JKBMS Random High Cell Voltages

[nate_syd]

Here's the setup...
2 x 16S LiFePO4 bank
All balancing cables come to a junction block
From that block I have both the JKBMS, & NEEY active balancer

The voltages in the NEEY are spot on, consistent and correct (as per multimeter)
The voltage readings on the JKBMS are jumping around a LOT, by 100mv

See the attached pic, look at 15 & 16. 3 screen shots taken from the BMS app, at most a second apart.
The ~3.54 readings are correct in the last shot, the first 2 shots showing wild readings.

It cant be the bus bars to the terminal blocks, as the NEEY is perfect.
The cables for the NEEY & JKBMS are wrapped together - so it cant be the other side of the terminal block

Is the NEEY balancer interfering with the JKBMS?
The low JKBMS readings i'm thinking it is causing issues, as the NEEY pulling current from the BMS more than the battery.
But why the high readings???

 

[sunshine_eggo]

Is the NEEY balancer interfering with the JKBMS?
The low JKBMS readings i'm thinking it is causing issues, as the NEEY pulling current from the BMS more than the battery.
But why the high readings???

Yep.

It's been documented several times on this site that you can't use the same leads for sensing and a separate balancer.

The NEEY is calculating the voltage drop and applying a correction factor, so it's reporting a corrected voltage, but the JK actually only sees the voltage drop across the sense leads and has no clue current is flowing.

The NEEY is an active balancer, pulling from one cell and "pushing" to another, so you will get low on the cell being drained and high on the cell being charged.

Reduce the current in the NEEY balancer, and you should see the deviations scale proportionally.

 

[zedconnor]

Here's the setup...
2 x 16S LiFePO4 bank
All balancing cables come to a junction block
From that block I have both the JKBMS, & NEEY active balancer

The voltages in the NEEY are spot on, consistent and correct (as per multimeter)
The voltage readings on the JKBMS are jumping around a LOT, by 100mv

See the attached pic, look at 15 & 16. 3 screen shots taken from the BMS app, at most a second apart.
The ~3.54 readings are correct in the last shot, the first 2 shots showing wild readings.

It cant be the bus bars to the terminal blocks, as the NEEY is perfect.
The cables for the NEEY & JKBMS are wrapped together - so it cant be the other side of the terminal block

Is the NEEY balancer interfering with the JKBMS?
The low JKBMS readings i'm thinking it is causing issues, as the NEEY pulling current from the BMS more than the battery.
But why the high readings???

Wow my case is the exact opposite of yours.

In my case the Balancer's voltages are jumping around when it's balancing.
And my setup is similar to yours as well.
Single cable going out to batteries on one side of terminal block. Both bms and balancer cables go into the block from the other side.

 

[zedconnor]

MIne is DALY bms and balancer like in this pic.
Notice that they do use a Y cable. And even sell these cables.

 

[nate_syd]

The NEEY is an active balancer, pulling from one cell and "pushing" to another, so you will get low on the cell being drained and high on the cell being charged.

Agree with with low voltages... but...
I'm seeing HIGH voltages on Cells that its pulling from... the high voltage cells are too high...
the only thing the NEEY is doing is pulling from those - so why are they above the normal voltages.

Is the correction factor causing an increase in voltage?

 

[nate_syd]

Wow my case is the exact opposite of yours.
In my case the Balancer's voltages are jumping around when it's balancing.

No - that's what i'm seeing too, whacky when balancing!

Maybe we put a heap of capacitors across all the cells to smooth it out lol

 

[zedconnor]

No - that's what i'm seeing too, whacky when balancing!

Maybe we put a heap of capacitors across all the cells to smooth it out lol

I am seeing High voltages too when the balancing is ON. Like 100mV higher that it actually is.
But during this, the voltage readings on the BMS is absolutely normal.(via bluetooth or the lcd or laptop)

For example highest cell shows 3.5V on both Multimeter, BMS display, the cell itself.
But the balancer will display it at 3.6v+
It is as if the balancer virtually added 100mV out of thin air to the cell. Maybe it has something to do with it's balancing algo?
Maybe it is lying to itself that this cell have very high voltages? Idk.

 

[sunshine_eggo]

Again:

It's been documented several times on this site that you can't use the same leads for sensing and a separate balancer.

Each device should have its own leads. Your exact scenarios have been documented multiple times on this site. All were fixed by: 1) reducing active balance current to yield acceptable voltage discrepancies or 2) stop sharing leads.

 

[nate_syd]

Again:
Each device should have its own leads. Your exact scenarios have been documented multiple times on this site. All were fixed by: 1) reducing active balance current to yield acceptable voltage discrepancies or 2) stop sharing leads.

I appreciate your feedback.
I am after the WHY - a deeper understanding as to exactly why it is happening, to work on a feasible solution.

Reducing the balance current is a really, really bad solution. You install a balancer for a reason, reducing its ability to be a balancer goes directly against the fact that you put one in place.

 

[Makeurown]

Think of a two cells and a 2 port active balancer:

When the Active balancer if off:

There are two long meter leads running from the A-Balancer to the Cells and you can correctly measure the cell voltages.

When Active Balance turns on, you effectively connect a small power supply across the meters inside the active balancer.

The Balancer readings now show the output of that power supply - you are no longer reading just the cell voltage at the end of the long leads.

When you read the voltage across the cells with a separate meter, you are not reading the cells at the same point and under the same conditions as the balancer.

The lead length, the balance current and voltage across and through the leads, and the fact that some balancers are pulsing the balance voltage, so it may not be a DC stable voltage, therefore your meters will not read it correctly, all or some of these conditions will cause the readings to jump around and be out of sync with the BMS and terminal readings, and that's provided you have used separate leads for the BMS and the Active Balancer.

 

[nate_syd]

When Active Balance turns on, you effectively connect a small power supply across the meters inside the active balancer.
The Balancer readings now show the output of that power supply - you are no longer reading just the cell voltage at the end of the long leads.

The lead length, the balance current and voltage across and through the leads, and the fact that some balancers are pulsing the balance voltage, so it may not be a DC stable voltage, therefore your meters will not read it correctly, all or some of these conditions will cause the readings to jump around and be out of sync with the BMS and terminal readings, and that's provided you have used separate leads for the BMS and the Active Balancer.

100% get that - & when you're charging a cell through a balancer - i'll read high, perfect.
What i'm seeing is that when cells are being DRAINED by the balancer, the voltage is reading HIGH on that cell (or its neighbour?).

Have a look at the middle balancing picture...
15 is low...
BUT! 14 & 16 are HIGH

So when a balancer is draining a cell, on a parallel connected BMS - the drained cell will read low, AND its 2 neighbours will read high.
I'm guessing its pronounced at high cell voltages when its balance time, as the battery charging current gets close to the balance current. So when a BMS does some factoring of voltages due to current - the "significant" increase in current due to balancing at those cells, throws a wobbly.

Given the JK BMS has 2A balance current, i'm guessing it has low input impedance... and the voltage sensing circuit is not buffered and taken off that low impedance path?!?

 

[RCinFLA]

All BMS's momentarily stop any cell balancing current when they make cell voltage sample measurements to avoid balancing current induced voltage drop on small gauge sense wires. Even 50 mA resistor dump BMS balancers do this.

JK active balancer BMS also does this. The wire resistance test is just to show up poor connections which are more likely when dealing with higher balancing current of the active balancer. When dealing with 2-4 amps of peak balancing current it is very easy to have a problem with small lug crimps.

When you connect multiple balancers they are operating asynchronously from each other without knowledge of what the other balancer is doing.

Creates a 'rats nest' but you need each balancer to have its own sense wires to cells.

 

[sunshine_eggo]

I appreciate your feedback.
I am after the WHY - a deeper understanding as to exactly why it is happening, to work on a feasible solution.

The NEEY is calculating the voltage drop and applying a correction factor, so it's reporting a corrected voltage, but the JK actually only sees the voltage drop across the sense leads and has no clue current is flowing.

You are pushing "high" amps for the wire gauge, so there is a significant voltage drop. Remember, high voltage + low current equals less voltage drop. You're pushing comparatively high current at very low voltage through thin wires.

Reducing the balance current is a really, really bad solution. You install a balancer for a reason, reducing its ability to be a balancer goes directly against the fact that you put one in place.

Agreed. That's why the balancer needs its own leads rather than using the BMS's SENSE leads.

 

[RCinFLA]

The JK active balancer BMS has an active buck switcher to charge the 2.5v supercaps from the highest voltage cell. It then has a boost switcher which has a constant current output to push the supercap's 2.5vdc up to the lowest voltage cell.

The boost switcher is limited on what the highest voltage output can be and still achieve its constant current sourcing capability. If you have too much sense wire path resistance, the boost DC to DC switcher will not produce its current limited output level.

The buck and boost DC to DC switchers are mux'ed through power MOSFET's with one 'from' cell at a time to one 'to' cell at a time. All based on what the highest and lowest cell voltages that were measured during all cell voltage measurement sampling.

 

[nate_syd]

OK - think I figured out whats happening...
Here's my diagram as to why HIGH voltage cells are causing neighbouring cells to read HIGHER.

Of course its most likely that its the balance cables and the BMS (NOT the balancer) Internals that are getting pulled high & low, due to the impedance of the BMS circuits being lower than the battery cells, but hopefully this helps to visualise the issue thats occurring.

 

[nate_syd]

So whats a solution...
As the BMS is no longer balancing, you could increase the impedance of all the wires by inserting resistors to EVERY wire. This would force the Balancer to draw from the cells & not the BMS internals.
Add some capacitors in parallel with the cells, hopefully they would be lower or the same resistance as the BMS, holding the voltage more stable during balance operation.

I'll test out the latter today!

 

[zedconnor]

You are pushing "high" amps for the wire gauge, so there is a significant voltage drop. Remember, high voltage + low current equals less voltage drop. You're pushing comparatively high current at very low voltage through thin wires.




Agreed. That's why the balancer needs its own leads rather than using the BMS's SENSE leads.

Please let me drop my question here as well. As I am having similar problems.
So you're basically implying that this advertisement is a very bad example of how to connect bms and active balancer to battery?



(The balancer in the pic is a dedicated one. It has no comms with the bms.)

 

[upnorthandpersonal]

So whats a solution...

Use the separate leads, Luke.

 

[sunshine_eggo]

Please let me drop my question here as well. As I am having similar problems.
So you're basically implying that this advertisement is a very bad example of how to connect bms and active balancer to battery?

(The balancer in the pic is a dedicated one. It has no comms with the bms.)

Yes.

If the harness is heavier gauge and can handle the balancer current without problematic voltage changes, it could work.