Balancing the battery

[stopcrazypp]

Stopcracypp, with great respect, that book is copyright 2010. Lithium chemistry has changed A LOT since then. Specifically, the Panasonic 18650 wasn't yet on the market when this book was written. The first press releases for the version of 18650 chemistry used in Tesla cars were coming out about the time that book went to press. Having said that, all the discussion about "Top" v "Bottom" balancing still hold, to the best of my knowledge, with 18650s.

Chemistry has changed but fundamentals have not. For the Panasonic cells, 3.6V nominal and 4.2V max has still not changed (although future chemistries are poised to change that). While the exact capacity/variant cells Tesla used was not existent in 2010, a lower capacity version (2900mAh NCR18650) had been in commercial use since 2006:
http://news.panasonic.com/press/news/official.data/data.dir/en091218-2/en091218-2.html

How much "mid" vs "top" balancing is done by the Tesla BMS? Doing some 'mid' balancing allows lower currents in any given BMS, and 'top' balancing can still be performed by the BMS (for all the reasons so well articulated in that book), even after some 'mid' balancing has occurred. I mention "...allows lower currents..." because one thing we know as a physical fact from dissected packs is that the wires and circuit board traces are physically small and could not source/sink very much 'bleed' or 'transfer' current.

At the risk of getting further chastised for "speculating": I imagine a bulk of it is mid balancing and maintenance balancing. That is sufficient for daily use. However, you do notice the caveat that the pack must start out top balanced and must have a very accurate SOC to Voltage map and/or SOC history. As the pack ages however, that mapping is going to become less accurate and a top balance will always remain as the most accurate option.

- How much 'rest' time is allowed before 'top' balancing, when the car sits plugged in for hours after a range charge? In plainer language, how much of this is actually balancing and how much measuring?

Just resting a pack at a high SOC will not help you measure the pack capacity better. It's a large SOC delta that does that (the discharge to lower SOC and charge to high SOC). Even the explanation from a Tesla rep explained that point. Therefore, I don't really buy the claim that the higher numbers from a pack plugged in at 100% charge is primarily from better measurements. Given that 100% SOC is a prime SOC to do balancing (esp. with the car plugged in so there's no concern about draining the battery) I have a hard time believing the Model S's BMS would not take advantage of that (esp. given we know the Roadster's DID).

So while there is no direct evidence from Tesla that balancing happens at 100% SOC, there is no direct evidence that no balancing happens at 100% SOC either. And all the facts we have so far point to the probability of the former being higher than the latter.

 

[J1mbo]

So, does anyone have info on when the Model S does its cell ballancing?

1) at the end of the charge, regardless of SOC target?
2) only at the end of a 100% charge?
3) after the vehicle is shut off or put into sleep mode?
4) while driving down the road?
5) some other time?
6) all of the above?
7) none of the above?

GSP

My guess: all the time that the HV contactors are open. It is so slow (based on the hardware found in the teardown) that there would be no benefit of having the hardware there in any other scenario. No point in trying to balance when under load (i.e. when driving).

 

[WarpedOne]

Hmm, balancing when off... could this be the famous vampire sucking away some part of that kWh lost over night?
Would also explain why some experience hungrier vampires than others.

 

[Danal]

I stand corrected... the 18650 was certainly on the market. My original post was extremely unclear in the way I phrased things. What I meant to say was that the 18650B was not on the market when that book was written. And, I fully agree that the 4.2 volt spec for charging did not change; however, charge current and charge cutoff current both did change with the "B". And they didn't just change in absolute numbers (which can be misleading since the capacity of the cells is different), they changed in terms of "C" ratio as well. Even small changes to the charge algorithm are very relevant to the design currents of a battery management system. Furthermore, the Gravimetric energy density changed quite a bit between the 18650 and the 18650B at 220 and 243 Wh/Kg respectively. That's a little over 10% change. This is not particularly relevant to a BMS, it is just a further indication of much these cells have evolved. All this is from the data sheets on Panasonic's site.

Having said all of that, I fully agree that most industry sources, and most commercially available Battery Management Systems, encourage "top" balancing. For example, the Orion BMS (which is one of the more flexible for EV use) is configurable and recommends that the "Start Balancing" voltage be set within 5 to 10% of the maximum (fully charged) voltage. However... it is interesting to note that the Orion (and most other BMSes) allow configuration to many other balance algorithms, including "mid" and mixtures. So there is still enough debate in the industry today for manufacturers to include multiple balance options in their firmware.

Interesting thought: Since a given Tesla may be charged (by driver setting) to xx%, and not charged to 100% for months at a time... the BMS will ultimately have no choice but to balance at non-100% state-of-charge.

And... We still don't know what Tesla is really doing. Their dissected packs have unlabeled chips in the BMS. We can also speculate that they may be doing something ?better? than an "off-the-shelf" BMS. Maybe.



Speculation on all of this is fun... but it seems this would be very easy to nail down with a few measurements. Except that the packs are more-or-less sealed, and that pesky "no reverse engineering" document. Oh well... Perhaps the next person to get a salvage pack could partially open it, and charge it with the salvage cars systems a few times, taking measurements as they go.

 

[AmpedRealtor]

Sure they do in many threads, especially if they are passing on "conventionally accepted (even if not proven) wisdom" when commenting. Myself included.

However, in this specific thread, the OP asked:



Hence you'll notice responses are being qualified as "anecdotal experience", "it's conjecture", or "speculation" & "assumption"

When you come along and make a direct assertion that contradicts the admitted conjecture here in this thread when the OP specifically asked if people are speculating, then you should probably expect people to question what authority provided the info.

Whatever. I'm just saying that anyone who thinks the car is balancing at 100% is not basing that opinion on any official citations from Tesla. That's my point. This whole fable about the car balancing at 100% because the car sits for an hour at 1 Amp is nothing more than a myth because there are no citations to support that view. It started as opinion based upon how the Roadster worked, yet no real evidence. That's all I'm saying. Many people bought into a rumor that cannot be proven, and it's far more reasonable to see the behavior for what it most likely is - the very end of the taper curve. All information regarding the taper curve would support this observed behavior.

We can nitpick and split hairs all we want, but the basic fact remains that anyone who says the car balances at 100% has no facts to prove that assertion.

 

[Denarius]

Whatever. I'm just saying that anyone who thinks the car is balancing at 100% is not basing that opinion on any official citations from Tesla. That's my point. This whole fable about the car balancing at 100% because the car sits for an hour at 1 Amp is nothing more than a myth because there are no citations to support that view. It started as opinion based upon how the Roadster worked, yet no real evidence. That's all I'm saying. Many people bought into a rumor that cannot be proven, and it's far more reasonable to see the behavior for what it most likely is - the very end of the taper curve. All information regarding the taper curve would support this observed behavior.

We can nitpick and split hairs all we want, but the basic fact remains that anyone who says the car balances at 100% has no facts to prove that assertion.

If sitting at 1 amp was part of the taper curve it would do it every single time, but it doesn't. Sometimes the car finishes right away, and sometimes it takes over an hour. Therefore it is a fact that *something* is happening and not every time. It is also a fact that the Model S battery pack has hardware in it to balance. It is also a fact that it is easiest to balance batteries in general while at 100%. It is a fact that the roadster battery, also designed by Tesla balances at 100%.

Logic, based off of the Roadsters behavior, suggests that balancing is what causes that extra time. That conclusion is the result of empirical evidence. While your assertion that it is a myth is based simply on the lack of a statement by Tesla.

 

[luster]

my total guess and conjecture would be that the pack is constantly balancing (albeit very very slowly).

I honestly doubt any intervention (start charge) or charging trigger (charge reaches xx%) would begin the process.

With the likely current of the BMS, it wouldn't make too much sense for it to not be continually balancing. At least until the point that the pack is in fact balanced.

To balance at 100% would be counter-intuitive as we would be trading off a balanced pack with degradation of the pack.

 

[Denarius]

my total guess and conjecture would be that the pack is constantly balancing (albeit very very slowly).

I honestly doubt any intervention (start charge) or charging trigger (charge reaches xx%) would begin the process.

With the likely current of the BMS, it wouldn't make too much sense for it to not be continually balancing. At least until the point that the pack is in fact balanced.

To balance at 100% would be counter-intuitive as we would be trading off a balanced pack with degradation of the pack.

I don't believe the pack only balances at 100%, it is my opinion that any charge 90% or higher will trigger a balance.

 

[scaesare]

Whatever.

Excellent retort.

I'm just saying that anyone who thinks the car is balancing at 100% is not basing that opinion on any official citations from Tesla. That's my point. This whole fable about the car balancing at 100% because the car sits for an hour at 1 Amp is nothing more than a myth because there are no citations to support that view. It started as opinion based upon how the Roadster worked, yet no real evidence. That's all I'm saying. Many people bought into a rumor that cannot be proven, and it's far more reasonable to see the behavior for what it most likely is - the very end of the taper curve. All information regarding the taper curve would support this observed behavior.

We can nitpick and split hairs all we want, but the basic fact remains that anyone who says the car balances at 100% has no facts to prove that assertion.

And I was simply asking for clarification if your statement was based on any more authority than the posts you consider to be baseless opinion. It wasn't.

Actually as other posts have pointed out, other similar systems (the roadster), do behave this way when balancing, and many whitepapers & BMS implementations do similarly. So there may indeed be at least precedent for the conjecture that the Model S does similarly.

Determining which opinions (for or against) hold more weight is left as an exercise for the reader.

 

[Todd Burch]

Here's a more important question: unless balancing really is completely automatic, done in the background, and performed regardless of whether one range charges or not....why has Tesla told us nothing about this, officially?

If a full 100% charge initiates balancing, there should at least be a blurb about it in the manual--and an indication of how often it should be done.

I understand that Tesla may be trying to "keep it simple" for the masses, but that doesn't mean withholding information about the proper maintenance of our cars when we request it. If balancing is only done at 100% charge, that means someone who always charges to 90% never balances their battery--so they end up having less effective available capacity. If the battery needs to be range charged for x amount of time every y weeks to stay balance, then tell us that.

I don't understand why they are officially silent on the issue.

 

[stopcrazypp]

If balancing is only done at 100% charge, that means someone who always charges to 90% never balances their battery--so they end up having less effective available capacity. If the battery needs to be range charged for x amount of time every y weeks to stay balance, then tell us that.

I'm getting confused now. I don't think anyone is claiming balancing *only* happens at 100% charge. That would not be consistent with Roadster behavior either (which balances also under standard charge which is ~87%SOC).

The only claim being made is that balancing *does* happen at 100% and that the balancing that happens during this time is the most effective (something that will bring an extremely out of balance pack back into balance quickly). In other words, similar behavior to the Roadster. And it is that balancing at 100% that will squeeze that last couple of miles of range out of the battery (as an explanation of why sitting a long time after range charge tends to yield higher numbers).

For daily use, a less accurate maintenance balance or mid-balance is all the regular consumer needs. Even something like 1% out of balance is less than 3 miles of rated range (and a maintenance or mid-balance likely can get to that kind of accuracy). No need at all to mention it in the manual as it only means a few miles of rated range off the battery (plus it conflicts with the advice about not letting the battery sit too long at high SOC). However, there are people out there that stress out about not seeing that last few miles of rated range, which is where this topic came up.

 

[dpeilow]

I was told a couple of weeks ago by the head of service in the UK, who has been servicing Roadsters since their launch, that to balance my pack I need to fully standard mode charge the car leave it a couple of weeks (e.g. next time I might be away).