the best advice...

 

[Joaozinho]

Thanks for all the advice.

 

[decimal]

I would advise that you do what works for you and don't worry too much about how you charge. I charge to 80% most of the time because it's easy to set that option on my early Leaf, and that's all I need each day. Sometimes I'll do a 100% charge, and I'll run the car down to zero when needed. Nothing will make the battery regain lost capacity. The BMS will do what the BMS gotta do, and I just ignore the guess-o-meter range estimations. When planning a journey, I look at my capacity in kWh remaining, then multiply by 2.4 to get the worst-case, usable capacity in miles. This seems to be accurate for my car, style of driving, and hills where I live!