Strange behaviour of BMS cell voltages

[izeman]

I got one of those xiexiang Bluetooth BMS connected to a 16s solar battery (connected to a photovoltaic system)

And for one cell it's cell voltage monitor shows super strange values. Under load the voltage goes down massively, and while charging it goes up.
Would be totally fine, but measuring the voltage with a Fluke DMM directly at the cell terminals shows no voltage change at all (well slowly raising/declining of course during charge/discharge).

I'm just monitoring the highest and lowest cell in the screenshot as those trigger the HVC/LVC.

As you can see the lowest cell fluctuates minimal (as expected) and this one cell (which shows no fluctuation directly at the cell terminals) jumps from 3.2V and 3.7V. 3.7V would be super critical for LiFePo4 if it was real - but as said - it's not.

Any idea what could cause this strange behaviour? I don't have a second BMS here to test this out, but maybe I will need to get one.

And: This phenomenum got worse over the last days. Voltage was jumping, but not from 3.2V to 3.75V with a super low 10A charge. And immediately dropping to 3.2V as soon as the load goes away.

Internal resistance of all cells is also as expected between 1-2mOhm.







Here's a pic showing how it became from yesterday to today:

 

[Jrbe]

That trace looks like the result of a poor connection to me. Could be a poor crimp, broken wire, cold solder joint, etc. If you watch it while carefully wiggling things you might be able to narrow down its location.

 

[xfrankie]

---

As you can see the lowest cell fluctuates minimal (as expected) and this one cell (which shows no fluctuation directly at the cell terminals) jumps from 3.2V and 3.7V. 3.7V would be super critical for LiFePo4 if it was real - but as said - it's not.

Any idea what could cause this strange behaviour? I don't have a second BMS here to test this out, but maybe I will need to get one.

And: This phenomenum got worse over the last days. Voltage was jumping, but not from 3.2V to 3.75V with a super low 10A charge. And immediately dropping to 3.2V as soon as the load goes away.

---

I have seen this exact behavior in my own diy battery. The reason was "weak" contacts between battery terminals and BMS-voltage-sense-wires. In a way it acted as a high resistance connection...
With no current across the cell (positive or negative) the reading was fine. Fluctuating by a millivolt or two, but fine.
Any current above 15-ish Amps would send the voltage numbers flying over the safety cutoffs, disabling the battery. Which was fun. Especially fun when working regen is essential.... Sooo much fun...

Anyway, as soon as I reworked those contact points, this behavior disappeared. Only one cell is acting up, instead of like eight or so... In the last cell I suspect the wire itself is damaged/pinched. Which will be fixed "soon"...

So my advice would be to thoroughly check or replace the bms wiring harness, redo the contact points where it meets the battery, hard-wire or solder those instead of crimped connectors, etc...

In my first battery the sense-wires are soldered on and it never had any problems. Second battery (roughly double the size) is made with crimped ring-terminals, screwed to the battery electrodes, and that one is the problem-child...

 

[izeman]

That trace looks like the result of a poor connection to me. Could be a poor crimp, broken wire, cold solder joint, etc. If you watch it while carefully wiggling things you might be able to narrow down its location.

Didn't think of that. But sure. Could be, still the balance wires measure voltage at the terminals and see absolutely no voltage drop if current goes through the power wires.

I attached a picture of the setup.

 

[izeman]

In my first battery the sense-wires are soldered on and it never had any problems. Second battery (roughly double the size) is made with crimped ring-terminals, screwed to the battery electrodes, and that one is the problem-child...

Thanks. Will double check. Ordered a new BMS anyway, as the one I use is rated for 40A only, and I pull 50A for a longer period, and it has 0.6A balance capability.
Will be able to tell in a few hours i some connection is loose.