Voltage threshold for balancing

Topics concerning OEM and open source BMSes
Post Reply
User avatar
catphish
Posts: 954
Joined: Fri Oct 08, 2021 11:02 pm
Location: Dorset, UK
Has thanked: 93 times
Been thanked: 179 times

Voltage threshold for balancing

Post by catphish »

I am finishing work on a BMS controller for some lithium ion battery packs. Everything works correctly, but I have a fundamental question: when should balancing actually occur? My current algorithm will bleed any cell that meets BOTH of the following criteria:

1) Voltage > minimum cell voltage + 10mV
2) Voltage > 4.0v

Is 4.0V sensible, or should this voltage be closer to the full charge voltage (ie 4.1V or 4.14V)?
User avatar
joromy
Posts: 371
Joined: Fri Jun 28, 2019 12:56 pm
Has thanked: 1 time
Been thanked: 3 times

Re: Voltage threshold for balancing

Post by joromy »

I start top balancing 3.8V, but it depends on the quality of cells, the size of bleed resistors..
You should also consider if you want to charge them to 100% all the time, if you charge to 80-90%, you may have to start balancing at lower voltage.

If your batteries need much balancing, then they need longer bleed time or more watt bleeding.

I have batteries from Model S, and they need very little balancing.
Thomas A. Edison “I have not failed. I've just found 10,000 ways that won't work"
User avatar
catphish
Posts: 954
Joined: Fri Oct 08, 2021 11:02 pm
Location: Dorset, UK
Has thanked: 93 times
Been thanked: 179 times

Re: Voltage threshold for balancing

Post by catphish »

Thank you, I would prefer battery longevity over range, so for me, it definitely makes sense to charge only to around 80%. What voltage do you charge to? I just looked at some voltage/SoC curves, and it probably makes sense to charge to between 4.05V and 4.10V.

I suspect that after the first balance, the batteries will require very little balancing, and at lower states of charge, the voltage curve is very flat, so they're likely to appear balanced anyway. With that in mind, I now suspect that the exact voltage to begin balancing is not hugely important, as long as it's low enough that it is invoked when charging, but not so low as to accidentally bottom balance. I will likely stick with 4.0V for now, as I imagine most packs will be charged above this point at fairly regularly.
arber333
Posts: 3241
Joined: Mon Dec 24, 2018 1:37 pm
Location: Slovenia
Has thanked: 74 times
Been thanked: 223 times
Contact:

Re: Voltage threshold for balancing

Post by arber333 »

joromy wrote: Thu Dec 09, 2021 10:01 pm I start top balancing 3.8V, but it depends on the quality of cells, the size of bleed resistors..
You should also consider if you want to charge them to 100% all the time, if you charge to 80-90%, you may have to start balancing at lower voltage.

If your batteries need much balancing, then they need longer bleed time or more watt bleeding.

I have batteries from Model S, and they need very little balancing.
I think balancing before 3.95V is a waste of energy. From what i observed cell internal resistance goes up only after 4V and so your balancer cannot "grab" cells at lower voltages. When under load cell iR variation will cause different sensory inputs and your BMS might not see correct voltages.
Also note that i stop charging after calculated 4.08V per cell = 392Vdc.

Lately i started to use those Neey active balancers and i think they are just wonderfull. I have my cells balanced up to 0.005V delta which incidentaly i cant measure with my normal multimeter :).

I also think cells should never be discharged lower than 3.0V per cell. Its just my experience with cell longevity and already accounting for lower external temps... After 3.0V per cell cell ageing gets much more pronounced if you discharge them lower.
User avatar
joromy
Posts: 371
Joined: Fri Jun 28, 2019 12:56 pm
Has thanked: 1 time
Been thanked: 3 times

Re: Voltage threshold for balancing

Post by joromy »

arber333 wrote: Fri Dec 10, 2021 8:36 am
I think balancing before 3.95V is a waste of energy.

I also think cells should never be discharged lower than 3.0V per cell.
But what if I always charge to 80%, when should I then balance? (I usually charge to 4.00V, and they rest at about 3.95V)

So what you actually mean is I should charge to higher SOC and better avoid going low in SOC?
Thomas A. Edison “I have not failed. I've just found 10,000 ways that won't work"
arber333
Posts: 3241
Joined: Mon Dec 24, 2018 1:37 pm
Location: Slovenia
Has thanked: 74 times
Been thanked: 223 times
Contact:

Re: Voltage threshold for balancing

Post by arber333 »

joromy wrote: Fri Dec 10, 2021 8:53 am
arber333 wrote: Fri Dec 10, 2021 8:36 am
I think balancing before 3.95V is a waste of energy.

I also think cells should never be discharged lower than 3.0V per cell.
But what if I always charge to 80%, when should I then balance? (I usually charge to 4.00V, and they rest at about 3.95V)

So what you actually mean is I should charge to higher SOC and better avoid going low in SOC?
Good point! In that case i dont balance untill 4Vper cell. If cells went their separate way i would take some time and charge little higher say 4.1V per cell where i would rebalance them.
That would leave cells (if they are healthy) with lower SOC where they wouldnt differr much as long as you dont pull any of them too low. Then you would unbalance singular cells at "lower knee" SOC and so "higher knee" SOC will not be the same anymore. That would need more "higher knee" balancing.
I would certainly not charge them over 4.1V per cell. This would cause further "higher knee" imbalance every time when you try to balance them too high.

But like i said it seems active balancers have solved much of my "higher knee" worries now since they take from higher cells and give to the "poor". Very Robin Hood... 8-).
User avatar
johu
Site Admin
Posts: 5684
Joined: Thu Nov 08, 2018 10:52 pm
Location: Kassel/Germany
Has thanked: 153 times
Been thanked: 960 times
Contact:

Re: Voltage threshold for balancing

Post by johu »

Just a word on the upper charge limit: I learned this at a battery seminar. High voltages wear out the electrolyte more quickly, the longer the more. So charging to 4.2V is fine if you plan to discharge right away, but leaving them sitting for days at that voltage damages electrolyte. Also it means not upper SoC is key, but upper voltage.

It also means this problem barely exists with LFP because voltage hardly changes with SoC.

Balancing itself is a rather special topic. Ideally you know capacity (aka SoH) and SoC of each individual cell. Then at any point in the charge curve you can balance the cells to have equal left over capacity. Basically bottom balancing on the go. Or the other way around balance them to have equal capacity to full, i.e. top balancing on the go. The Nissan Leaf BMS *seems* to do the latter.

If you don't know SoH you can balance at some defined voltage, e.g. have all cells sit at 4.000V. Of course as you discharge they will come apart again because of the difference in capacity.
Support R/D and forum on Patreon: https://patreon.com/openinverter - Subscribe on odysee: https://odysee.com/@openinverter:9
arber333
Posts: 3241
Joined: Mon Dec 24, 2018 1:37 pm
Location: Slovenia
Has thanked: 74 times
Been thanked: 223 times
Contact:

Re: Voltage threshold for balancing

Post by arber333 »

johu wrote: Fri Dec 10, 2021 10:02 am If you don't know SoH you can balance at some defined voltage, e.g. have all cells sit at 4.000V. Of course as you discharge they will come apart again because of the difference in capacity.
Exactly. However in practice i see that if you dont discharge below certain SOC you will not need to balance on the top unless you charge over certain SOC. I set my limits between 20% and 90% and it seems to serve me well so far. Even without BMS intervention Kokam cells in Mazda have set themselves at varying voltages to delta 0.05 but they are there EVERY TIME. As long as i dont overcharge or underdischarge.

This applies as long as i keep to my limits, so this is my safety.
Attachments
brain1.jpg
User avatar
catphish
Posts: 954
Joined: Fri Oct 08, 2021 11:02 pm
Location: Dorset, UK
Has thanked: 93 times
Been thanked: 179 times

Re: Voltage threshold for balancing

Post by catphish »

To answer my own question here, I decided to use the following algorithm in my BMS:
* If any cell is > 4.0V AND 10mV higher than the lowest cell in the pack, discharge it.
* Discontinue charging when any cell is above 4.05V

This provides a nice window for top balancing, and keeps the overall pack charged to about 90%.
tom91
Posts: 1273
Joined: Fri Mar 01, 2019 9:15 pm
Location: Bristol
Has thanked: 97 times
Been thanked: 201 times

Re: Voltage threshold for balancing

Post by tom91 »

Depending on your chemistry you can tweak this, also got to be mindful of the fact you cannot have balancing fix an inherently imbalanced pack.

Did you map out how much capacity there is between lets say 3.9-4.05V compared to total capacity?
Founder Volt Influx https://www.voltinflux.com/
Post Reply