Tesla Model 3 BMS voltage issue - advice needed!

[nkiernan]

Hope this is the correct place to post this. I’m hoping someone with a better understanding of BMS boards can help identify an issue I’m seeing.
I will be using an Orion 2 BMS with Tesla Model 3 battery modules (4 modules, 24 cells per module). To connect to the BMS slave boards on each module, flying leads have been soldered to each tap pad to run to a connector.

Before any soldering, each of the cell tap voltages were recorded, and these were checked again afterwards before anything was connected. Cell 47 is flagging up an issue and I can’t see why. I’ve checked for shorts, or any visible issues with the solder on that pad and the two neighbouring pads.
I’m wondering if the symptoms suggest something with the BMS board, eg: a component got damaged with the solder heat, or maybe soldering iron tip shorted two pads during soldering. Someone more experienced in electronics might have some ideas on what to check? (really hoping this is not a costly mistake that can’t be fixed!!!).

The table below shows the measured tap voltages at the BMS slave board pads before and after the flying leads were soldered on. Note the new cell 47 voltage now shows lower than the cell 46 voltage, how could that happen? And then everything looks to correct itself from cell 48 onwards:

So I checked a little more around cell 47. 44/45 and 45/46 are spot on, but 46/47 and 47/48 appear to show very low cell voltages (which would suggest discharged/damaged cells I would think, except that from cell 48 onwards everything corrects and the overall module and battery pack voltages have not lowered in relation to these measurements)

TapVoltage2.GIF (3.96 KiB) Viewed 4474 times

So then I checked the voltages across a number of cells around cell 47, and these all appear to read correctly as if cell 47 is at the correct voltage (approx. 3.95V). This gives me hope that the overall battery module/pack is not seeing an issue, and something is just affecting my measurements on this one cells tap pad (so hopefully there are a couple of things I can check easily?)
Notice the variance in cell 47 voltages in the first table is 4.6V (in a negative swing) which matches the sum of the cell 46/47 and 47/48 voltages


Some photos of the flying leads and pad soldering:

[Pete9008]

Not quite clear how you measured those voltages, are they all with a multimeter or are some of them measured by the BMS (possibly silly question as I know nothing about the Orion system).

If with a multimeter do you get the odd measurements with just the flying leads connected or only when the flying leads are plugged into the BMS?

[nkiernan]

Hi Pete,
So these are measuered with a Fluke multimeter, and checked at both the flying lead end in the open plug and at the aluminium pad on the battery module and on the copper pad on the BMS slave board...just to be sure. These checks are before the flying lead plug has been connected to anything (to protect the Orion BMS before first powerup). The four modules were just connected together to give full pack voltage, multimeter negative lead connected to negative battery terminal and positive lead to each tap

So there should be very little that changed between doing the checks before and after other than soldering on the short flying leads. Hope that makes sense

[Pete9008]

In that case I think the cell 47 tap might be open circuit (I'm guessing the voltage you are reading is due the old BMS board providing a high impedance measurement return path)

Do any of the above pictures show the cell 47 tap? If so which one?

Edit - I'm not familiar with the Tesla modules and have assumed that the PCB shown is part of the original Tesla BMS and that the flexi circuit board goes off to the batteries. If that's not correct then neither is the above comment on the cell 47 tap.

[nkiernan]

Yes, the PCB on the module is part of the Tesla BMS system (so not used by the Orion BMS, just easier to solder to the copper pads on this board than the aluminium pads on the flexi, and also hoping at some point these can be reverse engineered and allow me to remove the Orion 2 BMS).

The orange flexi with aluminium pads goes to the battery cells with the two small alu wires per pad connecting this to the BMS slave PCB. So I thought if the PCB with copper pads got damaged during the soldering, or the alu wires to the flexi got damged, measuring the cell voltage at the flexi alu pads would show the correct voltage anyway? But this shows the same weird voltage as at the copper PCB pad.

Or if the small alu joinging wires to the pads are actually lifted off one of the pads, could that give the symotoms above?

Edit: This is tap for cell 47

[Pete9008]

Ok, that makes sense.

I'm hoping I'm wrong here but my suspicion is that while soldering two taps were briefly shorted. This would short circuit a cell and be likely to melt either a bond wire or pcb trace. The bond wires at the bms end look ok so it might be worth visually inspecting as much of the flexi pcb trace as you can?

Edit - if both ally bonding wires were lifted it could give that effect when measuring on the copper bms pads but measuring on the flexi pads should give the correct cell measurement. If the voltages are wrong on the flexi then it suggests that a flexi trace, or the far end bond wire, is the issue.

[nkiernan]

Thanks for your help Pete,
That is a plausaible scenario during the soldering, so will do some checks along the flexi and underside taps. Fingers crossed

[nkiernan]

Ok, so looks like you were spot on Pete. And hopefully I've had a lucky escape because the trace in question happens to be one of the shortest and closest to the PCB. Voltages direct to the cells are spot on but no continuity on that trace from cell to PCB tap.


Photos below show the suspected failure point! Thank you for your help, hopefully an easy fix

[Pete9008]

Glad you found it and that it's accessible for the repair

What's your plan for fixing it?

[nkiernan]

My plan is not to interfere with the small traces on the flexi too much, so easiest route I see is clean the point of the failure, then add a small guage wire soldered between the alu flexi pad near the PCB and the cell alu flexi connector and tape it down across the flexi

[Pete9008]

Makes sense but how are you going to solder to aluminium??

[nkiernan]

I'll be relying on the guy who helped solder the BMS boards, he knows his stuff and was happy to work with the alu originally so he'll be drafted in again

[Pete9008]

Didn't even know that soldering to aluminium was possible!

Any chance you could ask what kind of solder/flux he uses? The ability to solder to aluminium could be very, very usefull!

[nkiernan]

Yes, will see what I can find out

[Bratitude]

Didn't even know that soldering to aluminium was possible!

Any chance you could ask what kind of solder/flux he uses? The ability to solder to aluminium could be very, very usefull!

The problem with aluminum is that it oxidizes in the presence of air, if you can scuff it and keep it from oxidizing, it’ll solder just fine. There’s a few different liquids that help aid the process

[Pete9008]

The problem with aluminum is that it oxidizes in the presence of air, if you can scuff it and keep it from oxidizing, it’ll solder just fine. There’s a few different liquids that help aid the process

Any more details on the specific solders and flux here? I've tried this approach with standard tin lead solder before with very poor results and virtually no adhesion.

The only thing I have used successfully is something called lumiweld which did work but I've no idea whether it would be any good for electrical connections. It also needs a much higher temperature to work (380C).