Dumb**s alert: am I the first to drain it??

[Ingineer]

Actually in another post AndyH makes a good argument that the LiMn chemistry in the LEAF has a steeper curve throughout it's capacity and that balancing can take place anywhere along the curve.
http://www.mynissanleaf.com/viewtopic.php?f=31&t=2735&start=50

Balancing only takes place at the top, and can only happen at low current levels.

 

[JRP3]

That was my initial contention but I'm not as familiar with LiMn as I am with LiFePO4. You might want to chime in on the discussion which starts here: http://www.mynissanleaf.com/viewtopic.php?p=67333#p67333

 

[AndyH]

Actually in another post AndyH makes a good argument that the LiMn chemistry in the LEAF has a steeper curve throughout it's capacity and that balancing can take place anywhere along the curve.
http://www.mynissanleaf.com/viewtopic.php?f=31&t=2735&start=50

Balancing only takes place at the top, and can only happen at low current levels.

How do you/we know this?

As I've stated in other threads, I know that silicon and software exists to balance throughout the charge/discharge, and I know it's used, and I know it's best used with cells like LiCo and LiMn that have a slope charge/discharge curve. But I don't know how Nissan's balancing the Leaf pack.

But - decently matched cells in decent health don't need any balancing -- and if range is adequate and if failed cells are swapped, I can see how a pack could be built with zero balancing.

 

[ERG4ALL]

A friend of mine in Australia sent me an email from: http://www.smh.com.au/drive/motor-news/electric-cars-run-out-of-range-20110315-1bux4.html.

This is the essence of the article in my reply to him:

Hi Again,
Here's the quote as it appeared in the "MyNissanLeaf.com" forum: (http://www.mynissanleaf.com/viewtopic.php?f=27&t=2746.)

"Yes. I am out of power. http://yfrog.com/gz3b7nwj. Went from 17 to --- to turtle to dead in about 5 miles. 2.3 miles from dealer. 4.2 miles from home. Part of me is amused that I may go down in history as the first dumbass to drive the car into submission. But I am slightly shaky and upset as I thought there should have been no problem getting home.

Ah well my experience was less than a mile in turtle."

And here's the quote from the article:

"Part of me is amused that I may go down in history as the first dumbass to drive the car into submission," the owner wrote. "But I am slightly shaky and upset as I thought there should have been no problem getting home."

____________________________________________________________________________

I'm a little put off by someone selectively quoting from the forum without giving the whole story and many of the cogent comments that would help to explain it or at least to recognize that if it is something that needs fixing that it is an isolated incident.

The disparity of cost of traveling with electricity instead of oil is so great I'm starting to get paranoid that there are interests out there that will do whatever to discredit EVs. Now many people in Australia will disregard EVs as a possible transportation choice.

See my post on how can we get ICE drivers to adopt EVs.

 

[DaveEV]

Balancing only takes place at the top, and can only happen at low current levels.

How do you/we know this?

It appears that the Leaf is only able to shunt power (discharge) from cells which have higher voltage readings than other cells. Given that it doesn't use that power to charge up low voltage cells, it doesn't make sense to do any balancing except when the pack is near full and plugged in.

All speculation of course - but Ingineer has analyzed the service manual as you have.

 

[jcesare]

The disparity of cost of traveling with electricity instead of oil is so great I'm starting to get paranoid that there are interests out there that will do whatever to discredit EVs.

"Just because you're paranoid doesn't mean that they are not really out to get you."

 

[Ingineer]

Balancing only takes place at the top, and can only happen at low current levels.

How do you/we know this?

As I've stated in other threads, I know that silicon and software exists to balance throughout the charge/discharge, and I know it's used, and I know it's best used with cells like LiCo and LiMn that have a slope charge/discharge curve. But I don't know how Nissan's balancing the Leaf pack.

But - decently matched cells in decent health don't need any balancing -- and if range is adequate and if failed cells are swapped, I can see how a pack could be built with zero balancing.

The balancing system in the Leaf is basically a bypass MOSFET per cell that can be switched in to bypass small amounts of charge current, and according to the service manual it's triggered by voltage. In the Leaf, according to the service manual, this system is implemented in an ASIC (Application-Specific Integrated Circuit).

This is a classic top-balance system, and due to the inherent limitations of thermal dissipation in an ASIC, the balance current is never going to be able to exceed 1 watt, and will likely be well under that. That puts the maximum charge current at under 250ma, and based on readings people have seen near the end of charge, I'd say it's more like 100ma.

They are likely using something similar to the LTC6803, if you read it's datasheet you will better understand how top-balancing works. Note the similarity to the service manual diagram:

Because the voltage curve is relatively flat, there would be no way to (let alone any reason to) effect bypass balancing until the cells are near 100% SOC.

My take is the Leaf's cells are very well made and have consistency, so thus need very little balancing.

Here is an interesting article on Balancing that's worth a read.

-Phil

 

[AndyH]

The balancing system in the Leaf is basically a bypass MOSFET per cell that can be switched in to bypass small amounts of charge current, and according to the service manual it's triggered by voltage. In the Leaf, according to the service manual, this system is implemented in an ASIC (Application-Specific Integrated Circuit).

Yup - read that. [devil's advocate mode] But it doesn't say at what voltage... [/da]

Because the voltage curve is relatively flat, there would be no way to (let alone any reason to) effect bypass balancing until the cells are near 100% SOC.

Here's where we diverge. The LiMn curve is quite sloped - not unlike lead acid. I'm not saying Nissan DOES balance at an intermediate point, but the hardware they have is capable of doing it under computer control.

My take is the Leaf's cells are very well made and have consistency, so thus need very little balancing.

I agree 100% when the pack's new. Balancing won't be necessary until the pack starts to age - and it'll be more necessary over time.

Interesting balancing quote from a TI paper:
http://focus.ti.com/download/trng/d... Cell Balancing - What to Balance and How.pdf

Simplest algorithm is based on voltage difference between the cells. If difference exceeds predefined threshold, bypass is engaged. To resolve several problems of this simple method, more complicated modifications can be implemented if microcontroller is used to execute the algorithm:

• Balancing during charge only is used to save energy in portable applications.
• Balancing at high states of charge only is used to decrease the effect on SOC balancing that can come from impedance unbalance.
• Simultaneous multi-cell balancing makes decision on which cells have to by bypassed under considerations of the entire pack and not only neighboring cells as it is the case with comparator-based solutions

One of the advanced implementation of voltage-based algorithms using all above optimizations is used in bq2084 battery fuelgauge.

The LiMn2O4 curve provides voltage comparison throughout the charge/discharge process, and the computer control could easily pulse the FETs to bump the cells around as needed. It would reduce heating in the battery controller as well.

 

[Ingineer]

Ok, I'll see your Devil and raise you a Bypass! =)

The service manual makes mention of the term "Bypass" multiple times, which by definition can mean only during charge, as otherwise it becomes a "load" not a "bypass", correct?

It doesn't make sense when every watt-hour is at a premium, to deplete charge from the pack that's already been charged, so this is the rationale for charge-bypass rather than active depletion. The TI quote you posted "Balancing during charge only is used to save energy" also confirms this. Also, you'd be unnecessarily "exercising" the cell, which over a long period could reduce it's life.

Does it matter if the pack is unbalanced as long as all the cells are not near the ends? The only reason to balance the pack is to ensure the most usable capacity long-term. If you didn't have balancing then you'd be limited to only charge until the first cell hit top, and only discharge until the first one hit LVCO. If the pack is never pushed to 100% or LVCO, then minor balancing is a non-issue. If a major imbalance occurs, it's likely due to a fault, and the Batt ECU will set a code.

It's possible they could balance at any SOC while charging, but then they'd have to reduce the charge current to something the bypass FET could handle. A typical ASIC FET implementation is around 15 ohms, so this limits bypass current to about 250ma. If that is continuous dissipation, the package would quickly overheat, so in practice this would be lower. So slowing down the charge in the middle for no good reason when it could be done at taper?

-Phil

 

[DaveinOlyWA]

i think it does matter since on my Zenn and its lead acid batts, the imbalance multiplies over time. since the voltages are drawn from the pack and not by individual cell, that can cause weaker cells to be stressed first.

also ones that have lesser capacity will also charge up faster possibly shortening their lives with a greater potential to overcharge

 

[Ingineer]

i think it does matter since on my Zenn and its lead acid batts, the imbalance multiplies over time. since the voltages are drawn from the pack and not by individual cell, that can cause weaker cells to be stressed first.

also ones that have lesser capacity will also charge up faster possibly shortening their lives with a greater potential to overcharge

Comparing Lithium Ion to Lead isn't a good comparison, and the variability in lead cells can be quite high compared to even low-quality Lithium Ion. The Leaf uses high-quality cells with very accurate manufacturing so each batch is very well matched.

As I mentioned, as long as all the cells are kept "in the middle" (not near the ends) then balancing is not as much an issue. Since all cells see the same current draw, they all rise and fall together. The most common reason for Lithium Ion cell imbalance are manufacturing differences, which we don't really have in the Leaf. The second most common is self-discharge, which is a function of temperature. This means cells that are hotter will discharge more over time. This is not as much an issue in high-rate applications where the pack will not be subject to high delta-T over a long time, but if you look at something like a laptop pack this is a big deal.

If we balance at the top end of the charge, that will keep everything in line barring a failure.

-Phil

 

[AndyH]

Ok, I'll see your Devil and raise you a Bypass! =)

All right - I'll play your silly game! :lol:

It doesn't make sense when every watt-hour is at a premium, to deplete charge from the pack that's already been charged, so this is the rationale for charge-bypass rather than active depletion. The TI quote you posted "Balancing during charge only is used to save energy" also confirms this. Also, you'd be unnecessarily "exercising" the cell, which over a long period could reduce it's life.

Does it matter if the pack is unbalanced as long as all the cells are not near the ends? The only reason to balance the pack is to ensure the most usable capacity long-term. If you didn't have balancing then you'd be limited to only charge until the first cell hit top, and only discharge until the first one hit LVCO. If the pack is never pushed to 100% or LVCO, then minor balancing is a non-issue. If a major imbalance occurs, it's likely due to a fault, and the Batt ECU will set a code.

It's possible they could balance at any SOC while charging, but then they'd have to reduce the charge current to something the bypass FET could handle. A typical ASIC FET implementation is around 15 ohms, so this limits bypass current to about 250ma. If that is continuous dissipation, the package would quickly overheat, so in practice this would be lower. So slowing down the charge in the middle for no good reason when it could be done at taper?

-Phil

Hypothetically... It would be very possible to switch FETs in and out during charging to keep voltages somewhat close. We wouldn't need to reduce the charge current because we're still working in mid-charge. It's a horse race - we want them all to run full speed, we just want to tug on the front-runners once in awhile to keep the pack somewhat together. It'll generate heat, but we've got at least 8 hours to switch things in and out. We may have to reduce current near the end of charge if we need to continue balancing at the end, unless the draw is already at or below the 250mA max bypass.

Maybe early-on the car won't have to balance at all. Then down the road we'll need some top-balance time near the end of charge. Maybe later if we can't keep the pack in check without overheating the BMS we'll add some drag during the main charging period. Then we can throw-in the towel and flag the bad cell and take a nap.

Thanks Phil!

(man...is the car here yet? I need to take something apart!) :lol:

 

[tbleakne]

I made my statement:
"At the 100% level the voltage slope gives the car an accurate SOC value, but in the middle range the Leaf likely uses coulomb-counting. "
is based upon what Kelangst was told by Nissan:

feedback from Nissan:
a specialist from Los Angeles, and TWO engineers from Japan.
Bottom line: nothing is wrong with batteries, nothing is wrong with systems, nothing is wrong with car. I drove it until it had no power.

Nissan knows they have an issue with calculations of available distance and Human Interface issues under 2 bars of charge.
So my problem was that I first charged the car to only 80%, then without driving it I thought I might need more so I charged for another 45 minutes, then I drove it, then I charged for 70 minutes, then I drove it.... bottom line is that it had trouble given all of that and the mixed driving of both highway and city street.
>> a software update should be pushed to the cars in the US soon.
when you hit the last 2 bars, the calculation gets a little dicier.
I was at 2 bars at 17 miles left. The last 2 bars disappeared very quickly.

If the car was just using voltage to estimate SOC, why would this partial-charge history affect its estimate? Nissan's LiMn chemistry may well have more voltage slope than other chemistries, but I as well as others would like to see discharge curves for Nissan's actual technology.

Andyh, you have frequently expressed strong optimism about the performance and long life of the Leaf battery near extremes, and the sophistication of the Nissan BMS. You are certainly welcome to your opinion, but the unfortunate experiences of Kelangst and Klapauzius suggest to some of us that caution and prudence are in order.

 

[AndyH]

Andyh, you have frequently expressed strong optimism about the performance and long life of the Leaf battery near extremes, and the sophistication of the Nissan BMS. You are certainly welcome to your opinion, but the unfortunate experiences of Kelangst and Klapauzius suggest to some of us that caution and prudence are in order.

Actually, no - because we're not talking about the same subject. I'm talking about the cells and the management system. You're talking about wanting an accurate range display. Apples and oranges.

All cars stop when they run out of fuel. If the battery or fuel tank is empty - regardless of how many digits are displayed on a panel - the car stops. Nissan might choose to tweak the range estimate software, but that has nothing to do with the cells or the cell SOC, or how the car might sense or calculate SOC.

As for the cells and chemistry - I've shown data for medium- and high-quality LiMn2O4 cells. I've given my best impressions based on what I know about lithium (from books and papers, from hands-on testing, from building and tweaking battery management systems, and from a couple of years of on-road lithium powered EV experience) and what I know about LiMn.

While I agree completely that I'd like to get a couple of AESC cells on my bench, and while I also agree that there may be variations in performance, I stand firmly planted in my impression that all of the wringing of hands and grinding of teeth happening on this forum about the Leaf battery will be proven to have been wasted energy. Time will tell.

And no - I will not be upset with you if you choose to put me on your ignore list. :lol:

I recommend that folks take the car for a drive from 100% to --- turtle at least once so the computer can get a good feel for the battery. Part of my computer experience and A+ certification lessons is the reminder that electronic devices sometimes get out of sync with their battery. Windows tries to track battery capacity in parallel with the BMS in the battery and they will lose track of each other. From time to time, one must turn off all the power saving utilities on the laptop and let the computer run down completely until the battery dies. Once the battery's completely recharged, Windows and the battery will be in sync again. It might be a good maintenance practice for our Leafs as well.
http://www.brighthub.com/computing/hardware/articles/50844.aspx

Maybe the car cannot give it's best range remaining answer until one has driven thru a full tank of electrons at least once?

 

[ERG4ALL]

It concerns me a little that we are so engrossed in the technicalities of the batteries while more than one journalist is trying their best to see that there is no LEAF II. I'd rather see us trying to get the truth about the car out to the public that is only getting sound bites.

 

[Ingineer]

The Battery ECU keeps an estimation of the total Amp hours of the pack, and then they coulomb count (record amp-hours in/out) to determine SOC. The Amp-hour estimation is updated as the pack hits certain stages where they have a map that correlates open-circuit voltage, temperature, and SOC.

Andy is right, if you never go to 100% and never deeply discharge, it's possible the capacity estimation could drift and could be inaccurate.

-Phil

 

[abasile]

It concerns me a little that we are so engrossed in the technicalities of the batteries while more than one journalist is trying their best to see that there is no LEAF II. I'd rather see us trying to get the truth about the car out to the public that is only getting sound bites.

There is a purpose to discussing the technicalities of the batteries, as you put it. As one who is hoping to be able to take delivery of a LEAF next month, I would like to be well informed as to "best practices". Internal combustion engine vehicles have plenty of quirks, but we tend not to think about those since we've lived with them for a century. Yes, I could simply take delivery of a LEAF, charge it, and drive it. But I'd prefer to be well educated as to the particularities of LEAF/EV driving so as to get optimal use out of the car.

I am no battery expert. However, I think it is clear that, in any battery application, determining the precise amount of charge remaining is not an exact science, particularly when you have temperature variation and partial charge/discharge cycles. On top of that, even with perfect knowledge of the battery state of charge, the range estimation algorithm can do no better than provide educated guesses based on past and present usage. My feeling is that, on drives that mostly deplete the batteries, LEAF drivers would do well to pay more attention to the "bars" indicating charge remaining, and less attention to the estimated miles (range) remaining. And try not to eek every last kilowatt hour out of the battery; you wouldn't try to use every last drop left in the tank of a gasoline car. Used appropriately for drives that fit within its range with a reasonable amount of margin, the LEAF should perform quite satisfactorily. This is what journalists need to be telling people.

 

[AndyH]

It concerns me a little that we are so engrossed in the technicalities of the batteries while more than one journalist is trying their best to see that there is no LEAF II. I'd rather see us trying to get the truth about the car out to the public that is only getting sound bites.

I promise - as soon as my car arrives (and my friends lock my tools in another state so I can't take the car apart) I'll be all-over promoting the Leaf. :lol:

In the mean time, the best I can do is do is keep riding my electric motorcycle and do a battery diagnostic on my electric mower. Yes - both lithium powered and both have custom battery management. And no range anxiety for either one.

 

[Deleted member 1622]

Kelangst --

I think your feeling that people are piling on in this thread is valid. I have found similar reactions in other threads whenever I've expressed dissatisfaction with the range estimates. ANY complaint about a problem with the car is treated by a few people as a direct attack on the global alternative energy movement. Who cares about the Leaf naysayers -- those who drive this car know how great and revolutionary it is. And if Nissan has screwed up on a few things, then it's our job to point them out, not act as cheerleaders all the time.
I am not surprised that you had a rapid "miles-left burn" from 17 to zero in five miles -- it fits with similar burns that I've experienced, I just never did it at the end of the battery pack, thankfully. If I were putting money on it, I would say it's not a software glitch or battery capacity problem, just a fairly common Carwings estimate error.
I agree with all those that have said Nissan needs to act quickly to fix whatever it is that Carwings is getting so wrong. I think a more sophisticated SOC meter would be much better for driver's sanity than the Miles Left meter. The true equation for miles-left will ALWAYS depend on your route and future driving. Do we really want to punch in our destination every time we get in the car, and then expect Google maps to calculate every hill and dip in the road? I almost never use the Nav -- 99% of our trips are to places we already KNOW how to get to.
On the other issue of how we drive the Leaf to avoid full-drain: I have just under 1,000 miles on my Leaf at the moment, and I would like to suggest three tips to those getting the car or those who already have it to avoid a repeat of Kelangst's full-drain breakdown:
1. Keep a log of common trips' road mileage and compare it to what the Miles Left meter has subtracted for each trip. Over time, you will learn what the differential is. Your intuition is much better than the Carwings estimate.
2. Do not plan a trip where you will dip below 20 miles left. Besides being good for your stress level, this fits with the recommended Depth of Discharge of 20%. We really should not be driving these cars into the "dash"-mode or turtle mode on a regular basis.
3. Set a reasonable HIGHWAY maximum for all trips, and don't break the rule. My max is 50 miles on an 80% charge, 60 for 100%. Sure, some may call it conservative, I call it peace of mind.

Kelangst, I hope the negativity has not driven you away... I just take it with a grain of salt and have a laugh about some of the conspiracy theorists out there.

Best,
Josh

Well I have to say that I Am a little frustrated by the article. I was sent a PM asking to comment. My comment was:

I am going to post on this in a few, but the bottom line is I simply drove it out of power. They had engineers look at the car extensively and test everything, and the end result is that I drove too far for the charge I had. Simple answer.
  Sent: Wed Mar 09, 2011 5:01 pm

I took ownership of my driving mistake, with a slight complaint that the remaining miles was a little deceptive and the behavior unexpected. I didn't intend for my post to be taken as a complaint about the car, and in fact I think I clearly said I felt it was my fault, that Nissan responded well, and that there is a fix in the works for the display.

This started as a self-deprecating statement that I thought would be worth conversation on the board. Forgot the simple rule that once it is on Internet, it is out of my control how it is used. Guess I might not be posting my experiences again. :-(

 

[ttweed]

I am not surprised that you had a rapid "miles-left burn" from 17 to zero in five miles -- it fits with similar burns that I've experienced, I just never did it at the end of the battery pack, thankfully. If I were putting money on it, I would say it's not a software glitch or battery capacity problem, just a fairly common Carwings estimate error.

Does the fact that neither he or the dealer could get his car to SHUT OFF after it happened fit into your "common Carwings estimate error" theory? That doesn't add up to me. I am still waiting to hear what explanation the techs offered Kelangst for that.

I sure hope he has not abandoned the forum over perceived ill treatment. I don't think anyone really bad-mouthed him that badly here. The real damage was done by misinterpreting his words and taking them out of context on other sites.

TT