Wiki Sudden Loss Of Range With 2019.16.x Software

[islandbayy]

I find it strange we have gotten to this point.

I don't find it strange, I find it Sad :-( Though I'd love to get a replacement for my V1 90 pack. And I have a dying brick in my 75D thats 50mv out of balance below 30% SOC now, so that will probably get one eventually. Thankfully 4 years left for the 90d and 4 3/4 years left on the 75D MX battery warranty...

 

[_jal_]

That doesn't make any sense. There's no way to bypass a single parallel group of cells, and the output voltage of the pack must equal the sum of all 84 parallel groups in the series string.

So the modules are in series, and the modules have bricks in series themselves composed of groups of individual cells. So if one cell has a short, conceivably all the current could try to get through that one cell. Is there a way to disconnect that cell from the pack?

 

[MP3Mike]

The REALLY odd thing was one cell was dead. Only 83 of the 84 cells reported voltage and the max pack voltage only reach 347 volts (instead of 352 that it should have with all 84 cells)

It didn't even report 0v? That sounds more like a problem with the CANbus logging/reporting to me.

 

[MP3Mike]

I don't find it strange, I find it Sad :-( Though I'd love to get a replacement for my V1 90 pack. And I have a dying brick in my 75D thats 50mv out of balance below 30% SOC now, so that will probably get one eventually. Thankfully 4 years left for the 90d and 4 3/4 years left on the 75D MX battery warranty...

Are you loading the current firmware on your Model X, or are you keeping it on an old version like your Model S? (I'm pretty sure that is what you have said.) Someone just recently got the warning that usually indicates that they are going to get a replacement but their ck had a 204mv in-balance. So 50mv may be a long ways from what Tesla considers a problem.

 

[omarsultan]

I don't find it strange, I find it Sad :-( Though I'd love to get a replacement for my V1 90 pack. And I have a dying brick in my 75D thats 50mv out of balance below 30% SOC now, so that will probably get one eventually. Thankfully 4 years left for the 90d and 4 3/4 years left on the 75D MX battery warranty...

Yeah, sad is probably more apropos. Assuming things get resolved on the P85, my attention is going to move over to my wife's X which also has a V1 90 kWh pack.

 

[Dave EV]

Maybe my case is unusual, I don't know. I am probably an unusual case in the sense that I don't use supercharging(only 25kWh DC charging on my battery).

View attachment 500109View attachment 500110

@ran349 Wow, that is one well-balanced and matched pack! Either you got a really good set of cells in your pack, or something you're doing is really helping to keep your pack well balanced. Can you share your typical charging habits?

View attachment 501022 View attachment 501023

Follow-up: stayed unplugged all week and ran the battery down--here is the battery at 50% and 10% - the same two clusters of modules are problematic. I think I will finally charge tonight and see what happens--I have only gone below 10% a couple of times ever and I don't want to go any lower now lest the battery heater decide to kick on and leave stuck somewhere.

@omarsultan Interesting here that in this snapshot balance hasn't gotten all that worse all the way down to 10%. Yeah, it's up to 150mV, but you were at 78mV at 90% which is really bad.

So, I got "the message" today. Interestingly, it popped up while running errands this morning (what Saturday morning is not complete without a trip to Home Depot). I'll call service and keep you apprised. TM-Spy data below. The one big change I see from the pics I posted 8 days ago (#9685) is the NomFullPack value has dropped another 5kWh. I'll have to go back later and see if I have screen grabs for a similar SoC to see if anything else is different.

View attachment 501803 View attachment 501805 View attachment 501806

So 2 days after discharging down to 10%, the imbalance has not only gotten much worse, but you finally also triggered the BMS into realizing that your pack has got a serious enough issue that it needs to be serviced.

I assume that's the root of the matter and it crossed some threshold. I am also curious about the 5kWh loss--if that came from more aggressive capping or something else happening in the module/brick.

Without doing a full charge/discharge cycle, it is very difficult to accurately estimate the actual capacity of the pack. So I think that discharging the pack down to 10% and then charging back up to 90% (I assume) allowed the BMS to finallize realize that one of the modules was significantly down in capacity compared to the rest of the modules. And since your pack is only as strong as your weakest module, once it realized this, it reduced the reported capacity of the entire pack.

It's pretty safe to say that there are probably a number of internal checks the BMS runs to determine the overall health of the battery pack. Keeping track of individual module capacity is one of them. Also keeping track of the overall balance and relative SOC of each module is another. One module getting significantly weaker than the rest, or perhaps getting significantly more out of balance than the rest, are likely triggers for generating error messages.

 

[islandbayy]

Are you loading the current firmware on your Model X, or are you keeping it on an old version like your Model S? (I'm pretty sure that is what you have said.) Someone just recently got the warning that usually indicates that they are going to get a replacement but their ck had a 204mv in-balance. So 50mv may be a long way from what Tesla considers a problem.

I bought the Model X with Firmware 9 already on it. I got it used last March. Since it had AP2 and FSD, I kept updating it, that was prior to knowing about Charge & Battery gate (Not hit with batterygate, but definitely chargegate).

I kept the Model S with AP1 on Firmware 8.1, one of the last released, because after seeing the complete crapstorm F-Up of a user interface of firmware 9, that would completly ruin the driving experience for me. At least the MX could take advantage of the new AP2.x++++ Features, the Model S would gain basically nothing.

The 50mv imbalance, while not yet too bad, at lower SOC's my rated range reported does jump around a bit, but no not enough to get a replacement pack or repair. But, with the lower voltage of the 75, heavier weight of the MX, and the heavy towing I do with it, I feel that imbalance will grow just from normal stress of using the vehicle. So guessing it's just a matter of time. When first noticed, it was only a 35mv imbalance, so it is growing.

The 90D, yah, just garbage battery to begin with (At least V1). I'm enjoying my blazing fast 94kw charge rate for now, but I know as soon as it gets updated to 10.x firmware, that all ends. As of right now, I can squeak out my annual cross country road trip from Milwaukee to Creator Lake, Oregon, (Actually 15-20 minutes south, we go to Train Mountain to camp and run our 7 1/2" gauge ride on train on their 40+ miles of track through the mountain, it's AWESOME! www.TMRR.org ) Once updated to V10, based on my MX performance, and difference in charging, I estimate that the 2 day drive (I do non stop other than to charge), will end up being 3 days each way, unless the route doubles in superchargers so I don't need 90+% at each stop. Or I drop my drive speed down to 55mph, which still adds a crazy amount of drive time, especially when the speed limit in route is 80mph.

 

[JRP3]

So the modules are in series, and the modules have bricks in series themselves composed of groups of individual cells. So if one cell has a short, conceivably all the current could try to get through that one cell. Is there a way to disconnect that cell from the pack?

Each cell is fused to the brick with thin bond wires. If current exceeds what the bond wires are rated for the wire melts and the cell is disconnected.

 

[DJRas]

It didn't even report 0v? That sounds more like a problem with the CANbus logging/reporting to me.

If it was just a reading problem then he wouldn't have seen the loss of range.
The total pack voltage reported by the BMS also only works with 83 cells. The BMS must have just skipped the brick. Looking at the brick to brick differential the missing brick is in module 7.

 

[MP3Mike]

If it was just a reading problem then he wouldn't have seen the loss of range.
The total pack voltage reported by the BMS also only works with 83 cells. The BMS must have just skipped the brick. Looking at the brick to brick differential the missing brick is in module 7.

As far as range loss it could be a phantom loss that isn't actually there if they actually tested capacity. (If the BMS thinks only 83 of the bricks are functional and they actually all are the range estimate would be off.)
​

If it isn't just a reading/reporting problem, i.e. the brick is actually at 0 volts, for the car to still be functional all of the cells that still have their fuses intact, would have to be completely shorted out. (Which is highly unlikely.) I guess the most likely cause would be a failed BMS module, or wire/connection, I would have thought Tesla would detect that kind a fault and present an error, but maybe this is a case that they didn't consider likely and didn't handle it.

Though that seems like it could potentially be a very dangerous situation, if the BMS doesn't know the voltage of that brick it could potentially over charge it, which would be bad. You should have that person open a case with Tesla, if they haven't already, with the details that it appears that the BMS doesn't know about one brick and see what they say. I'm really interested to hear more about this case and what happens.

 

[MP3Mike]

I tested a 2018 75kWh Model S that had a "sudden range loss". Interestingly enough, the MaxV for one cell hit 4.199 at 99.5% there was higher than normal cell differential of 27 mv at 99.6% and 10 mv at 99.2% after settling.
The REALLY odd thing was one cell was dead. Only 83 of the 84 cells reported voltage and the max pack voltage only reach 347 volts (instead of 352 that it should have with all 84 cells)

Additionally, the charge current was still 5kW at 99.6%

View attachment 501860

@wk057 Have you ever seen a case where the CANbus data didn't include readings for one brick?

 

[glhs272]

I tested a 2018 75kWh Model S that had a "sudden range loss". Interestingly enough, the MaxV for one cell hit 4.199 at 99.5% there was higher than normal cell differential of 27 mv at 99.6% and 10 mv at 99.2% after settling.
The REALLY odd thing was one cell was dead. Only 83 of the 84 cells reported voltage and the max pack voltage only reach 347 volts (instead of 352 that it should have with all 84 cells)

Additionally, the charge current was still 5kW at 99.6%

View attachment 501860

Don't forget that Tesla actually made a change to 75kwh packs where they deleted one bricks worth of cells in one of the modules. This could very well be that type of pack. I believe Tesla did this because of crash testing showing a vulnerability in the corner of the pack. Removing the cells from that spot was the cure. There are pictures of this type of module somewhere around here on TMC. It's an oddball module where they removed the cells and put wires in their place to bridge between the bus bars.

 

[DJRas]

Don't forget that Tesla actually made a change to 75kwh packs where they deleted one bricks worth of cells in one of the modules. This could very well be that type of pack. I believe Tesla did this because of crash testing showing a vulnerability in the corner of the pack. Removing the cells from that spot was the cure. There are pictures of this type of module somewhere around here on TMC. It's an oddball module where they removed the cells and put wires in their place to bridge between the bus bars.

That must be it.
I couldn't find anything like that before I posted.

 

[mthoring]

That must be it.
I couldn't find anything like that before I posted.

It was speculated that 75-packs which included a B on the sticker contained fewer cells.

 

[supratachophobia]

That doesn't make any sense. There's no way to bypass a single parallel group of cells, and the output voltage of the pack must equal the sum of all 84 parallel groups in the series string.

I would assume whatever reader he's using skipped some CAN frames.

 

[Leeboy22001]

The scale can be deceiving, this my 60kW pack at 30% SoC, at two scales

 

[DJRas]

It was speculated that 75-packs which included a B on the sticker contained fewer cells.

 

[Ferrycraigs]

View attachment 502140

Now that is interesting. They can’t be designing these new variants just because some 75 or 85 batteries are failing. To design new variants you would imagine there would need to be a substantial demand.....

 

[VT_EE]

Now that is interesting. They can’t be designing these new variants just because some 75 or 85 batteries are failing. To design new variants you would imagine there would need to be a substantial demand.....

Regardless of the current warranty issues, there will be demand for replacement batteries as these vehicles age. Tesla would be stupid not to have something people could purchase to keep their cars running. Nobody paying for a replacement pack out of pocket wants a refurbished pack. I'm sure they'll do the same in 4-5 years for Model 3s and Ys.

 

[swegman]

View attachment 502140

What does the 5B on the sticker designate?