It ensures the battery cells can't be over-charged or over-discharged.

 

Incorrectly estimate the amount of energy you have available in your battery.

 

Battery management system - Wikipedia

en.m.wikipedia.org

Any questions?

 

Also monitors the battery module temperatures, provides SOC data for the dash, probably also for the GOM.

 

It tells the DC rapid charger how many or how few amps to give your battery, depending on temperature, the phase of the moon, various other hidden factors.

 

Does a BMS control charging at 80% charge ?

 

It controls all charging and on some EVs you can set it to stop at 80%, otherwise charging will continue to 100% and will be slower and slower the higher above 80% it gets.

 

BMS stands for Battery Management System; its job is to monitor and manage the state of the battery.
Whenever you're charging, the BMS will oversee this, monitoring things like temperature, charge speed, etc and managing these. If you've set the charge limit to 80%, the BMS will be what manages this and calls a halt to charging when 80% is reached, even if the vehicle is still connected to a charger.

 

The reason i asked is i said this on another forum.

The reason it takes longer to charge to 100% is for the last 20% the BMS (BATTERY MANAGEMENT SYSTEM) will slow the charging rate down, 1 to let the battery cool, 2 to balance the battery cells so they are all charged to the same level, an EV should be charged to 100% at least once a month just for balancing. No need to charge 100% every time, because I only charge once a week I charge to 100% but if I needed to charge every day I would only charge to 80%.
I have only used a public charger once and charged at 90 kWh for 20 mins just to make sure everything worked.
The Corsa EV will charge at a rate up to 100 kWh and it's the car that has the charger built-in, not the charge point that's just a power supply.

 

The reason i asked is i said this on another forum.

The reason it takes longer to charge to 100% is for the last 20% the BMS (BATTERY MANAGEMENT SYSTEM) will slow the charging rate down, 1 to let the battery cool, 2 to balance the battery cells so they are all charged to the same level, an EV should be charged to 100% at least once a month just for balancing. No need to charge 100% every time, because I only charge once a week I charge to 100% but if I needed to charge every day I would only charge to 80%.
I have only used a public charger once and charged at 90 kWh for 20 mins just to make sure everything worked.
The Corsa EV will charge at a rate up to 100 kWh and it's the car that has the charger built-in, not the charge point that's just a power supply.

Wrong.
On a DC charger, the 400V DC goes direct to the traction battery whilst being monitored by the BMS. It is this that will command the charger to alter the charging current at stop at 80% or 100%.

The onboard charger is only used when connected to an AC source.

You need not have any charging regime as the BMS will take care of the battery - that is what it's for.

 

and this is a reply that i got

Sorry to disagree. It is not the BMS that limits the charge rate. It is the nature of charging in that it is CCCV.

Constant Current Constant Voltage.

A battery can be charged at up to a specific current maximum. It can also only accept a charge voltage up to it's maximum charge voltage.

This explanation is not just aimed at you, but for anyone interested in what I mean by CCCV.

Lets use an imaginary battery so that the maths is easy.
The battery is flat at 2V and full at 4V and it is a 1AH battery.
The charge voltage is directly related to the current. So if you increase the charge voltage you increase the charge current proportionately.
Let's say the internal resistance of battery is 1Ω and the max charge rate of the battery is 1C So the maximum charge rate would be 1AMP.

If the battery is flat and showing 1V, we cannot charge at the full 4V maximum charge voltage because 4v-2v = 2v (difference) and 2v / 1Ω = 2Amp which is double the maximum charge current.
So you set the charge voltage at it's maximum which is 1Ω x 1A = 1V above battery voltage.

At the start when the battery is empty at 2V the charge voltage would start at 3V. As the battery charges it would increase to stay 1V ahead of the current battery voltage as it charges.
The charge voltage would stay 1V ahead of the battery voltage until it reaches 4V at which point it couldn't get any higher.
That is the constant current part of the charge cycle. The fast bit and usually around 80%.

We then move to the constant voltage part where the charge voltage is capped. The charge voltage cannot go any higher than 4V so as the batteries voltage continues to rise the difference between them falls and the charge current falls proportionally.

Here is a quick rough and ready table. The 1st two show constant current and the rest show the slow fall in charge current when it enters constant voltage mode.

The BMS will kick in when one cell reaches the maximum voltage. It will apply a resistance between the cathode and anode of that cell to drain it a little while the rest catch up. rinse and repeat.

 

and this is a reply that i got


Sorry to disagree. It is not the BMS that limits the charge rate. It is the nature of charging in that it is CCCV.

Constant Current Constant Voltage.

A battery can be charged at up to a specific current maximum. It can also only accept a charge voltage up to it's maximum charge voltage.

This explanation is not just aimed at you, but for anyone interested in what I mean by CCCV.

Lets use an imaginary battery so that the maths is easy.
The battery is flat at 2V and full at 4V and it is a 1AH battery.
The charge voltage is directly related to the current. So if you increase the charge voltage you increase the charge current proportionately.
Let's say the internal resistance of battery is 1Ω and the max charge rate of the battery is 1C So the maximum charge rate would be 1AMP.

If the battery is flat and showing 1V, we cannot charge at the full 4V maximum charge voltage because 4v-2v = 2v (difference) and 2v / 1Ω = 2Amp which is double the maximum charge current.
So you set the charge voltage at it's maximum which is 1Ω x 1A = 1V above battery voltage.

At the start when the battery is empty at 2V the charge voltage would start at 3V. As the battery charges it would increase to stay 1V ahead of the current battery voltage as it charges.
The charge voltage would stay 1V ahead of the battery voltage until it reaches 4V at which point it couldn't get any higher.
That is the constant current part of the charge cycle. The fast bit and usually around 80%.

We then move to the constant voltage part where the charge voltage is capped. The charge voltage cannot go any higher than 4V so as the batteries voltage continues to rise the difference between them falls and the charge current falls proportionally.

Here is a quick rough and ready table. The 1st two show constant current and the rest show the slow fall in charge current when it enters constant voltage mode.

The BMS will kick in when one cell reaches the maximum voltage. It will apply a resistance between the cathode and anode of that cell to drain it a little while the rest catch up. rinse and repeat.

Sorry that's rubbish.

 

The BMS DOES control DC charging, just as it controls AC charging. The only difference is that when AC charging everything is within the car, so the BMS just communicates to the OBC directly, whereas with DC charging the BMS communicates with the charger.

Simplistically, what happens is that as the highest voltage cell/cell group approaches shunt activation voltage (where charging current to that cell/cell group starts to be shunted past that cell/cell group) it signals to either the charger (for DC charging) or the OBC (for AC charging) to reduce the current. It does this because the shunts only have a limited power dissipation, so cannot operate at very high charge current without generating a fair bit of heat.

An out of balance pack will throttle back the charge current more quickly than a well-balanced pack. This is one good reason why an AC charge is a good idea every now and again, as the lower charge current allows the pack to balance towards the end of charging (that very slow stage over the last couple of percent), and a well-balanced pack can then take a rapid charge for longer before throttling back next time a rapid is used.

This issue can be resolved by using active balancing, something I've been doing for a while with my motorcycle battery pack. Not sure if many EVs yet do this, but it does have the advantage that the pack will try and balance all the time. The system I used has "flying capacitors", that transfer small amounts of charge from the highest voltage cells/cell groups to the lowest voltage ones. Over time this brings every cell/cell group to the same voltage, and tends to maintain that all the time. It's also less wasteful, as there's not as much charge power wasted as there is with a cell/cell/group shunt system.

 

EV batteries certainly do charge CCCV. You can see it just watching the voltage and currents on LEAFSpy or whatever.

Its an intrinsic characteristic of Lithium Ion batteries.

Not sure why everyone seems to think the car is "balancing" at the end of the charge as its tapering, its clearly not, its just holding a constant voltage while the cells saturate.

Balancing happens at a tiny current. 10-20ma perhaps, and it happens over a much longer time than one charging session.

 

[QUOTE="Aragorn, post: 3065156, member: 8522"

Balancing happens at a tiny current. 10-20ma perhaps, and it happens over a much longer time than one charging session.
[/QUOTE]

Even my old electric motorcycle cell groups shunted around 300mA when I was using resistive shunt balancing at the end of charge. My electric bicycle charger balances at around half an amp, and that's just for a 20 Ah pack. The pack in my electric boat can balance at up to 1.2 A per cell group.

 

Tesla definitely use switched shunt bleed resistors. They announced a change in the design of their packs to reduce the shunt balance current a couple of years ago, IIRC. I've probably built well over a dozen packs now, using various cell chemistries, and only found one chemistry that will sort of self-balance, as long as you don't overdo the charge/discharge current. That's a pack using LiFePO4 cells (same as in some MiC Model 3s) and even they can only tolerate a few dozen charge/discharge cycles before they drift out of whack, and that was with some really careful cell selection to try and ensure that every cell group was as near as dammit identical in capacity.

I've switched one pack over to active balancing now, using a flying capacitor charge distribution topology, and that is definitely a LOT better at keeping cell group voltages equal, and also allows faster charging to first cell group voltage limit slow down/shut off, so the pack charges faster. Not having current bleed resistors also makes it very slightly more efficient, although that has to be balanced against the longer time that the active flying capacitor circuit is operating.

 

Will someone please post a document from their vehicle manufacturer which states:

1. their definition of cell balancing
2. if balancing happens and is it passive or active,
3. when it happens
4. what the owner has to do to enable it to happen

Cell balancing in some form must happen automatically during a trickle charge at constant voltage for that last few percent, at least at the bank level, without BMS involvement.
I accept BMS involvement if balancing is necessary at the cell level but it must add a whle new level of complexity.

 

Enjoy:

http://media3.ev-tv.me/TeslaModuleController.pdf

(Is Google not working for you?)

 

As someone that's been building battery packs for years, here's a brief explanation of what happens in the real world, as measured by yours truly. When you start to charge a series connected pack of cells (or paralled cell groups, they behave just like single cells) then current flows through the whole pack as it charges. Lithium chemistry cells have a pretty flat voltage right up until they reach almost full capacity, when the terminal voltage rapidly rises. For LiFePO4 cells that voltage is typically around 3.6 V to 3.65 V or so, for most other Lithium ion chemistries that voltage is around 4.15 V to 4.2 V. If the limiting terminal voltage is exceeded, there is a risk of cell damage and overheating. The margin is tight, just a 100mV over the maximum allowable cell terminal voltage during a high current charge may cause a cell to seriously overheat and will likely cause cell damage.

The problem is this. Manufacturers try to match the capacity of cells in packs, so they don't vary much, but there is always some small variation. This means there will always be one cell (for cell read cell group from now on) that reaches cut-off voltage before all the others. A consequence of this is that this cell has to trigger the charge to slow down and stop, as if it did not then there's a risk of damage, even fire, as that cell voltage continues to increase. The snag is that when it does this the rest of the cells in the pack may well not be fully charged at all. You cannot allow current to flow through a cell that has reached cut off voltage, for fear of damage, therefore current cannot flow through it to charge the other cells. It is worth noting that the terminal voltage of every cell is measured by the BMS, primarily to allow it to detect the highest voltage cell during charging, but equally the lowest voltage cell during discharge (there's a similar lower voltage cut off that must be adhered to to keep cells safe and reliable)..

The original fix for this problem was to use cell shunt resistors that are individually switched across cells by the BMS when that cell reaches cut off. These resistors bypass the cell(s) that are fully charged, and allow current to flow to the remaining cells in the series stack. Obviously, the shunt resistors cannot pass a very high current, so charging slows down a lot as they activate. The BMS signals to the charger to reduce the current to a level that the pack can safely take at that time, using multiple sensors, including cell temperature, or can just use a crude CCCV method (this is far too crude for proper charge control, though). There are obviously flaws with using cell shunt balancing, it's not very efficient, for fast charging the shunts need to pass a high current (to reduce the time spent in the balance phase), the shunts will get hot as they dissipate power that would otherwise be heating up the cells, etc.

The method I now use (and I suspect that some EV manufacturers must be looking at using by now) is an active balancing BMS. This shuts off charging when one cell reaches it's set cut off voltage (and I shut mine off at 4.15 V, although the cells will take 4.2 V, makes the pack last longer), but, because the pack starts charging from a very well balanced state, it's pretty much balanced when the first cell shuts off the charge. The balancing works by continuously (in a cycle around the pack) switching a capacitor across each cell in turn, charging the capacitor from the highest voltage cell(s) and discharging it into the lowest voltage cell(s). After some time (off charge) the cells will all be at essentially exactly the same cell voltage, and hence SoC. This system is less wasteful of charge power (none gets wastes in switched cell shunt resistors), but it is slow (best done when the pack is idle) and also means that the pack doesn't quite ever get charged to a true 100%. This latter point can be a benefit, as packs tend to last a lot longer if rarely charged to 100%.

There are other ways of cell balancing. I had a long online row with the late Jack Rickard around 15 years ago, as he was insistent that "bottom balancing" was the only way to do it. It does work (Jack used it for some packs he built) but it's not any more efficient, and allowing cells to reach their lowest acceptable terminal voltage regularly is as bad as allowing them to reach their highest acceptable terminal voltage. IIRC, Jack did move away from doing this, although he never had the decency to admit he was mistaken.

 

As someone that's been building battery packs for years, here's a brief explanation of what happens in the real world, as measured by yours truly. When you start to charge a series connected pack of cells (or paralled cell groups, they behave just like single cells) then current flows through the whole pack as it charges. Lithium chemistry cells have a pretty flat voltage right up until they reach almost full capacity, when the terminal voltage rapidly rises. For LiFePO4 cells that voltage is typically around 3.6 V to 3.65 V or so, for most other Lithium ion chemistries that voltage is around 4.15 V to 4.2 V. If the limiting terminal voltage is exceeded, there is a risk of cell damage and overheating. The margin is tight, just a 100mV over the maximum allowable cell terminal voltage during a high current charge may cause a cell to seriously overheat and will likely cause cell damage.

The problem is this. Manufacturers try to match the capacity of cells in packs, so they don't vary much, but there is always some small variation. This means there will always be one cell (for cell read cell group from now on) that reaches cut-off voltage before all the others. A consequence of this is that this cell has to trigger the charge to slow down and stop, as if it did not then there's a risk of damage, even fire, as that cell voltage continues to increase. The snag is that when it does this the rest of the cells in the pack may well not be fully charged at all. You cannot allow current to flow through a cell that has reached cut off voltage, for fear of damage, therefore current cannot flow through it to charge the other cells. It is worth noting that the terminal voltage of every cell is measured by the BMS, primarily to allow it to detect the highest voltage cell during charging, but equally the lowest voltage cell during discharge (there's a similar lower voltage cut off that must be adhered to to keep cells safe and reliable)..

The original fix for this problem was to use cell shunt resistors that are individually switched across cells by the BMS when that cell reaches cut off. These resistors bypass the cell(s) that are fully charged, and allow current to flow to the remaining cells in the series stack. Obviously, the shunt resistors cannot pass a very high current, so charging slows down a lot as they activate. The BMS signals to the charger to reduce the current to a level that the pack can safely take at that time, using multiple sensors, including cell temperature, or can just use a crude CCCV method (this is far too crude for proper charge control, though). There are obviously flaws with using cell shunt balancing, it's not very efficient, for fast charging the shunts need to pass a high current (to reduce the time spent in the balance phase), the shunts will get hot as they dissipate power that would otherwise be heating up the cells, etc.

The method I now use (and I suspect that some EV manufacturers must be looking at using by now) is an active balancing BMS. This shuts off charging when one cell reaches it's set cut off voltage (and I shut mine off at 4.15 V, although the cells will take 4.2 V, makes the pack last longer), but, because the pack starts charging from a very well balanced state, it's pretty much balanced when the first cell shuts off the charge. The balancing works by switching a capacitor across each cell in turn, charging the capacitor from the highest voltage cell(s) and discharging it into the lowest voltage cell(s). After some time (off charge) the cells will all be at essentially exactly the same cell voltage, and hence SoC. This system is less wasteful of charge power (none gets wastes in switched cell shunt resistors), but it is slow (best done when the pack is idle) and also means that the pack doesn't quite ever get charged to a true 100%. This latter point can be a benefit, as packs tend to last a lot longer if rarely charged to 100%.

There are other ways of cell balancing. I had a long online row with the late Jack Rickard around 15 years ago, as he was insistent that "bottom balancing" was the only way to do it. It does work (Jack used it for some packs he built) but it's not any more efficient, and allowing cells to reach their lowest acceptable terminal voltage regularly is as bad as allowing them to reach their highest acceptable terminal voltage. IIRC, Jack did move away from doing this, although he never had the decency to admit he was mistaken.

Helpful, but with respect, we still don't know which strategy has been adopted by which manufacturers. (Exc Tesla)

Might it be the case by having a top buffer of 5%, means that the BMS may not need to use active balancing, just, as you say, a crude CCCV method?

 

Be useful to know if some EVs now use active balancing. I strongly suspect they do, because active balancing systems on a chip that will handle a dozen cells are now cheap, and that suggests they are in mass use, and that sort of cell number is way more than would be used in consumer devices like phones or laptops. Any BMS monitoring system, like Leaf Spy, should show this. If the cell voltages all periodically equalise off-charge then that's a very good indication that some sort of active cell balancing system is in use.

 

Here are a couple of screenshots taken just now. Car not on charge.
The dash shows 22% SOC so Leafspy's 32% must include the buffers.
The 96 cell voltages fluctuate continuously a small amount but I have no idea why.
At 100%, the battery voltage reaches about 406v

 

Big clue there, in that Leaf Spy refers to "shunts", so that strongly suggests the pack uses switched resistive shunt balancing, not active balancing. Also the cell voltages are too far apart for active balancing. Typically I find that active balancing gets all cells to within about 5mV of each other after an hour or so.

 

That makes sense. Never thought about the consequences of there being shunts but I can well believe Nissan using it's old tried and tested technology and conservative use of top and bottom buffers to minimise degradation.

 

All the EV BMS boards I've seen have a bunch of power resistors that look in 1-2 watt range, and mosfet/transistor switches for resistive balancing.
For example :

Volvo V60 battery pack - openinverter forum

Renault Zoe BMS - openinverter forum

The BMW i3 is similar

 

All the EV BMS boards I've seen have a bunch of power resistors that look in 1-2 watt range, and mosfet/transistor switches for resistive balancing.
For example :

Volvo V60 battery pack - openinverter forum

Renault Zoe BMS - openinverter forum

The BMW i3 is similar

Yep, my old cell balancers used a similar system. Here's a (very old, around 2004) BMS I designed, using individual cell shunts, with optocoupler signalling to the charge controller, just diode OR'ed. Instead of shunt resistors, I just mounted the TIP105 darlingtons on heatsinks and used them as variable current shunts. This was for my electric boat, so only needed to handle a few tens of amps maximum:

 

Ofcourse there are shunt resistors. but the power they can dissipate is very small. The BMS is not equipped with large heatsinking to dissipate hundreds of watts of power to actively shunt cells being charged at tens of amps... Its a few SMD resistors on the PCB to trim things at a very low current.

They're certainly not applying the shunts and running full charge current thru them to pull the rest of the cells up. The charging will stop when either the overall charge current drops to a predetermined value, or if any one cell hits whatever max voltage they specify. If one cell is somehow higher then all the rest, then the shunts will activate to pull it down, but that might take many hours or days. However because this is always going on, they never really get much out of sync unless theres a problem.

Similarly if theres one low cell, it'll stop charging with that cell still low, and switch on the balnacers on every other cell in the pack to try and pull them down to match.

Thus the actual charging process doesnt really involve balancing at all. The balancers are always active and working to keep the pack in check, and the charging is just charging the back upto a specified voltage (under the watchful eye of the BMS ofcourse) and then holding that voltage until the current drops to a specified value.