Wiki Sudden Loss Of Range With 2019.16.x Software

[Florian500]

The Linked thread is about the TeslaLogger but gets a bit off topic from time to time.
TeslaLogger can log a lot of CAN data with ScanMyTesla, see Open Source Teslalogger on Raspberry / Docker with ScanMyTesla integration

Around the linked post in the German forum cell voltage differences are discussed. There are some graphs and conditions under which they were recorded.

Let's try to get some content over here if it helps this topic.
And I asked in the German TFF forum to cross post plots and conditions here.

 

[Battpower]

The Linked thread is about the TeslaLogger but gets a bit off topic from time to time.
TeslaLogger can log a lot of CAN data with ScanMyTesla, see Open Source Teslalogger on Raspberry / Docker with ScanMyTesla integration

Around the linked post in the German forum cell voltage differences are discussed. There are some graphs and conditions under which they were recorded.

Let's try to get some content over here if it helps this topic.
And I asked in the German TFF forum to cross post plots and conditions here.

Yes. Tesla Logger and SMT is perfect for this. I am setting up for my car but not had time yet.

 

[First EV]

View attachment 572250

This is a screen from CanZE for my 2017 Zoe 40kwh.

My dad has similar model, 2015, 22kwh and his health shows 72%.

Seems like this kind of info should be standard on every EV and accessible without buying 3rd party tools.

Imagine the reduction in resale for used model S cars if there was a screen that showed battery health vs supposed range left on a charge like we currently have. I think people would be willing to pay a lot less if the battery health was visible and showed something like 62%. I think it could also reduce resale as a whole because the decreased life would be more obvious.

 

[Battpower]

So here, with no context, we see a car charging from 18:00 hours to 18:30 up to 80%. During charging, voltage difference (labelled as 'balance') sits at around 38mV.
Then car drives for around an hour discharging to 45%. Heavy loading around 19:10 shows huge spike in voltage difference at 800mV, but this is not actually showing imbalance of stored energy I suggest. Just difference in voltage.

We have to be careful interpreting numbers, especially when labels ('imbalance') could be misleading.

 

[Florian500]

"Cell imbalance" source:
SELECT $__time(datum), val as 'Cell imbalance [mV]'
FROM can
WHERE $__timeFilter(datum) and id = 27
ORDER BY datum ASC

There is not calculation / data transformation in Grafana, either this comes directly from the CAN bus or the transformation happens in SMT.
@bassmaster any ideas?

 

[Battpower]

Imagine the reduction in resale for used model S cars if there was a screen that showed battery health vs supposed range left on a charge like we currently have. I think people would be willing to pay a lot less if the battery health was visible and showed something like 62%. I think it could also reduce resale as a whole because the decreased life would be more obvious.

It cuts both ways. Who really thought they could get a 'machine' in which degradation is not effected by how hard it is used? I would see it as a big bonus to be able to show that my battery is in great condition and worth more. And for all cases, you surely need to be able to identify warranty issues where the car didn't deliver what could reasonably be expected / was guaranteed by warranty.

In any case, seeing a battery condition indication reminds you that although you may not be paying for gas, there are no free lunches.

 

[bassmaster]

View attachment 572265

So here, with no context, we see a car charging from 18:00 hours to 18:30 up to 80%. During charging, voltage difference (labelled as 'balance') sits at around 38mV.
Then car drives for around an hour discharging to 45%. Heavy loading around 19:10 shows huge spike in voltage difference at 800mV, but this is not actually showing imbalance of stored energy I suggest. Just difference in voltage.

We have to be careful interpreting numbers, especially when labels ('imbalance') could be misleading.

Cell Imbalance in Teslalogger is exactly the reading in ScanMyTesla. You can double check that by pedal to the metal and check readings of ScanMyTesla.

 

[Battpower]

Cell Imbalance in Teslalogger is exactly the reading in ScanMyTesla. You can double check that by pedal to the metal and check readings of ScanMyTesla.

The point I would make is that you can certainly display 'voltage difference' (say between highest and lowest brick) but this is not the same as charge difference, which is the point @wk057 was making I believe when he referred to 'charge counting'.

So all I am saying is that we need to be careful when we talk about 'imbalance'. We need to be clear 'imbalance of what?' In this thread at the moment, we are more interested in an imbalance of stored energy between bricks. When you use energy driving, the energy comes (ideally) fairly evenly from all bricks, so a brick with less energy stored will run out of energy first. Of course, this manifests itself by the brick voltage dropping to the minimum permitted level, but it is the brick having less stored energy than the others that causes its voltage to drop first.

In ideal / simple scenario, stored charge / energy is result of current flow for a specific period of time.

Anyone know about simultaneous differential equations?

One point of confusion could stem from voltage only indicating charge level when nothing else is happening and everything has had time to settle down. Every single cell in the battery is in parallel with other cells. As they deteriorate they may absorb and release charge at slightly different rates from eachother. The cells in parallel take a finite time for the charge to balance between them even though they all have the same voltage (within each brick).

 

[Battpower]

@Florian500, what would be interesting to find is a plot of invidual brick voltages during the charging phase and under heavyload (like 19:10 in the above chart). In fact, a multi-trace plot showing multiple brick voltages (to compare good and bad bricks) would be awesome. You could get a clue about which bricks to look at based on their voltages just at the end of charging. IE: The brick that reaches max voltage first could be a 'lazy' one.

Oh, and I havent said 'speculation' for a while - so "all speculation"! Hopefully with a lot of thinking.

 

[First EV]

It cuts both ways. Who really thought they could get a 'machine' in which degradation is not effected by how hard it is used? I would see it as a big bonus to be able to show that my battery is in great condition and worth more. And for all cases, you surely need to be able to identify warranty issues where the car didn't deliver what could reasonably be expected / was guaranteed by warranty.

In any case, seeing a battery condition indication reminds you that although you may not be paying for gas, there are no free lunches.

I agree it would be great to show your car is in great shape but also it could show it's not. I went without paying for gas, it was great until the charge rate slowed way down. Part of having a Tesla for many people is the quick acceleration. Not using the quick acceleration to avoid huge near term charge speed reductions and/or medium range reductions would defeat part of the purpose of having the car for those who enjoy the acceleration.

 

[Battpower]

Lots of cell voltage plots here: Teslalogger mit Raspberry Pi & MySql & Grafana & OSM

The Linked thread is about the TeslaLogger but gets a bit off topic from time to time.
TeslaLogger can log a lot of CAN data with ScanMyTesla, see Open Source Teslalogger on Raspberry / Docker with ScanMyTesla integration

Around the linked post in the German forum cell voltage differences are discussed. There are some graphs and conditions under which they were recorded.

Having had a look at the German site and (yet) another scan through the TeslaLogger thread as well as thinking about SMT (which I do have set up) it appears that the term 'imbalance' is being used too loosely and I suspect is at the root of quite a lot of flawed conjecture.

SMT does appear to be giving a voltage reading, which in some of the German examples clearly fluctuates in line with power being drawn from the battery. This would be more an indication of the apparent / effective internal resistance of each brick I think, and the idea being explored is that this is not the same / directly linked to the 'cell condition' with regards to energy storage.

 

[Battpower]

Not using the quick acceleration to avoid huge near term charge speed reductions and/or medium range reductions would defeat part of the purpose of having the car for those who enjoy the acceleration.

Those owners had better hope @wk057 replies to this thread proving my conjecture wrong and making me look a complete numpty. (Which I'm fine with! although I would feel really bad for wasting everyone's time)

Quite likely the better (closer matched in every way) you make the cells, the better they will stay matched, and so they will last longer and handle heavy use better. You know, 'million mile battery', battery day and all that stuff. That post a few back of mine and my Dad's Renault battery condition is interesting because both cars have done about 30k miles. So while mine (2017) is at 94%, his (2015) is at 72%.

Key points made by the Renault sales guy (I guess they train their sales staff differently from Tesla) with the first Zoe I bought that you make your battery last by:

Not fast charging. (Just charge at home 7kw)
Keep the power meter in the green area as much as possible while driving.
Don't let the battery SOC get too low.

and (although I don't know if this was correct)

Be sure to let the car charge to 100% indicated charge so the battery balances. It may be correct as their approach to balancing could be quite different from Tesla's.

Sorry to repeat, but once more people like @wk057 have confirmed (maybe the best we will hear is silence) or refuted this (imo somewhat obvious and inevitable) 'old news', then it will become clear how essential the warranty angle is. What's the point of 'paying to unlock extra acceleration' if it is at the expense of battery life (as well as drive shafts etc)? Any one who has a voice or route to EM / battery day Q & A (if there is such a thing) should certainly ask something about 'How does Tesla plan on guaranteeing long-term performance of batteries and what measure of degradation would be most appropriate?'

And, if by a miracle someone does have a line to EM, ask what his plans are for recycling / repurposing all the cells that come back from Tesla vehicles? Will they just end up being scrapped? How do owners get back the remaining value from their packs just because one or two weak bricks rendered the whole pack unusable? (Of course, the whole pack may have done a decent amount of work in its life. But not in every case I'm sure.)

 

[Battpower]

Anyone fancy starting up a third party battery refurbishment business?!

 

[Chaserr]

@Chaserr, if you are still tuned in, as the picture has (for me at least) clarified to a point that I feel I have a reasonable handle on most (though probably not all) that's going on, it does look more like 'life with lithium batteries' rather than some specific dangerous situation that took everyone by surprise. With gasoline, how long did it take for issues and compromises to work their way through the system? And even knowing far more and further reaching ones today, most cars driving round are still burning oil based fuels.

@wk057 believes it is a safety problem that needs to be recalled.

Intentionally limiting charge, removing things we paid for to save warranty cash and hiding short circuits (an obvious known danger) from us and the NHTSA is illegal on many levels.

This might be normal for Tesla but it is not normal for electric cars. We are covered by safety law if they need to remove what we paid for for safety, and they are dangerous for hiding it from us. We are protected by safety law because they can't reduce performance to hide a recall.

 

[feif]

I was asked to upload this, max "cited-maybe-inaccurate wording" Cell
Imbalance when flooring with 83% SOC and 50 C battery temperature right after SUCing on onramp to Autobahn
Cheers
Feif

 

[Battpower]

But on some cars the tanks do shrink. You can look at the Prius where they used a bladder inside the tank, in some cases the shrinkage reduced the capacity by almost 50%. (And someone started a class action which the judge threw out because it wasn't a "defect", it was a "design decision" to use a material that shrunk over time and when exposed to cold.)

On reflection, this was more relevant than I appreciated. If the bladder idea was an intrinsic design solution to a problem, (just as the Tesla battery system is) then it will of course have certain characteristics that may eventually reveal themselves to be undesireable.

Gasoline is highly flammable verging on explosive.
Hydrogen is explosive in air.
You can drown in water.
Lithium burns.

So far the evidence I see suggests:
Nothing that Tesla did (now I have thought it through) was for the purpose of depriving owners of anything.
Neither did they suddenly discover some new characteristic that they had to manage the consequences of.

All they did was carry on following the same processes that they had been doing all along.

We know what Tesla is worst at. Communication.

 

[Chaserr]

Nothing that Tesla did (now I have thought it through) was for the purpose of depriving owners of anything.

Tesla has already gone on the record admitting they deprived us. The purpose wasn't theft - theft was a side effect. They have avoided saying anything about the purpose because what they did is still illegal.

If they didn't have to do anything, they can revert us to before, return what was stolen, and everything is OK and the crimes can be forgiven.

If they had to do it to save warranty money, it's illegal.

If they had to do it for safety it's illegal.

They didn't have to do anything. What they did is illegal, and doing nothing would have been best. Undoing it would have been the next best option. They are in a profit seeking mode, so that option was probably far too expensive.

Leafs have normal degradation, it's accepted. That is different that Teslas in warranty.

 

[Battpower]

I was asked to upload this, max "cited-maybe-inaccurate wording" Cell
Imbalance when flooring with 83% SOC and 50 C battery temperature right after SUCing on onramp to Autobahn
Cheers
Feif
View attachment 572289

Thank you.

Is that your own car? What we considerring balance of energy / charge stored in the cell bricks. We are interested in being certain of the significance of the word 'balance' in SMT. It appears to be voltage balance rather than charge / energy balance. Your graph appears to show a close correlation between voltage difference and power, which is probably normal.

 

[feif]

yes this is my p85d which power fades after 15 sec flooring with insane+ ready

 

[Battpower]

If they didn't have to do anything, they can revert us to before, return what was stolen, and everything is OK and the crimes can be forgiven.

I (perhaps hastily) have posted similar to this myself. I can kind of agree that from an owner perspective it feels like theft. No car owner should have to wade through all this stuff to try and understand if something is wrong with their car. And without straight forward warranty terms and evidence to support battery condition claim, owners seem left no choice.

[edit: I think an argument of mis-selling might make more sense on the grounds that you were sold something that proved to be unable to meet the claims made of it.]

If they had to do it to save warranty money, it's illegal.

This case seems the most plausible argument to me at the moment. It is at least a happy coincidence for Tesla that while doing the right thing to look after batteries and keep (hopefully) the vast majority of owners fairly happy with their cars, they also stand to save themselves some money on warranty claims. Not sure which is chicken and which is egg.

If they had to do it for safety it's illegal.

This might come under a similar argument to Warranty i suppose, in that the actions they took could possibly be taken as (unavoidably / coincidentally) making the car safer, but I don't think that means it was particularly dangerous or faulty in the first place, except as in being able to drown in water or make gasoline explode.