Charge completion and allowing time for cell balancing

[svenomous]

Hi.  KS18XL.  Haven't allowed it to charge fully in a while, simply because it's basically been in storage for the fall/winter/spring (the wet months here in the Seattle area).  Recently I've started riding again, but hadn't done a full charge yet.  Today it occurred to me that it's really overdue for a good full charge with cell balancing.  So I connected the stock charger, with a meter in between for monitoring purposes, and got it going.  After pumping about 300Wh into the wheel (which was at 80% previously), the charger LED went green and the meter showed that charge voltage was at 84.27V.  There was still a trickle charge going into the wheel of course, and I let it keep going.  It's been another 5 hours or so, the battery voltage remains the same, and the charger has reduced to about 0.07A (5.8W) of current output.  Almost 500Wh total have gone through the meter.  My question is, what is the longest I can expect cell balancing to take, in a worst-case scenario?  Could the BMS still be bleeding the cells to equalize them, or am I by now just wasting electrons and dissipating heat due to the battery pack's internal resistance?  Keep going a few more hours, or is it done?  Any way to tell the difference between battery resistance and "still bleeding power through resistors?"

[Rywokast]

at this point youre just stressing out the cells and wasting electricity i think... if it has no issue getting to 84V, as in you dont need to leave it on green forever to get there.. then there is no issue.. afaik the majority of balancing takes place before you ever get to green.. seems your battery is perfectly good.. though im no battery technician lol

[Chriull]

7 hours ago, svenomous said:

and the charger has reduced to about 0.07A

According to batteryuniversity.com liion charging should be stopped around 3% of rated current.

So for most used cells that's about 3% * (.5 ... 1)*3500mA= 52 .. 105 mA. This times parallel cells one has in his pack.

Ks18xl has 20s6p? So charging is finished should be stopped at some 0,3-0.6A.

At it seems from batteryuniversity the second stage of charging (saturation stage, Christian nstant voltage) takes some 2-3 hours. Should be fairly independend from charge current.

Li Ion cells do not like trickle charging - so 5hours in this stage are already some 2-3 hours too much 

For balancing its better one charges fully until some 0.3 A. Rides a bit and charge again...

Cell balancing imho happens most at "the beginning" of the saturation stage.

https://batteryuniversity.com/learn/article/charging_lithium_ion_batteries

Edited May 1, 2020 by Chriull

[svenomous]

Ok thanks @Rywokast and @Chriull.  Wanted to make sure since my understanding is that this particular balancing circuit doesn't activate until battery voltage is very close to full charge, so about 83.9 or higher, I'm guessing.  And then it takes time to bleed the power out of the higher-voltage cells while the whole pack continues to come up.  I assumed that this should be taken into account regardless of normal charging characteristics of LiIon chemistry (which doesn't require or benefit from a trickle/float charge), as in "hold the full-charge voltage longer, so the BMS can do its balancing thing."

As far as I can tell, the factory charger never cuts off, it just keeps holding its set voltage at very low amperage indefinitely, which is not good for the cells.  Is that true?  If so, a factory charger should never be left connected for long, even forgetting it for a day or two could do some measurable damage.  Many modern LiIon charging circuits (phones, for example) are smarter than that, but this one seems to be pretty basic and needs manual intervention.  I believe the eWheels quick charger does cut off once it reaches full saturation at the 100% set voltage, but in that case has the battery pack had enough time to balance the cells?

[Chriull]

1 hour ago, svenomous said:

Ok thanks @Rywokast and @Chriull.  Wanted to make sure since my understanding is that this particular balancing circuit doesn't activate until battery voltage is very close to full charge, so about 83.9 or higher, I'm guessing. 

Exactly. Somewhere in this range.

If one starts charging at around 80-90% state of charge (~4 .. 4.1V per cell "resting Voltage", not while charging or directly after riding. So about 80-82V...) the charging current pulls the cells quite quickly up in this 84V region and balancing happens!

Quote

And then it takes time to bleed the power out of the higher-voltage cells while the whole pack continues to come up.

No - there is no power bleeded out of the higher voltage cells, but a "bleeding" resistor put in parallel so they get less charging current.

The lower voltage cells don't get this "bleeding" resistor so they can charge a bit faster with the full charging current.

Quote

   (which doesn't require or benefit from a trickle/float charge), 

They not only do not require or benefit from it, they get stressed by this!

Quote

As far as I can tell, the factory charger never cuts off, it just keeps holding its set voltage at very low amperage indefinitely, which is not good for the cells.  Is that true?

Yes. 

As the evil tongues would say, that's just to increase the turnover for battery packs.

"Li-ion cannot absorb overcharge. When fully charged, the charge current must be cut off. A continuous trickle charge would cause plating of metallic lithium and compromise safety. To minimize stress, keep the lithium-ion battery at the peak cut-off as short as possible."

from

https://batteryuniversity.com/learn/article/charging_lithium_ion_batteries

Most chargers have 3 trim potentiometers for adjusting:

Maximum charge voltage

Maximum charge current (first stage constant current)

Cut off current threshold

Don't know if this "cut off current threshold" is just set to low or some clever designer cut this functionality...

Quote

  If so, a factory charger should never be left connected for long, even forgetting it for a day or two could do some measurable damage.

Imho its just normal increased "aging". A couple of times should cause (edit: forgot the no before... ) _no_ real harm, doing it always should noticable reduce life cycle count.

Quote

  believe the eWheels quick charger does cut off once it reaches full saturation at the 100% set voltage, but in that case has the battery pack had enough time to balance the cells?

Afaik it has different settings for cut off? 

With some, like if the pack reaches first time 82-84V (as good as no) balancing happened.

 

5 hours ago, svenomous said:

reduced to about 0.07A

Coming back to your 0.07A (for a ks18xl with ?20s6p?): This current would mean 70mA/6=11.6mA per parallel "strain". Thats about 84V/20/11.6mA=360 Ohm. That's a value which afair the bleeding resistors have about.

So they are seem to be all active and they cells just waiting to be "released".

KS wheels also can be turned on while charging/have the motherboard active while charging - so this could be just the power consumption of the mainboard and the battery packs bms have already cut off...

Edited May 1, 2020 by Chriull

[Seba]

21 minutes ago, Chriull said:

Coming back to your 0.07A (for a ks18xl with ?20s6p?): This current would mean 70mA/6=11.6mA per parallel "strain". Thats about 84V/20/11.6mA=360 Ohm. That's a value which afair the bleeding resistors have about.

In KS-18XL they're actually 100 Ω and there is only one resistor per entire set of cells connected in paralell. Anyway we shouldn't forget that control board turns on during charging, so it will always draw some power from the charger and only a part of charger current goes to the battery.

[Chriull]

19 minutes ago, Seba said:

In KS-18XL they're actually 100 Ω and there is only one resistor per entire set of cells connected in paralell. Anyway we shouldn't forget that control board turns on during charging, so it will always draw some power from the charger and only a part of charger current goes to the battery.

So there are 2kOhm|2kOhm=1kOhm - 84/1kOhm = 84 mA plus the current consumption of the MB.

@svenomous - with 70mA the batteries don't  charge anymore. If the motherboard has no cut off mechanism in firmware to cut off the charging it could be that a cell group reached the ~4.28V overvoltage threshold and the BMS cut off...

Should be noticable in a current log whilst charging. Its a pity Charge Doctor is not available anymore - hope @Inductores proceeds with his 

 

[Planemo]

1 hour ago, Seba said:

 Anyway we shouldn't forget that control board turns on during charging, so it will always draw some power from the charger and only a part of charger current goes to the battery.

Out of interest, do Gotways have anything 'on' when charging? What is the usual cut-off current for a charger? Does it vary depending on how many cells are in series? I might try adjusting my charger if required. Or at least I will check it.

[Chriull]

15 minutes ago, Planemo said:

What is the usual cut-off current for a charger?

What's set with the according potentiometer.

Most/many/all chargers have three potentiometers as described here:

If you mean the amount, it should be something like this, depending on the used li ion cells in your pack:

3 hours ago, Chriull said:

According to batteryuniversity.com liion charging should be stopped around 3% of rated current.

So for most used cells that's about 3% * (.5 ... 1)*3500mA= 52 .. 105 mA. This times parallel cells one has in his pack.

Ks18xl has 20s6p? So charging is finished should be stopped at some 0,3-0.6A

 

Quote

Does it vary depending on how many cells are in series?

No. They all get and require the same current. Current through series components is equal for them all - it only divides for parallel components. (Kirchhoff's circuit law)

 

Edited May 1, 2020 by Chriull

[Planemo]

Sorry yes I meant the amount. OK, so around 3% which on a 6P should be around 0.3~0.6A. So, at some point below this, the charger current should suddenly cut off and drop to just about zero?

 

[svenomous]

I don't have a Charge Doctor, but I do have the ability to put a meter in series with the charging circuit, to monitor volts and amps flowing from the charger to the wheel.  I could do some testing with the eWheels charger, if anyone's interested, but the stock KS charger that came with the 18XL, at least over the time period of several hours that I tested, seems to taper-charge to its set voltage (which in my case is just a smidgen under 84.3V), and then holds that voltage indefinitley.  This means the stock charger should be monitored and disconnected manually (or via smart/timer plug capability), but not immediately when the light turns green (at which point the taper hasn't quite completed yet).  I'd say about 1-2 hours after the light turns green would be a good time to disconnect the stock charger.

My memory of the eWheels charger is that if it is set to 80% or 90% it bulk charges to a voltage setpoint and stops the charge without tapering; if set to 100% it reaches the setpoint, does a taper charge, and cuts off.  I'd have to rest again to be sure...did it last year, and wrote a message about it somewhere on these forums, but don't remember precisely.

Thanks for all the discussion and info on this thread.  I know quite a bit about LiIon in general, due to a past infatuation with RC stuff (helicopters, mostly), but the specifics of wheel charging that involve parasitic draw from the wheel mainboard being active (for KS anyway), plus the BMS balancing circuit diverting charge current from higher voltage cells, etc., is good info to keep in mind.  With stock charger I now know there's no point in letting it hold the charge cutoff voltage after reaching a certain amperage, which happens not long after the green light on the charger illuminates.  Based on how the balancing circuit works, once the pack voltage is topped off, and amperage has tapered to 3% or less of capacity, any balancing is done.

Great stuff!  Thanks @Chriull and others.

[mrelwood]

6 hours ago, Planemo said:

Out of interest, do Gotways have anything 'on' when charging?

No, the charging cables go directly to the battery. The mainboard does not know wether the battery is charging or not, so the wheel can be powered on and even ridden while charging.

6 hours ago, Planemo said:

What is the usual cut-off current for a charger?

Of the five included chargers I have had (GW, KS, Airwheel, IPS x2), none of them ever cut off the charge at any point, down to 9mA or lower. And only one of them had trimmer pots inside.

[Seba]

7 hours ago, Chriull said:

Its a pity Charge Doctor is not available anymore

There is something similar that is manufactured by @EUniCycles.eu - ChargeMan:

https://eunicycles.eu/en/home/204-chargeman-king-song-84v.html

[Planemo]

3 hours ago, mrelwood said:

No, the charging cables go directly to the battery.

Thanks, I was aware of that, I just wondered if the board was parasitic whilst the battery was charging.

3 hours ago, mrelwood said:

The mainboard does not know wether the battery is charging or not, so the wheel can be powered on and even ridden while charging.

Understood, but that would be parasitic.

3 hours ago, mrelwood said:

Of the five included chargers I have had (GW, KS, Airwheel, IPS x2), none of them ever cut off the charge at any point, down to 9mA or lower. And only one of them had trimmer pots inside.

Thats what I was trying to get to the bottom of. So...leaving the charger on after the green light has gone out *could* be good for balancing but *could* be bad if left on too long....I think you see where I'm going here...

Lions dont like saturation, so if we are using chargers that dont ever shut down...the last thing we want to do is leave them connected once the green light has gone out...unless the charger cutoff (green light) happens to be set at the perfect point *and* we unplug it as soon as it goes green...

[mrelwood]

 

8 hours ago, Planemo said:

Lions dont like saturation, so if we are using chargers that dont ever shut down...the last thing we want to do is leave them connected once the green light has gone out...unless the charger cutoff (green light) happens to be set at the perfect point *and* we unplug it as soon as it goes green...

I have read about several cases where the battery had developed a severe misbalance seemingly because the user always unplugged the charger as soon as the light turned green. I don’t have any actual first hand knowledge on wether the balancing process happens mostly at the beginning or at the end of the CV charging, but it does seem that leaving the charger plugged in even after it turns green could a good idea for the longevity of the pack.

[Planemo]

1 hour ago, mrelwood said:

but it does seem that leaving the charger plugged in even after it turns green could a good idea for the longevity of the pack.

And thats how I have always thought of it too, but I assumed that at some point the charger would shut down totally (once peak voltage had been met and no more than say 300mA was being drawn) but if that isnt the case then how long after the green light comes on is too long? Sounds like a bit of a lottery if you want to carry out a full balance charge but not harm the pack.

Pretty poor show really. Luckily I have never left my wheels plugged in for any considerable time (I think max is about 3 hours) after the green light but I bet a lot of people do (overnight charges for instance).

Or, are we saying that the big packs we are generally using can tolerate being fed 300mA for such lengthy periods that theres no need to build in a total shut off circuit?

[svenomous]

1 hour ago, mrelwood said:

 

I have read about several cases where the battery had developed a severe misbalance seemingly because the user always unplugged the charger as soon as the light turned green. I don’t have any actual first hand knowledge on wether the balancing process happens mostly at the beginning or at the end of the CV charging, but it does seem that leaving the charger plugged in even after it turns green could a good idea for the longevity of the pack.

Hmm.  OK, so we know that for best health LiIon shouldn't be saturated below 3% of 1C charge current.  I'm guessing that going past that by 1-2 hours wouldn't have a measurable impact on health/longevity, but 3% is the sweet spot.  We also know from Chriull's explanation that balancing (at least in this BMS type) is done by bleeding energy and dissipating it as heat via resistors, specifically from cells that are at the top end of their full-charge voltage.  If during charge you have 17 cells in a 20s string at 4.2V, and 3 cells still at 4.1V, the 17 higher-voltage cells will be slowly bled while the other 3 continue to try to come up to 4.2V.  Eventually, all the cells in the string should be at the same voltage. Do I have things right so far?  Here's hopefully my last question, which will help answer my original question:

So, @Chriull, is it possible to have for example 10 cells in a 20s string be at 4.1V, slightly below their "full charge" voltage, and the other 10 cells at 4.3V, slightly above the full-charge level, yielding a string voltage of 84.0V?  Or, does the BMS ensure that the higher cells cannot go above 4.2V?  If the scenario is possible, to me it means that "saturating" the pack at 84V (plus or minus a couple tenths) is a good idea, at least for a little while, as a "full" pack may still have some cells above and some below the right voltage.  If it is not possible, it tells me that when the pack reaches 84V the balancing process is by definition complete, because no string in the pack could reach 84V unless all the cells in the string have reached 4.2V.

[svenomous]

28 minutes ago, Planemo said:

And thats how I have always thought of it too, but I assumed that at some point the charger would shut down totally (once peak voltage had been met and no more than say 300mA was being drawn) but if that isnt the case then how long after the green light comes on is too long? Sounds like a bit of a lottery if you want to carry out a full balance charge but not harm the pack.

Pretty poor show really. Luckily I have never left my wheels plugged in for any considerable time (I think max is about 3 hours) after the green light but I bet a lot of people do (overnight charges for instance).

Or, are we saying that the big packs we are generally using can tolerate being fed 300mA for such lengthy periods that theres no need to build in a total shut off circuit?

I ended up leaving the OEM charger attached for about 5 hours past green light, long past when my meter said the charge voltage and amperage had stabilized.  Even after 5 hours, 0.07A (a little over 5 watts) was continuing to flow out of the charger.  As explained above, some of that was consumed by the mainboard (at least for my KS wheel, which is "on" while charging), and the rest either got eaten by BMS balance resistors, or possibly some by the internal resistance of the battery pack itself.  The charger never stopped on its own.

I know for a fact that the eWheels 5A quick charger completely stops when it reaches 80% and 90% setting, because when my meter is attached it loses power when the eWheels charger cuts off, indicating that voltage going to the wheel charge port goes to zero.  I don't remember now what the eWheels charger does when it is set to do 100%.  The behavior is different, because for 80% and 90% the charger just stops upon reaching a certain voltage, whereas for 100% it does perform an amperage-taper at about 84V.  I can't remember what happens as this charger gets to that "done" amperage of about 0.5A, though.  I guess I should test it....

[mrelwood]

20 minutes ago, svenomous said:

does the BMS ensure that the higher cells cannot go above 4.2V? 

BMS cuts the charging if any of the cell groups go above 4.25-4.3V, whichever the cutoff voltage is on any specific BMS.

 When the balance starts to get out of whack, this feature is what prevents the balancing to go on long enough for the cells to get back in line. And going further, this also happens in the posts about “not charging above 93%” etc.

 Even if the pack charges to “100%”, there can be a notable imbalance, since there is room for the other cells to have a little overvoltage.

Edited May 2, 2020 by mrelwood

[Chriull]

1 hour ago, svenomous said:

Hmm.  OK, so we know that for best health LiIon shouldn't be saturated below 3% of 1C charge current.

This was afair stated earlier at batteryuniversity. Now there is written 3% of rated current, which is for some/many li ion cells at about 0.5C.

1 hour ago, svenomous said:

  .  If during charge you have 17 cells in a 20s string at 4.2V, and 3 cells still at 4.1V, the 17 higher-voltage cells will be slowly bled while the other 3 continue to try to come up to 4.2V.

The 4.2V cells get charged slower, but are still beeing charged.

1 hour ago, svenomous said:

 So, @Chriull, is it possible to have for example 10 cells in a 20s string be at 4.1V, slightly below their "full charge" voltage, and the other 10 cells at 4.3V, slightly above the full-charge level, yielding a string voltage of 84.0V? 

Almost. Overvoltage cut off threshold for single cell (groups) should be somewhere around 4.25-4.28V. So 4.3V should never be possible.

Also the distribution of voltages could be a bit improbable.

Mostly, from reports here, it is just one or two cell groups that deviates in voltage - the others are more or less at the same voltage.

1 hour ago, svenomous said:

Or, does the BMS ensure that the higher cells cannot go above 4.2V?  If the scenario is possible, to me it means that "saturating" the pack at 84V (plus or minus a couple tenths) is a good idea, at least for a little while, as a "full" pack may still have some cells above and some below the right voltage.  If it is not possible, it tells me that when the pack reaches 84V the balancing process is by definition complete, because no string in the pack could reach 84V unless all the cells in the string have reached 4.2V.

At around 4.2V the bleeding resistor is put side n parallel to this cell group, so it gets charged slower.

At around 4.28V the BMS cuts off the charger.

[svenomous]

OK, thanks. My examples were far-fetched to make them easier/simpler, but the principle seems to stand: a pack with 20s strings can reach 84.0V (if that's the charger's set voltage) with some cells still being below 4.2V and some being above it.  Therefore, when the charger reports "full," some balancing may still be going on.  Going 1-2 hours past "full" on a charger that doesn't do its own charge cut-off seems like it has a potential benefit of giving the balancing a chance to be more complete.

[esaj]

On 5/2/2020 at 9:20 AM, mrelwood said:

 

I have read about several cases where the battery had developed a severe misbalance seemingly because the user always unplugged the charger as soon as the light turned green. I don’t have any actual first hand knowledge on wether the balancing process happens mostly at the beginning or at the end of the CV charging, but it does seem that leaving the charger plugged in even after it turns green could a good idea for the longevity of the pack.

 

On 5/2/2020 at 10:56 AM, Planemo said:

And thats how I have always thought of it too, but I assumed that at some point the charger would shut down totally (once peak voltage had been met and no more than say 300mA was being drawn) but if that isnt the case then how long after the green light comes on is too long? Sounds like a bit of a lottery if you want to carry out a full balance charge but not harm the pack.

Pretty poor show really. Luckily I have never left my wheels plugged in for any considerable time (I think max is about 3 hours) after the green light but I bet a lot of people do (overnight charges for instance).

Or, are we saying that the big packs we are generally using can tolerate being fed 300mA for such lengthy periods that theres no need to build in a total shut off circuit?

 

Well, I have no real answers, but just for a single reference:

I have had my KS16S / 840Wh (16S4P with Sanyo cells, don't know the exact model though) for over three years now (manufacturing date for the wheel itself is 17.03.2017 based on serial number, of course the batteries are made before that). I'd love to tell the mileage, but for some reason the non-music BT is very, very touch'n'go (usually no-go, like right now) with my current phone. I'm still using the old "green app". The firmware is still 1.00, never updated.  Not ridden very much, I'd guess less than 5000km / 3000 miles, but could be wrong. The usage pattern has been:

-mostly commuting (depending on where I worked each year, this could be 10 to 25km per weekday) and some errands for 5 days a week on average, between late April/early May and late September / early October (about 5 months per year)

-Occasional "fun rides" here and there, 10-50km at a time, but not very often

-Stored in room temp totally unused at 30-40% charge for rest of the year (about 7 months per year)

I don't remember the last time I checked the total capacity (might be a couple years), but at that time it was still at around the nominal 828Wh or slightly above (the 840Wh -number is calculated with rounded values, but every manufacturer does that, and on the few occasions I've had the chance to measure new cells, they've been above the nominal mAh -rating, for brand-cells that is).

I usually charge all the way to full on weekday evenings (when I know I'll ride it to work in the morning), that is, usually something like 10-70mA charge current at full voltage (measured with Charge Doctor V2) when the charging is stopped. Disconnect before I go to bed, most of the time it has easily reached <100mA charge current at that point, I'd guess usually something like 0.02-0.04A (20-40mA). The wheel is then left as is in room temperature until next morning. Against weekend, I don't charge the wheel on Friday-evening, unless I know I'm going riding on Saturday morning, usually I just charge "as needed" on weekends, and do a full charge on Sunday evening.

Since I have the Charge Doctor, I don't pay any real attention to the charger light, but the charger I use goes to green around 250mA. I have forgotten to unplug and left the wheel being charged overnight now and then, but not often... less than 5 times per year, I'd say, more like 1-2 probably on average.

No dead cells, no indication of any abnormal wear & tear on cells. I can still do the stupid "from zero to full speed so fast that the tiltback scares the shit out of you" -accelerations without issue. Of course, now that I say it, it might fail on me the next time... knock on wood.   And I'm lower weight than your average rider (about 60kg / 130lbs). But for all I can tell just by riding (haven't bothered to tear it down for accurate measurements), this usage pattern hasn't at least significantly hastened the doom of the cells. 

I'll try to do a full "ride it all the way to dry" -test some day and measure the total capacity of the cells. Likely they have degraded, since the cells probably lose at least a couple of percent of total capacity each year even when stored "perfectly". Let's hope I'm not in for a nasty surprise, but it's fairly likely that I couldn't break 60km/37 miles on these cells anymore....

Edited May 7, 2020 by esaj

[mrelwood]

@esaj, thanks for the detailed data point! It seems perfectly in line with my current beliefs.

And my current charging habits as well, actually. Good to know that I finally seem to be on the right track!