How to mitigate a badly out of balance pack?

My Nissan Leaf Forum

Help Support My Nissan Leaf Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

mwalsh

Well-known member
Leaf Supporting Member
Joined
Apr 22, 2010
Messages
9,781
Location
Garden Grove, CA
The other day I was having some real range issues. I got to work with ever so slightly less than I would normally see, but going home my range was falling WAY more rapidly than it should have. When I eventually got home (I had to take it easy the last 10 miles), I unplugged the Gidometer and plugged my ELM in instead, only to see that the voltage range between the lowest and highest cell was a whopping 68mV! It is HIGHLY unusual for my pack to be over 20mV, and it's generally between 15mV and 17mV. So pretty obvious why I ran low on charge much earlier than I would have expected.

But the question is...how can I mitigate this in the future?

Well, clearly it would help if I was using the ELM all the time, instead of the Gidometer. If I'd seen my pack was either out of balance already before leaving home, or noticed that it was edging towards becoming unbalanced at the mid-point, I might have been able to do something about it. But I do have my issues with the Android device - the first of which is I find it difficult to read the screen while wearing polarized glasses. Perhaps an OBDII splitter might be the answer to that (if there is such a thing?), so I can keep both devices going at once.

But the bigger question is...what exactly can I do if I found my pack appeared to be badly out of balance? Would any charging event attempt to re-balance it, assuming the BMS sees the imbalance (and hopefully it would!)? Would it need to be a charging event from a relatively low SOC? Would I have to just tough it out until the next time the BMS decides a rebalance is needed or could I force the BMS to rebalance?

Edit: Oh, one nugget I can give you is that my pack "seems" to be more out of balance on Mondays, especially after it's sat at a relatively low SOC all weekend because I haven't needed to use it.
 
mwalsh said:
But I do have my issues with the Android device - the first of which is I find it difficult to read the screen while wearing polarized glasses. Perhaps an OBDII splitter might be the answer to that (if there is such a thing?), so I can keep both devices going at once.
I tried to run two devices at once with a splitter. Didn't work and was told I shouldn't do that. My suggestion: Get the LeafDD. You can easily read Gids and pack temperature on screen 1 (even with polaroids on), can switch to screen 3 to see max, min and average cell voltages.

But the bigger question is...what exactly can I do if I found my pack appeared to be badly out of balance? Would any charging event attempt to re-balance it, assuming the BMS sees the imbalance (and hopefully it would!)?
Apparently it is balancing all the time, but the best balance if you need the full range is to charge to 100% and keep the charging going for several hours afterward, during which time it should balance. It may also be that the pack is out of balance because you have a weak cell which is exposed at low SOC. What is the balance like at 80% or 100% charge?
 
How often do you see LBW ?

How often do you see VLBW ?

Do you normally charge with L1 or with L2 ?

The more often you see a low SOC, the more often the pack will be out of balance. Every car is different, but my car will balance better if I run it down to LBW then recharge to 100% using the 120 EVSE. It also seems to help to let the pack cool off for several hours after driving it. End only timer works well for this.
 
KJD said:
How often do you see LBW ?
Every day.

KJD said:
How often do you see VLBW ?
Rarely (for now).

KJD said:
Do you normally charge with L1 or with L2?
L2. Could probably do L1, given that I have 15 hours from needing the car one day to needing it the next. Except that I'm too lazy to drag the L1 charger out every day, and the pack battery wouldn't get any rest/cooling off time.
 
I only had this happen once (range falling more rapidly than expected near the low end), so I attributed this to an imbalance condition. A couple of full charges (including a 4-hour "balancing" opportunity), seemed to do the trick. When balancing takes place you'll get a 2nd "charging complete" message if you have notifications set in Carwings. I got that notification on 2 subsequent attempts, but not the third, so it seemed I needed 2 sessions to get back to normal.

I have made it a point since then, to give the car a "balancing opportunity" once a month, which includes the 100% charge and at least 4 hours of plug time afterwards. Even if I don't specifically need the full charge. Sometimes I get the 2nd message; more often not. The car seems to know when a rebalance is needed, and just needs the opportunity to exercise it.

I don't have any kind of Gid-meter, fwiw.
 
Speaking to the root of the problem, Mike, I think that you only saw the large variance in cell voltage because you were at very low charge. In my experience, I've only seen a large variance at very low SOC. So my guess is that you wouldn't have seen that large gap earlier in your trip, even if you had been running the ELM app.

A couple of other comments. I run the app on an old Droid X2 and a newer Samsung SCH-I405, and neither one has a problem with viewing the screen with polarized shades. The screens on both of these phones seem to have the interference set up at a 45 degree angle, rather than on portrait or landscape orientations. You might keep an eye open for an opportunity to swap your Android device with someone else who doesn't need to view the display with shades on (slim chance, but I thought I'd throw it out there).

And lastly, I'd concur that the best balancing opportunity is to have the car charge to 100% with no end timer set and allow the car to remain attached to the J plug for a few hours after the charge is complete. Doing this for a couple of charge sessions has done the trick for me.
 
Boomer23 said:
Speaking to the root of the problem, Mike, I think that you only saw the large variance in cell voltage because you were at very low charge. In my experience, I've only seen a large variance at very low SOC. So my guess is that you wouldn't have seen that large gap earlier in your trip, even if you had been running the ELM app.
Yep, 68 mV isn't necessarily out of line, depending on exactly how low the SOC was. So what was it? Screenshots are a great way to get all the data across.

Boomer23 said:
And lastly, I'd concur that the best balancing opportunity is to have the car charge to 100% with no end timer set and allow the car to remain attached to the J plug for a few hours after the charge is complete. Doing this for a couple of charge sessions has done the trick for me.
Do you have any before/after data to show the effect?
 
With regard to splitting the CAN bus, the attribute of note is whether the device(s) just listen to the bus or whether they issue commands to get their information. I believe Gary's original meter is in the former group, and I can't speak for the others, but its likely most of them issue commands, since that's how a lot of the info is obtained. Especially where it asks one controller to go get data from another (in English: asking the car controller to get info off the EV controller and so on). Readers that issue commands can have collisions if there's more than one operating.

[Edit, based on Stoaty's subsequent message with a link to the relevant thread--the more conventional term for "command issuers" is "Active Devices". As in, you shouldn't attach two active devices to the bus. Especially if they are going to ask for the same data.]

As for the low cells, it would be instructive to note which cells are the offender, and whether its the same ones each time. It isn't really the 68 mV difference that's a factor; it's really the low cell compared to the average. After all, having a few "hot" cells doesn't really contribute to your expected range. But even so, the "penalty", so to speak, is that if one cell is 68 mV low then we have to pretend they all are. 96 times that is 6.6 volts. As usual, I can't find the chart that would show how much earlier you'd run out of battery if it were 6V low.
 
So, it has to stay plugged in after 100% is reached to optimally balance, and for at least four hours?

Why is that? Does it draw small amounts of power occasionally during this process? I've never noticed it draw anything after 100%, but then, haven't ever looked for it.

Like most of the general public, I don't have any extra meters (although if there was a LEAF Spy for the iPhone I would), and I don't think there's anything on balancing in the manual, but maybe there should be.
 
DNAinaGoodWay said:
So, it has to stay plugged in after 100% is reached to optimally balance, and for at least four hours?

Why is that?
When the battery sits without charging or discharging, the cells should eventually get to about the same temperature. That allows for optimum balancing.
DNAinaGoodWay said:
Does it draw small amounts of power occasionally during this process?
Yes, it will sometimes charge a little bit after the battery has gotten into an isothermal condition to fully balance the pack.

To OP: I find that running the climate control for an hour or so after the end of the L2 charge seems to bring all the cells up to their maximum voltage (4.1V) so that I have the best range for a trip. This works best when it is cold and the heater runs, which is typically when I need the boost. It's inefficient, but sometimes I can get an extra few miles this way.
 
DNAinaGoodWay said:
Why is that? Does it draw small amounts of power occasionally during this process? I've never noticed it draw anything after 100%, but then, haven't ever looked for it.
I've seen it draw power after "100%". Once the Chargepoint EVSE communications got fixed at my work, I am able to reliably see the power consumption graph but it's only at 5 minute intervals. Basically, when charging to 100%, once the draw drops to 0 kW, I've noticed that it periodically pulls some power, maybe ~0.4 to ~0.8 kW. It eventually seems to completely stop.

I don't think I've left it plugged in for 4 hours at 100% w/any sort of monitoring though.

The % SoC on the '13s also can have a jump/"mysterious move" (those who were present at the SF BayLeafs meeting in December 2011 will recall a "mysterious move" slide). I can be charging to 100% but there's still some time remaining, the power draw has ramped down a lot (but not 0 kW yet) and the % SoC (part of the b/w dash display) reads 98%. When I unplug and power cycle the car, % SoC often jumps to 100%.
 
RegGuheert said:
When the battery sits without charging or discharging, the cells should eventually get to about the same temperature. That allows for optimum balancing.

So, first it equalizes temperature, then the cell voltages equalize?

When I can get off this phone and to a desktop, I'll search old threads, I remember Gary Gid writing that most of the balancing is done in like 45 minutes or so. So maybe 2 hours is enough?

But then, Tony's range test would start with a car balancing for six hours iirc, maybe he should try your cc trick.
 
cwerdna said:
DNAinaGoodWay said:
Why is that? Does it draw small amounts of power occasionally during this process? I've never noticed it draw anything after 100%, but then, haven't ever looked for it.
I've seen it draw power after "100%". Once the Chargepoint EVSE communications got fixed at my work, I am able to reliably see the power consumption graph but it's only at 5 minute intervals. Basically, when charging to 100%, once the draw drops to 0 kW, I've noticed that it periodically pulls some power, maybe ~0.4 to ~0.8 kW. It eventually seems to completely stop.

I don't think I've left it plugged in for 4 hours at 100% w/any sort of monitoring though.

The % SoC on the '13s also can have a jump/"mysterious move" (those who were present at the SF BayLeafs meeting in December 2011 will recall a "mysterious move" slide). I can be charging to 100% but there's still some time remaining, the power draw has ramped down a lot (but not 0 kW yet) and the % SoC (part of the b/w dash display) reads 98%. When I unplug and power cycle the car, % SoC often jumps to 100%.

I'm on a ChargePoint right now at work for 8.5 hours. Started at one bar, should take ~5 hours to 100%, I'll check to see what it draws after that.
 
gbarry42 said:
With regard to splitting the CAN bus, the attribute of note is whether the device(s) just listen to the bus or whether they issue commands to get their information. I believe Gary's original meter is in the former group, and I can't speak for the others, but its likely most of them issue commands, since that's how a lot of the info is obtained. Especially where it asks one controller to go get data from another (in English: asking the car controller to get info off the EV controller and so on). Readers that issue commands can have collisions if there's more than one operating.
LeafDD and Leaf Spy don't mix:

http://www.mynissanleaf.com/viewtopic.php?f=44&t=12098&p=313237&hilit=leafdd#p313237" onclick="window.open(this.href);return false;
 
DNAinaGoodWay said:
I remember Gary Gid writing that most of the balancing is done in like 45 minutes or so.
There's a bit of history involved, such that in the beginning, the only clues we had about balancing were a page in the service manual that mentioned shunts, and the fact that the charging tapered off at the end. Plus a few reports of the charge coming back on after awhile. Later on, ingineer mentioned that balancing could take place at times other than the 100% charge. Now, with the LEAF Spy, we can see what appears to be balancing (shunting) going on all the time. But we must remember most of what we know is gained empirically, since Nissan won't tell us what actually is going on.

Your results will vary with the charge turning itself back on. Mine never does this, so maybe this battery doesn't get too far out of balance.
 
Back
Top