Wiki Sudden Loss Of Range With 2019.16.x Software

[Battpower]

This has nothing to do with fires. Nothing whatsoever.

I don't think you can blame people for making that connection, although I agree that the discussion muddles reasoning.

attempted to turn the range loss issue into a fire issue

While several strong arguments / explanations have been presented, I think the link to fires specifically could have benefitted from more openness on Tesla's part.

Even the Model 3/Y BMS uses the same tried and true algorithms.

I've read posts suggesting 3 / Y BMS is either simpler, technically more advanced, smarter design than S / X, but also that in some ways it is less sophisticated in its algorithms and doesn't work the same way. Which is the case? Any links to what you think represent accurate info?

 

[fbitz777]

As noted previously, the "Abundance of caution" update, which came down way after they publicly noted, barely did anything at all. A few thermal deltas were tightened a couple degrees, and the SoC vs allowed temperature table was tweaked almost insignificantly.

That's it. It's an update that has no real effect on anything in order to say they did something. While I think this is pretty dumb, and doesn't do anything to actually prevent fires... they're not technically lying. Tightening a few allowed ranges could be seen as being cautious... but in reality, this can't really do much besides make the thermal systems work harder than they need to, as some have noticed.

.

I case of Regen (or specifically lack of) they tightened by more than a few degrees! My car is Regen limited below 60F ..used to be ok at 50F. Regen is basically useless to me 8 months a year.
And to the point: have you yourself seen a pack that has been limited to say 4.15V/cell (i.e capped) and that has recovered yet? Or you are mainly dealing with packs on a bench that you are trying to diagnose?

 

[BigNick]

So that reports that it will cost Hyundai $1.8 billion to replace the the LG batteries in 54,090 vehicles in Korea alone. So about $33k per vehicle, and a Kona EV started at only ~$38k. It seems excessive that replacing the battery would cost ~85% as much as the original cost of the vehicle. (But I suppose that includes recycling/disposal expenses as well.)

I don't know if the Kona is designed as a skateboard, or if there is significant disassembly required to replace the battery. I'm thinking because it's not a dedicated EV, that it might require more disassembly than an S or X to replace the pack.

 

[wk057]

I don't think you can blame people for making that connection, although I agree that the discussion muddles reasoning.

While several strong arguments / explanations have been presented, I think the link to fires specifically could have benefitted from more openness on Tesla's part.

Have seen zero arguments for this being fire related that could be considered "strong"... everything about it in this thread is either raw speculation, links to unrelated information presented as if it were related, and other hand waving. Nothing substantial, and nothing whatsoever that even remotely ties the range loss issue with fires.

I've read posts suggesting 3 / Y BMS is either simpler, technically more advanced, smarter design than S / X, but also that in some ways it is less sophisticated in its algorithms and doesn't work the same way. Which is the case? Any links to what you think represent accurate info?

Honestly, no one else seems to have any real information. As far as I'm aware, I'm the only person in the world who has done extensive reverse engineering on the hardware and software and publicized anything about the same. Everything else seems to be pure speculation.

There are improved hardware aspects, for sure, mostly related to production efficiency. But the software aspects remain pretty much the same as the S/X.

I case of Regen (or specifically lack of) they tightened by more than a few degrees! My car is Regen limited below 60F ..used to be ok at 50F. Regen is basically useless to me 8 months a year.
And to the point: have you yourself seen a pack that has been limited to say 4.15V/cell (i.e capped) and that has recovered yet? Or you are mainly dealing with packs on a bench that you are trying to diagnose?

This sounds like tweaks to Range Mode, and changes made with relation to the Smart Preconditioning stuff, nothing to do with the "abundance of caution" update.

Tesla used to precondition the battery slightly in almost all cases when plugged in so that more regen was available. They ditched this in favor of the "Smart" preconditioning stuff, which can leave the pack colder than before in a lot of cases.

As for battery caps, not 100% sure what example you're referring to... but there are zero cars in my fleet today that have any voltage caps.

 

[Battpower]

have you yourself seen a pack that has been limited to say 4.15V/cell (i.e capped) and that has recovered yet?

I wondered the same thing. Various claims in posts but often without hard data.

there are zero cars in my fleet today that have any voltage caps.

So is that saying that some had caps and have since reverted to 4.2v?

 

[BigNick]

Is there any sort of way to determine the BMB hardware revision if our battery pack's part number is not on the list? Raw CAN data containing a HW ID?

 

[wk057]

I wondered the same thing. Various claims in posts but often without hard data.
So is that saying that some had caps and have since reverted to 4.2v?

As noted, mitigation of Condition X or Z would result in changes to the permitted voltage range in order to ensure voltages would be within a defined range despite a high margin of error present pre-mitigation. This can take the form of limiting voltage below normal maximum if required by the algorithms.

Is there any sort of way to determine the BMB hardware revision if our battery pack's part number is not on the list? Raw CAN data containing a HW ID?

If you're not on the list, have a car built after Q1'15, or before about Q2'13... you can be reasonably certain you don't have BMB v1.5. The further outside the date range your pack is, the more likely it isn't BMB v1.5. Additionally, if your pack is new enough to have the HV fuse access cover present on the underside of the pack (likely the case in your July '15 P85D, specifically), you don't have BMB v1.5.

There are ways to query BMB info from the BMS.... but I'm not going to get into that publicly. You can, however, get your pack HWID from 0x5D2 16-bit @ bytes 4+5 (0 indexed) when byte 0 = 0x0A ... assuming I'm reading my notes correctly. It's a low frequency message. This isn't super helpful information, but I believe HWID 70 (decimal) is the highest HWID affected. There are many below that which are not, though.

Edit: Looks like the pack HWIDs with BMB v1.5 in my database are 27, 31, 39, 41, 42, 49, 50, 57, 58, and 70. This likely is not an exhaustive list.

 

[Matias]

1031043-00-E is HWID 57, or upgraded to Ludicrous HWID 70. This is BMBv1.5.

If you have the issue, you can't access the range that would be "lost" with 100% certainty right now on any earlier update anyway. You risk a shutdown at range showing > 0 miles if the BMS gets an anomalous Condition Z induced reading while attempting to utilize that area of the voltage spread, which is quite likely if you have the issue. As noted, this is why I previously suggested folks update in the first place.

If your pack has the issue, after an update beyond 2019.16 (I'd suggest at least 2020.36 if you decide to do so) your gauge would show lost range soon after the update. You've only got a few % chance of that at worst, so, your gamble to make the call on. If so, it will take months and thousands of miles of use to recover, but the range shown will be somewhat guaranteed usable.

)

If I understand this correctly, before the update, condition Z cars had false range readings (because intermittent low voltage readings did not affect range algorithm) and car could shut down before battery reading reached zero. With the update range algorithm now registers random low voltage and now shows lower range. Now range display is more accurate and you can drive to 0. But in addition to this car also limits the maximum voltage, so you lose additional range from the top. Eventually after few thousand miles, you will get most range back (because car learns which voltage readings are false).

 

[Matias]

No, you are completely missing what he said. Here is an example: Pre-update your car reports 260 miles of range, but shuts off when it reports 30 miles are left. Post-update it reports 230 miles of range and shuts off at 0 miles left. You have lost no available range, the estimated range is just more accurate to what you can actually use. Then after the remediation has completed you might get returned to have 255 miles of range available. (This is assuming of course that your battery pack suffers from a BMB problem.)

You are not taking into consideration that before the update car could charge to 4.2V, after update to 4.1V (after few thousand miles you get most miles back8.

 

[Droschke]

Eventually after few thousand miles

the check is in the mail

 

[Battpower]

You are not taking into consideration that before the update car could charge to 4.2V, after update to 4.1V.

It needs to be made clear if cars are still capped at 4.1v.

Something I read seemed to suggest that the (displayed) range (may be not voltage) gradually increases as the system learns about the battery. But of course if the battery is capped at 4.1v then that means it is storing less energy than at 4.2v. And of course why would a faultless battery need to be capped at all?

 

[Matias]

That's exactly what it's doing after the update according to wk057.

Post update it allows lower voltages when discharging, but temporarily prevents higher voltages. After enough cycles/miles, it allows nearly all of the higher voltage range again.

Where in wk’s post it says that after the original update (2019.16)
it allowed lower voltages when driving near zero?

 

[Matias]

Assuming a pack suffers from Condition Z, and was reporting 260 miles of range when fully charged:

• Pre 2019.16.x it would shut down prematurely, at ~30 miles reported remaining. So you only had 230 miles of capacity that could be used.
• Post 2019.16.x the car only charges to ~230 miles of reported range, all of which can be used.

So in both cases you had ~230 miles of energy capacity you could use. In the first case it was shutting down before it hit the bottom, in the second case the bottom capacity is unlocked and the top is locked. Why is that so hard for you to understand?

Where in wk’s post it says the bottom range was unlocked in
2019.16?

 

[Matias]

Interesting. So they knew about condition Z in 2019.16 and immediately unlocked the bottom of the battery (ignoring erroneous low readings) but locked the top until they could figure out a way to reliably read the voltage?

This is the question I’m trying to ask. Because if they didn’t immediately unlock the lower end but immediately capped to 4.1V, then the range actually dropped (not just became more accurate). Of course this is mostly historical question now.

 

[cpa]

Thanks for your response in writing, to be honest I was surprised by the negative feedback I got on my post. I can understand your disappointment in Tesla and perhaps I would feel the same if I’d have that experience all the way from 2019. Now that @wk057 connected the dots I thought people would be more appreciative and less angry but that doesn’t appear to be the case.

When I talked to a Tesla Ranger about batteriesa while ago I noticed a strong and reserved reaction, the kind you’d expect from an organization that’s highly concerned with being sued. Their reservation seems to counterbalance the contempt you and others have. I can understand your ask for transparency but that ask varies wildly per individual. While you’d be happy with a t-shirt others are fine with the software solution and others want a brand new car. This ambiguity poses a great challenge for organizations, regardless of the fairness of your ask.

My original point I tried to make was aimed at accepting the risks of owning an innovation for a car. Given their mission it is to be expected that they’ll continue to dissapoint as they’ll focus on producing more instead of better vehicles. Some people here appear to have very unrealistic expectations of a company that tries to electrify the car industry. The goal of Tesla is to reduce your carbon footprint, not to make you happy.

Bring on the dislikes

No dislike in this quarter. My T-shirt comment was intended as sarcasm or as a sort of veiled suggestion that Tesla make nice with us customers. Perhaps something got lost as my message crossed the pond!

I agree that not all customers will view any olive branch extended by Tesla as wholly satisfying. There will be those that litigate; others will grouse and complain; while others will insist that Tesla is a company run by ex-felons looking to hoodwink the public.

However, I disagree about your conclusion from your discussion with the Ranger. I think his reservations and reluctance to share information with you has more to do with Corporate's tight control over their intellectual property. Rangers stateside have told me that anything that has to do with software or the traction battery/BMS and drive train is handled by Fremont. Several layers of supervisors need to sign off on major warranty or out-of-warranty repairs, and they are not done locally. Tesla does not want anything leaking out to the public. I think it has much less to do with being sued and more to do with the fact that the Rangers are almost as in the dark as we are. They have been admonished not to say anything to us customers, so they grope about to try to answer your questions sincerely without saying outright, "Corporate has told us to keep our lips zipped."

I also believe that we early owners knew we were accepting risks when we bought these cars. We expected things to go wrong. Early drive trains were defective after 10-25,000 miles. That said, Tesla never told us we were among a group of tens of thousands beta testers for their innovative product. (I did not know what a beta tester was until I joined this forum!) If we are going to shell out $100,000 to be beta testers, then Tesla had a choice to make: Either disclose up front that we would be subject to Tesla's research and data mining subject to their warranty terms, or to remain silent and let the chips fall where they may. The odds are quite good that if they opted for disclosure, sales would have been much lower in the early years. With my ignorance of these things, I would have delayed my purchase. So, they chose door #2. The chips fell, and Tesla got caught. We weren't buying a $950 Tesla smartphone!

Lastly, Tesla's mission statement is just a pithy turn of a phrase meant to appeal to younger and more liberal consumers who want Mother Earth to be in more of an equilibrium. Musk is a dyed-in-the-wool capitalist and libertarian who is pursuing celebrity and wealth. He neither cares about reducing our C footprint nor anything else as relating to climate change as much as he strives for fame. That Tesla's cars are exceedingly fun to drive and less stressful on the planet is coincidental, in my opinion.

 

[omarsultan]

Have seen zero arguments for this being fire related that could be considered "strong"... everything about it in this thread is either raw speculation, links to unrelated information presented as if it were related, and other hand waving. Nothing substantial, and nothing whatsoever that even remotely ties the range loss issue with fires.

Now you can have the conversation about causation vs correlation, but I don't think that is completely fair--the statement from Tesla in May, in response to the Hong Kong fires created the linkage between fire safety and SW updates:

As we continue our investigation of the root cause, out of an abundance of caution, we are revising charge and thermal management settings on Model S and Model X vehicles via an over-the-air software update that will begin rolling out today, to help further protect the battery and improve battery longevity.

At the same time, Tesla Service was using the same language ("improving battery longevity")when explaining the changes that resulted from 2019.16.x and beyond.

And, again, Elon could have used one of his hundreds (thousands?) of tweets he's made since summer 2019 to completely suck the oxygen out of the room for this thread.

 

[JRP3]

Tesla never told us we were among a group of tens of thousands beta testers for their innovative product.

I'd say it's fairly well known that early versions of new products are likely to have more issues. This was a new product in an entirely new market segment.

 

[wk057]

Don't get me wrong. I sincerely wish Tesla would have said something about this that made sense.

At the same time, I'm pretty sure we all know why they never did, and likely never will.

 

[wk057]

As we continue our investigation of the root cause, out of an abundance of caution, we are revising charge and thermal management settings on Model S and Model X vehicles via an over-the-air software update that will begin rolling out today, to help further protect the battery and improve battery longevity.

I want to point out that they're not lying here. They did update thermal management settings, and related charging settings. They tweaked the allowed thermal deltas a bit tighter, and they tweaked the thermal vs SoC table a bit tighter, both towards the side of what could be considered safer/cautious. The amount of change was minimal, though, and not likely to cause any real world improvements. The only thing this seemed to do in the wild was make the thermal system work a bit harder than it needed to (more pumping, more fans, more chiller, more heating, etc), and shifted the full-charge-power mark towards a higher temperature by a few degrees with the table change... which would result in slightly less regen at colder temps, and slightly slower charging (super or otherwise) if the pack was colder.

This update wasn't widely available for quite some time.

Keep in mind Tesla's versioning system. YYYY.WW.xx where Y is the year, W is the week. 2019.16 was developed and hit QA on the 16th week of the year, which would be April. Tesla announced their "abundance of caution" update in the middle of May... which would put the version at 2019.20-ish, which would generally hit the public a few weeks later... and according to the firmware trackers did exactly that. Same pace/delay as 2019.16

Which version first included the thermal and charge management tweaks noted? If you're been playing along at home you may have guessed it: 2019.20.

Just basic knowledge of the firmware release cycle and version numbering is enough to separate these issues.

 

[AustinP]

@wk057, spending too much time around TMC, seeing stories of cars shutting down at up to 17% (iirc), I started accepting that older batteries like our 85 are a risk to drive below 20%.

Reading your explanation about Z, I’m inclined to believe that the 0% is again correct and the risk of the battery shutting down when above 0 (*) is now mitigated.

Is this correct?

(*) and driving normally.