Tesla Model S missing cells

[jsimonkeller]

As I discuss in another post, I am having to swap out my 14 Tesla Model S modules and aquired a complete TESLA MODEL S / X LITHIUM ION HV BATTERY PACK 75kWh 350VDC. I took all of the modules out of the case and swapped out the BMS cards and took voltage reading on my BatteryGO device.

When I was doing this, 13 of the modules gave out readings between the 3.1 -3.2 range, but one of the modules gave off strange readings (only 5 cells showed up) and on closer inspection, 46 individual batteries were missing from the module.

Has anyone seen this before? They were right out of a pack, so not a one of purchase.

Assuming this is not normal or expected, can I add one more module from another pack to make up the 14 I need and just make sure to get it properly balanced in the first charge?

[tom91]

Yes there are random variants of modules out there. They went up in bigger capacity of 18650s so did not need all 96s groups so one module got one less cell group.

this is not an issue if you account for it in your cell tap wiring.

[muehlpower]

bigger capacity of 18650s so did not need all 96s groups

Can you explain that in more detail? I did some research and found out that the 5.3 kWh modules are 6s74p. I thought the 75 kWh battery packs had the same modules as the 85 kWh ones, but only 14 instead of 16. No matter how I calculate it, I can't come up with a reasonable distribution with only 46 cells missing; there would have to be 74 missing. Then it would be 83s instead of the usual 84s for the 75 kWh pack.

[tom91]

https://www.ebay.com/itm/204427547974

Not 100% sure how they done it exactly but there can be dummy cells in there to ensure they still get the series connections using the 6S busbars.
I believe I had seen more pictures but cannot find any more via google, internet seems alot more difficult to search lately.

[manny]

In the picture you can see 5 bondwires going to the dummy cells. Normal cells get one bondwire.

Interesting solution

[muehlpower]

You could be right. If you look closely at the larger image, you will find 28 cells that have a different color. Together with the 46 missing ones, that adds up to exactly the expected 74!

[jsimonkeller]

Thanks again guys. That makes me feel better.

Now, my other question. The readings on the 13 batteries without any dummies came out as follows:

Battery 1
3.132
3.141
3.134
3.156
3.136
3.148 18.847
Battery 2
3.139
3.158
3.146
3.167
3.143
3.159 18.912
Battery 3
3.126
3.146
3.121
3.144
3.119
3.16 18.816
Battery 4
3.164
3.176
3.166
3.173
3.171
3.174 19.024
Battery 5
3.165
3.172
3.166
3.172
3.168
3.174 19.017
Battery 6
3.167
3.176
3.17
3.172
3.168
3.176 19.029
Battery 7 3.169
3.178
3.173
3.177
3.168
3.179 19.044
Battery 8
3.155
3.148
3.153
3.149
3.146
3.166 18.917
Battery 9
3.154
3.161
3.157
3.165
3.156
3.159 18.952
Battery 10
3.135
3.157
3.140
3.157
3.136
3.152 18.877
Battery 11
3.151
3.147
3.142
3.145
3.144
3.161 18.89
Battery 12
3.170
3.180
3.170
3.178
3.171
3.177 19.046
Battery 13
3.130
3.142
3.132
3.154
3.132
3.161 18.851

Battery 14 with the missing/dummy batteries in it came out as:

Battery 14
2.395
3.934
3.166
0
3.183
3.171 15.843

Obviously, the low reading of 2.395 concerns me, but with the higher reading next to it of 3.934 makes me wonder if because this battery is wired differently than the other ones, my BatteryGO is reading those incorrectly and I don't have a dangerous condition on cell 1. Together, the two of them average out to 3.1645, which would be consistent with the other batteries I tested. Since these all came out of the same pack monitored by Tesla's BMS up until going out of service for an accident, I feel like the Tesla would have prevented a cell from dropping this low. Thoughts on how to prove my thereom?

[midway]

In Tesla modules, the BMS control wires often fall off

[jsimonkeller]

In Tesla modules, the BMS control wires often fall off

Do you think this is why I am getting the strange readings on Cell 1 and Cell 2?

Do I need to take a multimeter reading on all 340 active cells and see if they all land in the same range to confirm that this is a wiring or programming issue and not a bad cell?

[muehlpower]

I don't trust your Orion or your pack tester. You don't know how your tester handles the missing cable! The cable is probably free-floating, which ruins the measurement if you pretend you have 6 cells. Finally, use a multimeter, as you have been advised to do here several times, and measure the voltage between each two cables to find the correct order. You need it anyway to wire the Orion correctly. You now only have 83 instead of 84 in series, which you also have to set in the Orion. Measuring the 370 cells is pointless.

[jsimonkeller]

Finally, use a multimeter, as you have been advised to do here several times, and measure the voltage between each two cables to find the correct order

I appreciate the response and I have no objection to using my multimeter, but up until this week, I have not had access to the entire module to do so.

I want to figure this but I need to better understand when you say to measure the voltage between each two cables, do you mean this (for example)?

I am not sure how I find the correct order or if I am trying to find the 6 qudrants like the photo below, but if you help me understand what I need to measure using the battery photo, I will certainly do this on Monday when I am back in the shop.

Thanks!

[muehlpower]

image.png (263.75 KiB) Viewed 1721 times

Don't worry about the physical position of the individual cells in the module. What you need to check is the connector that connects the module to the BMS. Your picture shows a Molex connector with 7 wires. This fits a module with 6 cells in series. The seven cables should be connected to the negative pole of the first cell, the positive pole of the last cell, and the respective connection points of the cells. Leave your tool plugged in during measurement and insert the measuring probes into the connector from behind to avoid short circuits, as this would certainly destroy the BMS cables in the module!!!! It is better to make a measuring adapter with the appropriate Molex connector and route the cables to easily accessible measuring points without the risk of short circuits, possibly even with 100mA fuses.
The order you need to find is :

module with 6 cells
1. 1neg
2. 1pos/2neg
3. 2pos/3neg
4. 3pos/4neg
5. 4pos/5neg
6. 5pos/6neg
7. 6pos

module with 5 cells
1. 1neg
2. 1pos/2neg
3. 2pos/3neg
4. 3pos/4neg
5. 4pos/5neg
6. 5pos

start with the black cable from your multimeter at any connection and use the red measuring cable to find the highest voltage. That is then point 7 (or poit 6 at 5 cell module) on my list. Now leave the red cable at the point with the highest voltage and use the black measuring cable to find the highest voltage again. This is then point 1 on my list. Now leave the black measuring cable at point 1 and use the red measuring cable to find a voltage of about 3.16V. This is then point 2 on the list. Continue searching with the red measuring cable for a point with about 6.32V. That is point 3. Points 4, 5, and 6 are then each 3.16 volts higher.

[jsimonkeller]

Hi muehlpower. Thank you for this detailed explanation of how to take the voltages. I will plan to do this tomorrow when I have time. I watched a video over the weekend about taking the voltages directly from the PLATES with the battery on its side with red cable on the top plates and the black cable on the bottom plates, taking reading across the plates. I am not sure if this does the same thing you were saying about, but here is how it came out. Each of the plates is numbered below randomly) for the 5 cell module.

Results:

1 to 5 = 3.15
1 to 6 = 12.66
1 to 7 = 6.33

2 to 5 = 3.15
2 to 6 = 6.33
2 to 7 = 0

3 to 5 = 6.33
3 to 6 = 3.15
3 to 7 = 3.15

4 to 5 = 12.66
4 to 6 = 3.15
4 to 7 = 9.48

5 to 1 = 3.15
5 to 2 = 3.15
5 to 3 = 6.32
5 to 4 = 12.66

6 to 1 = 6.33
6 to 2 = 0
6 to 3 = 3.15
6 to 4 = 9.48

7 to 1 = 12.66
7 to 2 = 6.33
7 to 3 = 3.15
7 to 4 = 3.15

Let me know if these readings conform to what you were looking for. If not, I can start over tomorrow.

I am guessing these readings do not help us determine how to correctly wire the molex connector. Feels like your solution was much more efficient, but I missed it this morning before going to the shop.

Thanks!

[jsimonkeller]

OK. So I went to the shop and followed muehlpower's instructions. On the 5 cell battery, here is how it came out:

There was never a 3.15v option, but two 9.48v readings side by side. I am not sure if this was expected or means something is wrong.

When I put the positive pole on the module positive connection and then touched the negative pole to each of the molex connectors, I got this:

Thoughts?

[muehlpower]

Two things bother me. In your first measurement, the top left is labeled “GND,” which makes sense. However, in the second measurement, 15.82V should be measured there. In the bottom right corner, the first measurement should be 3.13V and the second 12.85V. But both times it is 0V. From this, I conclude that the line to the cell is interrupted. Check the ribbon cable on the underside of the module, especially at the point labeled “1”. According to my logic, this point should be connected to the lower right pin of the Molex. 3 and 4 appear to be bridged, which I think is OK.

Translated with DeepL.com (free version)

[jsimonkeller]

Thanks muehlpower. I will head back to the shop today or tomorrow and do some more investigating with the ribbon cable and retake the voltages to make sure that the top left in the second set of voltages was actually 0. Is it telling that in my "plate to plate" measurements, I never got a reading of 15.8 in any of the combinations? There were a couple of points that read ZERO (2 to 7 and 2 to 6), which makes me wonder if that is a clue to anything...

I guess some positive news is that the plate to plate readings all showed consistent cell voltages, so this would be a wiring issue most likely.

[muehlpower]

The more I think about it, the more I am convinced that the connection from cell 1+ to the board is interrupted. Your previous measurement with your tool says exactly the same thing. 3 bridged with 4 and line 1 free floating! I have marked which plates should be connected to which positions of the Molex connector (in yellow). I assume you are using a board like this!

[jsimonkeller]

I am using one of the new Felten branded BMS swap cards like the one in your photo. I had 14 of them, so I did try more than one on this 5S battery to make sure the readings were not from a faulty BMS swap card.

I will use your numbers above to check out the readings on the plates today and report back.

Here is a photo of the bottom connector to the BMS swap card from the battery. It does not look like anything is torn, but it feels like the cable it not fully connected to the connector piece.

I thought maybe because it is a 5S that the whole ribbon is not hooked up.

[muehlpower]

I can clearly see damage. The dark parts are the conductor tracks. One of the three is completely torn. It can't work like this, and my suspicion has been confirmed. You don't need to take any further measurements.

[jsimonkeller]

Well . . . I was afraid of this when I took the photo yesterday. Sorry I missed the tear originally.

Now, the internet is full of advice for what I can do at this point in time, but I always prefer the thoughts from this group of learned folks.

1. Fix the connection. How realistic is this at the connector vs. where it connects to the plates.

Additional photos of the torn connector:

2. Buy a like battery. We have a link to another 5S battery for sale on EBAY that was posted above. From that battery, I can either:

(i) steal the working non torn cable and attach it to this battery, but I will need to be able to make the internal plate connections, which I know is more advanced work; or

(ii) use the 5S battery in place of this one (I KNOW THE REASONS WHY THIS IS NOT ADVISABLE BUT WOULD LIKE TO WALK IT THROUGH). My car setup would allow me to put 6 full 5.3kWh batteries in one box, 6 full 5.3kWh batteries in the second box and one module under each seat. For the 14th module, if it showed higher cell voltages, I would hook up the 13 batteries using a 76 cell profile and bring all of those batteries up to within .01 of the new 14th battery and then shut everything down, hook up the 14th battery, reprogram the orion for 83 cells and then boot back up and continue to charge and balance. I know that the wear on the new battery could be more of less than the other 13, but is it enough to concern myself about shortened battery life at this point?

3. If I am going to swap out batteries, I get a full 5.3kWh vs. 4.7 kWh battery so I have use of 84 cells on my rebuild. Again, I understand the reasons why modules from different cars is not advisable, but wondered if I took the necessary steps ABOVE, should I go for more cells using the same battery architecture.

4. Go with 13 batteries and less total pack voltage at 76 cells, which is my last resort.

[muehlpower]

Flat cables are easy to repair. Find a spot where the cable is wide, remove the insulation, and solder a cable in place, which you then connect directly to the plug.

[jsimonkeller]

Thanks. I will give this a shot next week. I fixed my Dad's old Black Knight pinball machine and had to deal with a lot of solder work on the analog copper wiring for the bumpers and such, so it will not be too unfamiliar to me.