2017 M3 left for 2 years! what steps should I Take?

[Tronguy]

Lead acid batteries is very sensitive.
They start to degrade when discharged to 50%, just like the LV 12V lead acid batteries used earlier in teslas.
They sulfate from low SOC, and should not be discharged below 50%, and preferably they are held at 100% all the time.

Its a completely other battery with very few common things to lithium ion

Good; and that explains that. At the big disaster mentioned above, which took down Wall Street for some time, there was gear drawing substantial load at 35V or so and some equipment still trying to run at 25V. On a 48V battery plant.

Yeah: 2' in diameter, glass walls, 5' high is one of those; and the plants I've been in had those stacked double high, then rows and rows of them, filling up a good-sized room. With another good-sized room Somewhere Else with the always-on Standby plant. The bus bars were copper, something like 6" square, and we're talking thousands of amperes. Fun. One does not drop a wrench across said bus bars.

The original PUB (it's been madly modified over the years) called for equipment to go to zero load at 39+-1V going down, and coming back to life at 42.5V+-0.5 coming back up. One of my fun jobs over the decades was testing just that.

What's even more fun is making sure that the multi-shelf system software itself can recover from these events. Not a whole lot of fun when some subsidiary electronics goes down due to power loss like this and the upstream control software thinks that it's a permanent fault. You guys don't want to know the number of times I've had to take a group of newbie CS majors and explain the Facts of Life to them. ("But.. but.. computers are always on, right?" "Um. See these GPIO pins right here? The ones that have a pull-up on them? When it's Zero, it's good, but when it's One it means, POWER_FAULT?" "But.. computers are always on, right?" Sigh.)

(In their defense, the time constants for a POWER_FAULT might be measured in the 50's of milliseconds, which is Forever to a CPU that's chugging along at 10ns per instruction or something. So it's kind of critical when a fuse blows to try and get that POWER_FAULT signal out there before VCC decays to the point where the remote equipment isn't working. Ulcers.)

 

[AAKEE]

But Teslas reserve about 5% of the pack at the bottom for this very case.

No, it is not like that.

The 4.5% buffer is the part of the battery below 0% on the screen.
But these 4.5% can be used to drive on.

The shut down level is most certainly below most of these 4.5%, and the shut down limit is 2.5V/cell, which would mean 240V on a 3/Y pack.
We can see the BMS stating min pack voltage 241V on my former M3P:

I have driven that car down to -1.8% twice, for some tests/calibration purposes.
While I haven’t driven the car until it stops, I’m certain that the BMS shut down no later than the first cell group reaches the 2.5V min voltage.
This gives room for protection.

My guess is (but I do not know) that the BMS uses very very little energy, or none when the Contactors open because of low voltage limit is reached.

My complete MSP and the former M3P did use around 4W when parked longer time. This includes waking up amd charging the LV Batt and also keeping the internet connectivity running etc.
Innmy world the consumption insude the battery should be very low when disconnected and the car is dead with the LV batt drained.

 

[brkaus]

All L2 charging is slow at about 1/10 C. There's no difference between slow and really, really slow.

Oh, I agree. My thought (perhaps incorrect) was the super duper slow might allow the balancing to keep up.

Not really a battery cell protection thing.

 

[gearchruncher]

My thought (perhaps incorrect) was the super duper slow might allow the balancing to keep up.

No system I have ever heard of attempts to balance until the first cell is at max voltage. There's really no point, and voltage is an imperfect measure of the state of the cell until you hit the max voltage.

 

[gearchruncher]

No, it is not like that.

The 4.5% buffer is the part of the battery below 0% on the screen.
But these 4.5% can be used to drive on.

Yeah, that was my point. Unless this car was parked with less than 0% indicated, it had at least 5% of the rated battery left to keep it healthy when it sat idle, and likely much more. I don't see anywhere where I claimed you couldn't drive on this buffer.

 

[AAKEE]

Yeah, that was my point. Unless this car was parked with less than 0% indicated, it had at least 5% of the rated battery left to keep it healthy when it sat idle, and likely much more. I don't see anywhere where I claimed you couldn't drive on this buffer.

What i mean is that the day the car did shut down (last comms with Teslas server, 10/23 2022) the battery was drained to the level where it shuts down, probably close to 0% or 2.5V in the lowest cell.
So, the battery did not have 4.5% at that time.

Still, self discharge is very very low.

 

[Zuikkis]

No system I have ever heard of attempts to balance until the first cell is at max voltage. There's really no point, and voltage is an imperfect measure of the state of the cell until you hit the max voltage.

Tesla does balance at every voltage level. It calculates a capacity (Ah) value for each cell group. Then it knows if a certain cell is "overcharged" even if the voltage is lower than some other cell!

It's like every cell group has its own voltage target curve, for every SOC. For a given SOC, Tesla BMS knows what every cell voltage should be.

@AAKEE probably can give a more technical explanation.

 

[Ray Brown]

good morning all my testa mentors.
we are now at 191 miles and still charging

 

[Cosmacelf]

good morning all my testa mentors.
we are now at 191 miles and still charging

I take it you don’t have a faster charger there? No 240V option? It wouldn’t hurt to charge it at 240V, won’t make a difference to the battery.

 

[Ray Brown]

we have a 240 link but it's not accessible because the garage has two 2010 roadsters also parked in there. I am told those batteries will be toast. they are all in a vineyard in Napa valley that's why there's no wi-fi also

 

[Ray Brown]

this property nor these cars are mine except in my dreams

 

[gearchruncher]

Tesla does balance at every voltage level. It calculates a capacity (Ah) value for each cell group. Then it knows if a certain cell is "overcharged" even if the voltage is lower than some other cell!

If this is true, then why does the last 1%-5% of a charge slow down so much and take so long, and why is it so variant? If it was balanced all along, all the groups should arrive at full charge at the same time.

Just because it knows they are out of balance does not mean that it is balancing the whole time while it is charging. It would be particularly hard to balance during fast charges. and it's silly to balance when not charging as it just throws away energy.

 

[brkaus]

If this is true, then why does the last 1%-5% of a charge slow down so much and take so long, and why is it so variant? If it was balanced all along, all the groups should arrive at full charge at the same time.

Just because it knows they are out of balance does not mean that it is balancing the whole time while it is charging. It would be particularly hard to balance during fast charges. and it's silly to balance when not charging as it just throws away energy.

The only balancing available is relatively small bleed resistors. Just because it slows down at the end, it doesn’t mean it wasn’t trying all along.

 

[Ray Brown]

at 70%

 

[Ray Brown]

we moved the charge bar to 100% for 1st charge only

 

[RayK]

at 70%

That would imply that a full charge would have a range of 291 miles.

 

[mborkow]

we have a 240 link but it's not accessible because the garage has two 2010 roadsters also parked in there.

Two 2010 Roadsters! Nice problem to have

 

[AAKEE]

Tesla does balance at every voltage level.

Naah, it doesnt.

There has been rumors about changed logarithms, but for my 23 Plaid I se around 10mV at 50-55% all the time. Charging higher reduces the imbalance, as it is balanced higher up in SOC. Also leaving the car for a while at higher SOC reduces the imbalance = balancing happening.

Its very possible that the BMS can balance at lower SOC when needed, but I see nothing of it.

My M3P 2021 initially had the same CAC on all cell groups, so CAC min/max/avg was the same. The imbalance did most often be 4mV anyway around 30-55% SOC and at higher SOC it got a little more spread.
I tested slowly increasing 10% steps and waited two hours between each. No difference except when exceeding 90% (stayed at 92% from
Memory and there the cells balanced with decreasing imbalance. This was spring 2021.

To not go too deep here, i’m quite sure the pack newer balances after reaching the low level limit.
Teslas solution is burning of excess voltage with resistors. This would mean to readuce the higher cells to a lover value. What would the purpose be off this?

In general there is no need to balance cells if the spread is not too big.
Balancing cells at low SOC (or any other SOC than high SOC) would mean that you actually increase the balance at high SOC. Next full charge would then need more balancing to compensate forcthe balance at low SOC.

Balancing is purely a matter of filling up pr drainging cells to get them to the same voltage ( = same SOC).
In Teslas case, only draining the higher voltage cells is possible.

When ( after) top balancing all cells are full, and we have the maximum energy stored.
Also, the spread is pushed to the bottom which means more imbalance in bottom, but this will be hidden inside the buffer on a tesla.

Cells in the pack, or group of cells actually have differebt capacity.
This means that if we are top balanced like the lower part of the pic below, the smaller cells will loose voltage faster during the discharge. Cell (groups) are connected in series so the drain in amphs is the same.

If we balance at for example 60% like the upper part of the picture, we will have a spread at higher SOC where the smaller cell (groups) get fullt chargerd first, remdering the largest capacity cell (groups) with the lowest voltage.

Same at low SOC, the smaller cells will loose SOC/voltage faster and will be the ones that reach the low voltage limit first and thus setting the shut down point for the BMS of the pack.

it knows if a certain cell is "overcharged" even if the voltage is lower than some other cell!

Voltage = SOC.
The true SOC is measured when the battery is ”offline” contactors open, and Voltage is the thing.
A cell can not get overcharged witout having a too high voltage.

 

[AAKEE]

If this is true, then why does the last 1%-5% of a charge slow down so much and take so long, and why is it so variant? If it was balanced all along, all the groups should arrive at full charge at the same time.

The reason for a lithium battery to take long time/charge slow in the end of the charge, is that you can not/are not allowed to increase the cell voltage over the 100% SOC voltage.
For these and most other lithium batteries (not LFP) it is 4.200v/ cell.

Charging a lithium ion cell means using CCCV (Constant current- constant voltage)
You charge as you like on the CC part but at the CV part you only use that max voltage limit on the cell (and pack).

When charging it is the voltage difference between the battery or cells and the charging voltage that set the flow of current.
When you reach 4.20V per cell, the voltage of the cell increases which means that the voltage difference reduces continously and the current goes down. (The internal resistance of the cell or pack decides the current at a certain voltage difference.

In the end of the charge the cell reaches very close to 4.20V and the current will be very small.
Many chargers for charging lithium ion have a set currebt limit mile 1/20 (or 1/10 or so for fast charging) of the set current as the cutoff where the charger says ”charging finished”.

 

[AAKEE]

Side note:
I had to help my parents with a dead Automower (robot lawn mower) today.
Could not be started, did not take charge.

Most probably they forgot to shut it down and maybe even to get it fully charged before the winter. (Stowed away ~ 9 months ago). Battery did look “dead”.

Opened the robot, took out the battery (5 18650 in series). Each cell was 0.9-1.1V which is waaay below the 2.5V end of discharge for these cells.

I charged with one of my many hobby lithium chargers. Circumnavigated the BMS and charged right on the end if the cells. After filling up ~50% of the charge I put the robot together and started it, and charged in the charging station.
All good, filled up to 100% as as far as it looks it has not lost any capacity from last year.

Charging at a safe place, just in case it was anything wrong causing a fire.