Proper process for balancing two Dyness batteries

[leon.davibe]

Hey everyone

We have two Dyness 5.12kWh DL5.0C batteries in our solar system. One was installed back in March 2024 while the other was added roughly a year later.

There are so many conflicting views about how to properly balance batteries but I think whatever our installer did was incorrect.

Although many online claim that the BMS on the two batteries will sort out the balancing automatically, others say they MUST be connected initially at the same level of charge.

I am now often finding that the newer battery has one fewer battery level indicator light on than the older battery.

For example, when the total SoC is 29%, the older battery is still at 2 lights while the new one drops down to 1 light.

This has me especially worried that I am dropping below the minimum rated SoC on the new battery when discharging to as low as 20%.

Some have recommended that I should charge the batteries to 100%, one at a time. Would it be possible to do this just by switching off one of the batteries with its toggle and changing the necessary amp ratings on the Deye inverter and then moving over to the next one? Or are more steps necessary?

If you charged full and keep the voltage there untill both is 100% full they are balanced

ie change float to match bulk
Once both ceases to cgarge , can change float to normal setting

How is the batteries connected
If the cables aren't equal lentghs

the one with shortest distance will supply more of the power and be slightly ahead of the other both charging and discharging


25%soc will be 2 lights
24% will be 1 on 4 light setup

So using the lights is folly

So after both is full having both reset to 100% soc , you can check behaviour

Need better data , soc % from bms
But even then they drift away from reality even in just one day

 

[RedViking]

2 Spoons of Apple Cider Vinegar on each battery will balance them properly.

 

[leon.davibe]

The right way is to charge the batteries to full one at a time and then connect them together. You also should have them on the same firmware ideally but if not have the newer firmware as the primary.

As things stand now switch your inverter to keep the batteries full charging off the grid for 2 days and then switch back to your regular settings.

That being said its been cloudy for a while are they charging up to 100% and staying there for a few hours to equalise every couple of days?
Which is the problematic battery? Old or new?

Doesn't actually need days to equalise

The battery that is full first
Will go into protection mode

The other will then accept all the charge once it hits protection levels

It will be close enough to be regarded balanced

It will be closer than it will be in the very next discharge cycle imo

 

[leon.davibe]

If it was LA batteries I could tell you for sure. Resting voltage matters when determining SOC. lithium I don’t know. Do they report 100 when connected together or disconnected? And does lithium need to be left standing for a couple of hours to verify the SOC like LA? I haven’t researched my on how BMSs report their SoC.

If one says 40 and the other 60 when discharging then they most definitely couldn’t be at the same SOC to begin with. Or there is a problem with one being discharged quicker than the other. Or one has a faulty cell and is working as a 15 cell battery?

Will powess had all his batteries connected at the same SOC then he started charging and discharging them. And because the cabling wasn’t done in a way to properly cajrge and discharge them equally, some batteries drifted during the one charge/ discharge cycle quite a bit. One battery was pulling for example 20A and another 10 amps.

Doesn’t matter if they are in parallel, they will never balance and share the load

These bms's. Tend to have a blind spot under a amp

So if the load vs solar means batteries are essentially slow chargjng especially if one is slow charge/discharge then the SOC gets out of wack quicker

I have a small system so try and use as much solar as possible the soc can drift 5% in a day in this specific scenario

So if combined i'm chargjng/discgargjng 2A one battery maybe 1.2A and the other 0.8A then the bms doesn't register this 0.8A and the soc indicator stays stagnant even though the soc is changing the same happens when it gets to the top

If one is pulling more amps cause shorter wires and the other slow charges it drifts again

Yes once they hit protectikn voltsge 3 65v cell in my case the counter resets and both are 100%

Even though the soc indicstors drift they don't actually drift that far apart just what the bms thinks drifts

You can use how long on stops charging before the other to have an idea how far they actually drifted apart

ie check charge rate and the time it takes till both are reported full from 1 inficated full

So if charge rate is 500w then if it takes 1hr for 2nd battery to also report full then one battery drifted 10% apart as thst is 0.5kwh

If it tskes 30 min then it is o.25kwh
So 5%

If the 2 are reported full within a few mins then the drift is nowhere near what bms's report

Naturally charge rate matters 5 min of 3Kw chargjng is 5% of 5kwh battery
While 5 min at 300w is 0.5%

I have found that my bms's report 100% soc and can sometimes charge more than 5% capacity after this point

So i take soc from bms with a bag of salt

If you don't spend as much timebin bms deadzone the drift soc vs reality may be closer

 

[leon.davibe]

If it was LA batteries I could tell you for sure. Resting voltage matters when determining SOC. lithium I don’t know. Do they report 100 when connected together or disconnected? And does lithium need to be left standing for a couple of hours to verify the SOC like LA? I haven’t researched my on how BMSs report their SoC.

If one says 40 and the other 60 when discharging then they most definitely couldn’t be at the same SOC to begin with. Or there is a problem with one being discharged quicker than the other. Or one has a faulty cell and is working as a 15 cell battery?

Will powess had all his batteries connected at the same SOC then he started charging and discharging them. And because the cabling wasn’t done in a way to properly cajrge and discharge them equally, some batteries drifted during the one charge/ discharge cycle quite a bit. One battery was pulling for example 20A and another 10 amps.

Doesn’t matter if they are in parallel, they will never balance and share the load

Yes lithium the bms regards the lifepo4 cells full once it acell hits 3.65v

Amd then it coulomb counts the current leving the battery

And does the math of where the soc is at at any given point

Problems can't count any losses in the battery and many of the bms's aren't accurate at very low amps or the software ignores low currents

I know mine does not count anything under 1 amp in or out

So f you spend lots of time there
Trying to optimise your solar setup

You can drift quite a bit
I have calculated that i can easily drift 5%+ in a single day


Because the voltage curve of lifep04 is that flat

Sure when they go to the extremes where the voltage curve does the steep moves they will shift the load drasticly from battery to battery

But in the flat section the voltage difference isn't much so the one with the shortest wires will do most of the work

 

[wingnut771]

Yes lithium the bms regards the lifepo4 cells full once it acell hits 3.65v

Amd then it coulomb counts the current leving the battery

And does the math of where the soc is at at any given point

Problems can't count any losses in the battery and many of the bms's aren't accurate at very low amps or the software ignores low currents

I know mine does not count anything under 1 amp in or out

So f you spend lots of time there
Trying to optimise your solar setup

You can drift quite a bit
I have calculated that i can easily drift 5%+ in a single day


Because the voltage curve of lifep04 is that flat

Sure when they go to the extremes where the voltage curve does the steep moves they will shift the load drasticly from battery to battery

But in the flat section the voltage difference isn't much so the one with the shortest wires will do most of the work

You can't say this for sure.

 

[leon.davibe]

You can't say this for sure.

For a while i was obsessed with monitoring the behaviour

Many wrong conclusions was initially made

And some may persist who knows

Everyday we learn something new

The way ro know the battery is full, it has to hit 3.65v on cell

So the bms will say it is back to 100% due to coulomb count

Then i would check the time and the charge rate, and calculate the kwh charged after it reported as full

And have often had charge sessions of o.5kwh +

The bms count problem under 1A also consistant

Tried to get a battery to balance the cells via bms

It only balances while charging (which is dumb imo)

So i tried to extend yhe charge time and thus extend the balance time

Problem is my inverter only has eskom charge options 2/10 or.more in 10A increments

So selecting 10A i felt was still to quick to balance as it only balances when over 3.45v and until it hits 3.65v

And the 2A would be divided between 2 batteries and thanks to losses each actually gets less than 1A the bms regards as under 1A as not chargjng and thus its logic for balancing does kot work

The battery still charges full the only thing is the soc counter doesn't change as charge is seen as 1A or kore or nothing

Once it hits 3.65v that if else logic kicks in and the soc counter jumps to 100%

Eventually i realised this flawed if else approach is never gonna get the cells balanced , i ripped out a halogen bulb opened the battery and did in 2 days what the bms could not do over 6 months of tryjng (doing all i can to work around the if else flaws on the shoto bms)

And seems many batteries uses this pace type bms

The actuve balancer in the jkbms really made a big difference for me
Since it will balance wether it is charging or not as long as any cell is over the voltage set for balancing

Imo any passive balance BMS that only balances while charging means nothing and you are essentially at the mercy of cell selection and guarantee

If the cell selection was kak you will test the guarantee if the drift happens quick enough if not you are stuck with the problem

 

[B-1]

You can't say this for sure.

Thread has deteriorated into an mybb expert post dump.
The usual "cables must be the same length" when dyness comes with premade cables at specific length.
The batteries equalise nonsense shows there is clearly no understanding how bms's work and need to charge/discharge individual cells over a long period to balance internally before the eternal balance matters.
Generic voltages assumed.

 

[wingnut771]

For a while i was obsessed with monitoring the behaviour

Many wrong conclusions was initially made

And some may persist who knows

Everyday we learn something new

The way ro know the battery is full, it has to hit 3.65v on cell

So the bms will say it is back to 100% due to coulomb count

Then i would check the time and the charge rate, and calculate the kwh charged after it reported as full

And have often had charge sessions of o.5kwh +

The bms count problem under 1A also consistant

Tried to get a battery to balance the cells via bms

It only balances while charging (which is dumb imo)

So i tried to extend yhe charge time and thus extend the balance time

Problem is my inverter only has eskom charge options 2/10 or.more in 10A increments

So selecting 10A i felt was still to quick to balance as it only balances when over 3.45v and until it hits 3.65v

And the 2A would be divided between 2 batteries and thanks to losses each actually gets less than 1A the bms regards as under 1A as not chargjng and thus its logic for balancing does kot work

The battery still charges full the only thing is the soc counter doesn't change as charge is seen as 1A or kore or nothing

Once it hits 3.65v that if else logic kicks in and the soc counter jumps to 100%

Eventually i realised this flawed if else approach is never gonna get the cells balanced , i ripped out a halogen bulb opened the battery and did in 2 days what the bms could not do over 6 months of tryjng (doing all i can to work around the if else flaws on the shoto bms)

And seems many batteries uses this pace type bms

The actuve balancer in the jkbms really made a big difference for me
Since it will balance wether it is charging or not as long as any cell is over the voltage set for balancing

Imo any passive balance BMS that only balances while charging means nothing and you are essentially at the mercy of cell selection and guarantee

If the cell selection was kak you will test the guarantee if the drift happens quick enough if not you are stuck with the problem

We are talking about dyness batteries here.

 

[leon.davibe]

Thread has deteriorated into an mybb expert post dump.
The usual "cables must be the same length" when dyness comes with premade cables at specific length.
The batteries equalise nonsense shows there is clearly no understanding how bms's work and need to charge/discharge individual cells over a long period to balance internally before the eternal balance matters.
Generic voltages assumed.

Batteries has to arrive internally balanced if not you are fuxked

No internal balancing has to happen on a new battery

So for the end user only corect setup matters for external balance


Only maintenance to prevent internal drift

I am of the opinion if cells are out of balance no amount , of trying will get them balanced again

They only get out of balance in two ways

The cell selection exceeds the bms balance capabilities

So if you weren't stupid enough to fall for the 20-80% bullshit and thus broke your own battery

And an imbalance presents itself
The bms cannot fix it , only warrantee option left

And even then it may not be truly fixed

If they take the shortcut and merely rebalance ,the problem will return later again

 

[wingnut771]

Charge it to 57.6V, let it float overnight at 57.6V then reduce float back to 53.6V.

If the leds are still out of wack then that's just the way it is.

 

[leon.davibe]

We are talking about dyness batteries here.

Just about all batterie have a passive balancer and a bms to prevent it from going out of range

So all batteries in parralel will likely behave the same

There is no smart tech that selects how much power comes from a certain battery (bar the overcurrent protection in the event of parralel connection of uneven batteries )

The logic of the bms is actually quite symplistic

Once it is in the game either chargjng or discharging

How much power each takes and gives is down to cable lentghs and battery resistance

No amountnof settings fiddling can affect it

 

[wingnut771]

Just about all batterie have a passive balancer and a bms to prevent it from going out of range

So all batteries in parralel will likely behave the same

There is no smart tech that selects how much power comes from a certain battery (bar the overcurrent protection in the event of parralel connection of uneven batteries )

The logic of the bms is actually quite symplistic

Once it is in the game either chargjng or discharging

How much power each takes and gives is down to cable lentghs and battery resistance

No amountnof settings fiddling can affect it

 

[leon.davibe]

View attachment 1861940
Charge it to 57.6V, let it float overnight at 57.6V then reduce float back to 53.6V.

If the leds are still out of wack then that's just the way it is.

Yes then you have to check cable lentghs to see if the one isn't being over worked cause it's closer to the inverter

If that is equal then it is battey resistance and nothing can be done about it

Unless you introduce a bit of extra resistance on the battery being overworked ie longer cable causing voltage drop and thus getting them to load balance better

But yea have not played around with this idea

It may be that the voltage drop on the flat section simply isn't enough to get it balanced draw unles you have silly amount of cable on the one

 

[wingnut771]

Yes then you have to check cable lentghs to see if the one isn't being over worked cause it's closer to the inverter

If that is equal then it is battey resistance and nothing can be done about it

Unless you introduce a bit of extra resistance on the battery being overworked ie longer cable causing voltage drop and thus getting them to load balance better

But yea have not played around with this idea

It may be that the voltage drop on the flat section simply isn't enough to get it balanced draw unles you have silly amount of cable on the one

I wouldn't bother.

 

[denmendez65]

Yes then you have to check cable lentghs to see if the one isn't being over worked cause it's closer to the inverter

If that is equal then it is battey resistance and nothing can be done about it

Unless you introduce a bit of extra resistance on the battery being overworked ie longer cable causing voltage drop and thus getting them to load balance better

But yea have not played around with this idea

It may be that the voltage drop on the flat section simply isn't enough to get it balanced draw unles you have silly amount of cable on the one

No just cable lengths. B1 mentioned that dyness already supplies the batteries with equal length cables.

Thats not the only thing that matters, also where the inverter cables connect to the battery bank will determine which one gets discharged harder than the rest.

 

[leon.davibe]

View attachment 1861942View attachment 1861943

Yea 1 light difference can be both sides of a threshold 2 not so much

I assume they both charge full regularly

Then i would measure cable maybe with a rope trace the canlenon wach battery and compare

If the same ,i woild try and do a capacity test on one batyery at a time, ie switch one battery off do capacity test may have yk change inverter to gell profile if you don't want to change dol switches

Do try and do the switch off not at bulk voltage anything under float voltage should be fine 20-99% soc
To prevent current rush from one to the other

Also try and see if you can see individual cell voltages to be sure both batteries are in good balanced state internally (but do check this when it is at bulk voltage and 100% soc)

Trying to check this at any other point n the voltage curve will mean nothing

 

[wingnut771]

No just cable lengths. B1 mentioned that dyness already supplies the batteries with equal length cables.

Thats not the only thing that matters, also where the inverter cables connect to the battery bank will determine which one gets discharged harder than the rest.

 

[leon.davibe]

No just cable lengths. B1 mentioned that dyness already supplies the batteries with equal length cables.

Thats not the only thing that matters, also where the inverter cables connect to the battery bank will determine which one gets discharged harder than the rest.

Yea imagine cable comes from inverter to first battery same lentgh cable goes from one battery to the next ?

That is double the distance

Engineers can try and solve the problem , end users will find a way

 

[wingnut771]

...

Also try and see if you can see individual cell voltages to be sure both batteries are in good balanced state internally (but do check this when it is at bulk voltage and 100% soc)

Trying to check this at any other point n the voltage curve will mean nothing

How do you do that?