Jk BMS shut down and unable to restart

[Skypower]

Are that washers under the BMS cables?
I would place them directly onto the busbars keeping locknuts and nuts on top.

Those washers have been flattened to remove the sheared wipe and the edges buffed off as to not cut or stress the top laminate of the buss. The washer puts pressure over a larger area of the buss. Above that washer is the ring terminal then a Bellevue washer and finally the nut with a smooth/flat bottom. Normally you’d be right putting the terminal right on the buss, but there’s only 2 amps, no measurable resistance and I put clamping area ahead of terminal position. You might notice that the edges of all the flexible buses are have an orange sealer(not silicone) so that with time and moisture they will be less likely to delaminate.
Yeah, I know, I got carried away.

 

[DIYrich]

When problems go away and I’m not sure why, just gets me more concerned. Can It happen again?

One of two possibilities:
1) cell fails and BMS protects.
2) BMS fails and shuts off battery

Either way, you are OK, just don't have the battery.

 

[fx99]

Interesting update:
now that I am able to do some proper load testing with the Inverter, I decided to screen record the BMS app while doing so.
Here is the video, I had to upload it on YouTube:

Watch the discharge load and the voltage at cell #7, which is the new position of the faulty #6 that was always the problem.

Currently the BMS measures the cell to be back to normal, but I guess the cell needs to be replaced after all, right?

 

[Skypower]

Is there any heat at that terminal? Bad cell or bad connection. It jumps back hella fast

 

[fx99]

Is there any heat at that terminal? Bad cell or bad connection. It jumps back hella fast

you are right, they jump back very fast.
I have done some more testing and it's always instantly back to normal once the load goes away.
there is a bit more heat at the terminal, 5 degree celcius after a few minutes. For the experiment I placed the temperature probes of the BMS onto the plus terminals of cell #7 and randomly selected cell #2.
in the following video you can also see that 3 more cells start to drop in voltage as the load increases to 2000w.
The video is a bit longer to see the temperature change.
for the voltage measurements i have put timestamps with short descriptions in the comment section, so you guys can quickly skip through the video instead of having to watch the whole thing.

 

[Skypower]

I’m thinking you may have some oxide on your terminal. With a good volt meter you could probe from the buss top to the center post of the cell and see if there’s any variation under a constant load. Probably something under.005 volts with a 50 amp load would be my guess from memory. The main thing is that they are alike. That will instantly show a bad weld and/or buss connection. If it’s a bad buss connection, I’d remove all the busses, studs and antioxidant then buff them with Scotch Brite. You will see an immediate aluminum color change even if it looked good before. Carefully blow out any git from the holes not at the burst vent. Careful with those busses, they are the perfect length for trouble. I put all the studs in and finish one at a time with at least the one nut loosely at each end so it can’t get flipped or swung off.

 

[fx99]

I’m thinking you may have some oxide on your terminal. With a good volt meter you could probe from the buss top to the center post of the cell and see if there’s any variation under a constant load. Probably something under.005 volts with a 50 amp load would be my guess from memory. The main thing is that they are alike. That will instantly show a bad weld and/or buss connection. If it’s a bad buss connection, I’d remove all the busses, studs and antioxidant then buff them with Scotch Brite. You will see an immediate aluminum color change even if it looked good before. Carefully blow out any git from the holes not at the burst vent. Careful with those busses, they are the perfect length for trouble. I put all the studs in and finish one at a time with at least the one nut loosely at each end so it can’t get flipped or swung off.

maybe you are right, i didn't use proper scotch brite pads previously, just some coarse kitchen sponge and pressure. In hindsight that sound pretty stupid, but I thought since scotch brite sponges are pretty common, any coarse sponge and a bit of elbow grease will do. Looking back that's probably the only step I cheaped out on, because scotch brite pads are so expansive and I thought aluminum oxide is relatively easy to remove.
so yeah... if that thought was way off I feel really stupid now ?

 

[Skypower]

Don’t beat yourself up it’s only a guess at this point. If you do it, don’t use any power tool. You want keep it flat and smooth not valleys and scratches. But check before you take it apart with a good volt meter.

 

[Skypower]

What I use. Don’t use steel wool ?

 

[fx99]

Don’t beat yourself up it’s only a guess at this point. If you do it, don’t use any power tool. You want keep it flat and smooth not valleys and scratches. But check before you take it apart with a good volt meter.

okay thanks you, I will check with the multimeter first. The best one I've got is a Metex ME-31. I borrowed it from my dad, it's pretty old, but I think it was quite expansive once. Don't know if it's good enough for the job, but I will just try it out tomorrow.
found following infos online about the Metex:

Manufacturer Specs​

DC Voltage

400㎷ to 1000V ?0.5% + 3dgts

Whats the best coarseness of scotch brite pads for cleaning the terminals?
Are these very fine ones okay?
3M Scotch-Brite 64926 Handpad 7447 PRO Very Fine A 152 mm x 228 mm Red (Pack of 20)

https://amzn.eu/d/2bvYITX

 

[Skypower]

okay thanks you, I will check with the multimeter first. The best one I've got is a Metex ME-31. I borrowed it from my dad, it's pretty old, but I think it was quite expansive once. Don't know if it's good enough for the job, but I will just try it out tomorrow.
found following infos online about the Metex:

Manufacturer Specs​

DC Voltage

400㎷ to 1000V ?0.5% + 3dgts

Whats the best coarseness of scotch brite pads for cleaning the terminals?
Are these very fine ones okay?
3M Scotch-Brite 64926 Handpad 7447 PRO Very Fine A 152 mm x 228 mm Red (Pack of 20)

https://amzn.eu/d/2bvYITX

That looks to be the right stuff.

 

[Skypower]

Ok, I put a 60 amp load on one battery and got some measurements. You will need a good volt meter to do this test. Mine reads four places past the decimal (.xxxx). With load, center post of cell to top of buss .0002 to .0007 From center post of one cell to center post of another cell (through the buss) .0025 volt typical. I couldn’t remember so I ran the test. You could probably have three times these numbers and it would probably be inconsequential. What you are looking for is something like .100 or more, likely a volt.
IF that’s your problem. When I assemble, I use an internal resistance meter to verify everything, but it’s useless when the battery is in use and connected to the bms so a good volt meter can work on the fly.

 

[fx99]

What you are looking for is something like .100 or more, likely a volt.

Wow seems like you are right ??

Here are the measurements I just took
BMS good cells: #2 #3 #4 #5
BMS problematic cells: #1 #6 #7 #8

1000W load through inverter, ~44A

center post of cell to top of buss
#1 12,6mV (+ side); 116mV (B- side)
#2 24 mV (- side) ~22mV (+ side)
#3 5 mV (+ side) 3,8mV (- side
#4 6,9 mV(- side)
#5 4,5mV (+ side);
#6 52,3/130/330/650mV (-); ~50mV (+)
#7 6,6mV (+ side); ~150 / 500mV (- side)
#8 40mV (- side); 33mV (B+ side)


From center post of one cell to center post of another cell (through the buss)
#5 to #6 58,8mV
#6 to #7 >300mV
#7 to #8 38,4 mV

Some measurement, especially those with higher voltages, weren't consistent, so I noted voltages from multiple measurements.
I didn't take all measurements for all cells because some were hard to get to with the BMS connected. But the data is sufficient, isn't it?

 

[Skypower]

That’s a lot of variance. Bad connections, oxidation or delaminating buss? Clean off the anti oxidant, buff both contact surfaces very (lightly on the buss, it can bend and the coating is very thin) for both ends of the worst and reassemble and test for improvement. Fingers crossed.

 

[Skypower]

Why it was changing was the junction was getting hot. Some people with huge chunky billet four hole buss were having trouble too. But that was probably a combination of oxidation and bar stock not always being flat. When you torque down a 6mm to 4Nm that’s about 700 pounds of force. That’ll flatten to conformance the conventional and laminate stuff pretty well but not the chunky stuff.

 

[fx99]

I think it's finally done ?
I got the Scotch Brite delivered and really took my time with removing the oxidation layer and cleaning everything like 10 times over. After about 5 hours I had everything reassembled and i have just tested it for about 5min at 40A to 90A.
With the load applied the BMS measured a cell voltage difference of 20mV to 40mV.
I hope that's okay ?

From center post of one cell to center post of another cell (through the buss) .0025 volt typical. @Skypower

From your cell to cell measurements I assume that your total difference accross all cells is still alot lower than mine.
Are my measurements still acceptable?

Edit: I found a video from Offgrid Power Solutions where his BMS measures 30mV to 55mV at 150A (12v battery). That seems comparable to my results.

 

[Skypower]

You’re probably fine. At a low state of charge with high loads, the cells can show what the bms is telling you and that’s normal. In the flat of the curve you’ll see the best delta. It might be a little out of whack at the upper knee. After a couple of trips to full charge and active balancing, I got a feeling it’ll be fine now. I typically first clean the surfaces with acetone to remove any film, buff with scotch bright till I see that uniform bright clean color, blow off the dust with air hose or dust off, apply a thin bead of anti oxidant and assemble. Perhaps a couple hours for 16 cells, clean, prep, grease, assemble buss&balance leads, torque and test resistance.

 

[fx99]

You’re probably fine

Great to hear that. Thank you so much man ??
I would have probably gone crazy without your help ?

At a low state of charge with high loads, the cells can show what the bms is telling you and that’s normal. In the flat of the curve you’ll see the best delta. It might be a little out of whack at the upper knee

You mean the curve of state of charge on the x-axis and battery voltage on the y-axis right?
I measured following battery voltages without load:
26.71V Multimeter =3.339V/cell
26.59 BMS =3.324V/cell
3.33V/cell = 90% SoC
3.35V/cell = 99% SoC
Which would mean that, considering that I am at the beginning of the upper knee of the curve, the voltage delta between the cells should be normal, correct?

After a couple of trips to full charge and active balancing, I got a feeling it’ll be fine now.

That would be great ? How much cell voltage difference do you think is considered very good (at different SoC)?

I typically first clean the surfaces with acetone to remove any film, buff with scotch bright till I see that uniform bright clean color, blow off the dust with air hose or dust off, apply a thin bead of anti oxidant and assemble.

I went with a similar procedure ? didn't see a huge color difference on the aluminum but a bit was noticeable. I also torqued down the nuts on the busbars a bit more, that might have helped for me too. the terminal threads with the helicoil inserts are actually pretty strong.
btw does your scotch brite also make a lot of dust when sanding the terminals? probably because of the relatively sharp edges of the terminals.. I used scotch brite pads for the first time and was a bit annoyed by it

 

[Skypower]

From what I’ve seen, the upper knee starts at about 3.4 volts and right about 3.42 is where you usually start to see when the cells are out of balance and start to run if it hasn’t been done before or many cycles of not getting into balance voltage. 3.42 is where I set the start balance on the bms. The system charges then slows to 3.45 volts per cell, usually holds it there for about 15 minutes and releases all current until 3.4 volts float so long as there’s sufficient solar. The lower knee seems to be around 3.15 volts.
I have three batteries and I just ran a test under load (dryer+house) about 20 amps per battery. They are at 90% state of charge and the voltage sagged to 3.333 volts. Two had a delta of.005 and one was.007. If the load was doubled you could almost double the delta number. The delta at 20% SOC is much worse with a high load. You will may see little hiccups with delta (.015) even without load at different points usually while passing back and forth through the upper knee. Since all three batteries are from different sources the all do things at different points. I don’t really look at voltage vary much except to see if it makes it to balance voltage several times a week. Mainly just the state of charge. I have Jk that recently has not reported full state of charge even tho I know it’s been there(about 8% off) so I have to look into it. Yes the 3M cloth makes a mess if you run it over the cornerstone/edges but I just try to keep it on top. You’re the second person to claim that they have helicoils. I don’t understand why unless they were fixing bad threads or mistake to begin with. There’s no need and added cost. Weird. I’m not so much worried about stripping out threads as turning the terminals post in the cell. So I limit to 4Nm.

 

[fx99]

Interesting stuff, good to know at which SOC some voltage delta is normal so I don't have to worry.
You really seem to know your battery system extremely well ?

I have Jk that recently has not reported full state of charge even tho I know it’s been there(about 8% off) so I have to look into it.

Do you mean that your JK wanted to overcharge the battery? sounds dangerous ?

You’re the second person to claim that they have helicoils. I don’t understand why unless they were fixing bad threads or mistake to begin with. There’s no need and added cost. Weird.

I have heard that helicoils are stronger than the original intact thread, even in a stripped thread. Just sounds like good marketing on one hand but on the other hand I don't think EVE throws their money out of the window, they probably did their testing.

I’m not so much worried about stripping out threads as turning the terminals post in the cell. So I limit to 4Nm

I didn't think about that, but luckily nothing happened. Good to have in the back of my head for next time. ??