one cell group wont fully charge

[m3vuv]

Hi all,i have a homebrew 13s 10p pack with a chineese smart bms over bt,i notice cell groupe 9 never charges more than 4.108v any ideas why this is? all other groups of cells seem fine,cheers.

 

[docw009]

I guess I would disconnect the BMS harness, and hook up my RC charger to group 9, and see if it will hit 4.2V alone. After that, if it still didn't go any higher, I would be sad because there's a bum cell in that 10P group limiting the charge. I think it has to come out. It must be really pulling current at 4.1V if it holds the other 9 cells back.

 

[m3vuv]

the cells are samsung 15q,controller is 500 watt motor is 1kw,pack has about 100 cycles on it,its only been used until the lvc about 5x,there are diodes in the throttle signal wire to cut max signal voltage to 3v max using a rotary switchin each group there are 9x 15q of 1.5ah cells and one samsung 2.5ah cell per groupe,cant remember the part number tho,the pack only charges to 53.4v max for some reason.

 

[m3vuv]

imwaiting on delivery of a litkala lii500 then will test the cells,i am going to leave my bike on charge overnight and check pack voltage on the throttle display and with the bt app,,need to take the guts out of the battery to fit a uart interface anyway so will tear it apart then.

 

[m3vuv]

i dont see how cells with a high ir would afect the other cells in a group on charging but can see cells with a high ir not delivering current under load?,am i right on that?

 

[m3vuv]

this is it it at mo on charge,this is from the bms phone app.

 

[m3vuv]

the cells are in paralell tho in each group so should ballance themself equaly as i see it!?.

 

[m3vuv]

the cells were not new but supposedly unused,i used repakrr program on the net to match the cells when doing the build.

 

[docw009]

1. 4.200 8. 4.151
2. 4.165 9. 4.124
3, 4.147 10. 4.201
4. 4.152 11. 4.123
5. 4.129 12. 4.153
6. 4.187 13. 4.135
7. 4.155

I'm sorry. I thought you had 130 new cells of high capacity. I wouldn't use a 15R, but it seems like a goood high current cell. I withdraw my earlier comment. These results of. 420 to 4.13 volts, max -min,are what I have seen when I made a battery out of old cells. You also have that Samsung 2.5AH cell in each row, and the combo of the high IR's in the 15R's might cause some funny happenings.

It would be interesting to know the true AH capacity of your battery. On paper, it's 16AH in a perfect world. On an ebike, I would guess 10 AH. I've found that the capacity measurements from my Littohala 500i, which are done at 500 ua are always much higher then what I get out the cells in a battery running an ebike. However, if your 500W controller is only 10 amps, you night do a lot better. Do you own a wattmeter, or can the BMS app measure the total AH as your ride?

When you run the battery down to where the bike or battery shuts off, what's the voltage on the battery afterwards? Can you watch the voltage sag on your BMS app when the bike is running? Is the sag pretty minimal with 10P? Are the groups staying balanced as the capacity drops?

Lot of questions. Don't need the answers, but if you get some time to find out. the rest of us can learn.

 

[amberwolf]

The difference mamp-hr cells within each parallel pack will by nature discharge at different rates causing them to be at different voltages within the same parallel groups.

It's impossible for paralleled cells to be at different voltages, because they are directly wired in parallel with such a low-resistance connection it is essentially none, under normal conditions.

(under extremely high current flow in an improperly interconnected pack, that causes voltage drops across high-resistance interconnects, it could happen during the high currents, but not under any normal scenario).

They could be different capabilities, different capacities, but always the same voltage across all the cells in a single parallel group.


FWIW, this kind of thing means that a single defective cell, failing the right way, within a parallel group can drag down the entire rest of that group to a lower voltage, causing the type of problem the OP of this thread has.

I haven't heard of anybody else building a battery using different mamp-hr cells within the same battery enclosure.

Many cheap batteries are made exactly that way, using literal recycled garbage cells. Often enough they are not even all the exact same chemistry because the sources vary, could be recycled laptop cells, toolpacks, factory discards, etc.

People here on ES have also done this (I leave judging success or failure of each such build to the reader ).

Other cheap batteries are made of new cells, but not tested or matched, so they are different capacities and capabilities.

Others are made of new cells but not all the same brand/model/etc.


If you poke around the many battery troubleshooting/repair/build threads, you can see examples of all of these and more.

 

[m3vuv]

a couple of pics after the pack has been on charge overnight.

 

[From-A-To-B]

a couple of pics after the pack has been on charge overnight.

If it were my battery, I would not be satisfied with the 4.13-4.20 voltage spread. That pack’s state of charge is spread between 92-100% charge. Looks like your BMS has been accurately stopping the charge when the first group hits 4.2V. If you can easily access the balance plugs, I’d look to manually balance the rest of of those cells.

That is, of course, in addition to addressing the issue of a suspected bum cell.

 

[From-A-To-B]

Sure! Here’s how I’d go about it, noting that other people many have different (or better!) ways to go about it.

The easiest thing to do would be to top-balance: charge the pack on your bulk charger, then use a hobby charger to bring up the groups to 4.2V, one string at a time. Others have advised that people can use 5V, 1A cell phone chargers (wall wart), but you absolutely must sit with the charger and monitor the charge so one can avoid an overcharge.

Alternatively, you could drain your high groups to match low ones. Any sort of resistive load would work, whether automotive bulbs, a heat lamp, a power resister, what have you. Go through the pack one cell group at a time and drain the high ones to match the low ones.

You can do either of these low-power balance procedures through the balance leads as they enter the BMS. (Use a small gauge wire to probe the balance connector). Alternatively it could be easier to open up the pack, strip the shrink, and access the buses between the cell groups directly. (Small alligator clips are useful).

Based on your two before/after photos of the battery that spent the night on charge, it is unclear to me whether your BMS is shuffling energy around to balance the pack. The battery isn’t in much better balance after a night on the charger.

 

[eMark]

If it were my battery, I would not be satisfied with the 4.13-4.20 voltage spread. That pack’s state of charge is spread between 92-100% charge. Looks like your BMS has been accurately stopping the charge when the first group hits 4.2V. If you can easily access the balance plugs, I’d look to manually balance the rest of of those cells.

A variance of 72mV (4.200V to 4.128V) is probably typical for a DIY battery pack of used cells that were originally balanced, but gradually become more unbalanced with more cycles. Thus the reason why "active balancer boards" were invented ... https://www.amazon.com/Lifepo4-Equalizer-Balancer-Inductive-Transfer/dp/B09C3H3HB3/ref=sr_1_1?gclid=CjwKCAjwsfuYBhAZEiwA5a6CDCi70-Bw8RTetLdpWOTXMEwN2MsLMRIlPYe-jLqUXxYj3i8fVMEGuhoCaL4QAvD_BwE&hvadid=600895420700&hvdev=c&hvlocphy=9019693&hvnetw=g&hvqmt=b&hvrand=5708651524475718811&hvtargid=kwd-1569122188046&hydadcr=7690_9903750&keywords=13s%2Bactive%2Bbalancer&qid=1663001223&sr=8-1&th=1 ... worth getting one if only for some experimenting. One that i bought (amazon) for experimenting was faulty so returned it ... was no problem to return the bad board for a good one.

With expense of new brand name cells DIYers are buying used (Grade A/B) cells that will need occasional balancing ...

That is, of course, in addition to addressing the issue of a suspected bum cell.

If the voltage variance between 12 of your 13 p-groups is only 72mV, but one of your p-groups is say 3.889V (311mV variance) then a weak/bum p-group cell(s) is certainly to be expected.

If this was my DIY battery build i'd do the following as much as a learning experiment and for solving the problem ...

• 1. Discharge p-group 1 (4.200V) to 4.130V resting volts

• 2. Discharge p-group 10 (4.200V) to 4.130V resting volts

• 3. Discharge p-group 6 (4.185V) to 4.130V resting volts

• 4. Discharge p-group 2 (4.171V) to 4.130V resting volts

Then charge the pack to 4.20V and report back to us the variance of the 13 p-groups. My guess is the variance between the 13 parallel groups will then be no greater than 30mV. This is a typical rebalancing procedure with any DIY build of used questionable cells. Doing so will tell reveal the overall condition of your battery parallel groups. Your battery may not be bad if all it needs is occasional balancing of the p-groups when more than 50-70mV out of balance. There's only so much a BMS can do with a DIY build using cells of questionable quality.

 

[amberwolf]

In a correctly built pack being used correctly within it's limits, the resistances between cells are so low as to effectively prevent differences in voltage. (realistically there is always some resistance so there might be microvolts difference simply from the interconnects themselves, but for our purposes and practical usage, this doesn't matter).

If the pack is poorly built with high enough resistance between the cells to see a voltage difference across the parallel interconnects during current flow thru the group, then there can be a difference between the cells

It is very easy to test this yourself by instrumenting with individual identical voltmeters (or a logging volt measurement setup) a parallel group of cells with different characteristics, and a correctly-installed paralleling bus with as low a resistance as possible that is essentially zero for the designed level of current flow within the pack. Start at zero current flow thru the group, either charging or discharging, and increase it until you have reached the maximum allowable for the cells.

.

To see the full effect of this, you'd need to instrument identical voltmeters across each cell (at the cell casing), and across the cell-to-interconnect interfaces at each end, and across the interconnects between each cell. This way you can measure the voltage drops across every part of the group, to see exactly which parts are caused by pure resistive loads and which are caused by the cell characteristics.

But in a generic 18650 pack you can see the basic effect by simply connecting meters to the interconnect spotweld area at each end of each cell in a single parallel group.



If there is no voltage across the interconnects between the cells, there is also no difference between the cell voltages, because the voltage must be developed across some resistance by current flow thru the resistance for there to be a difference between the voltage sources.

Even if it were a pure resistive setup, with no voltage / current sources within the resistances replacing the cells, then even with different resistances in parallel, there would still be the same voltage across every one of them in parallel during current flow, assuming that the interconnect resistances are essentially zero. If the resistance of the interconnects is high enough you will see voltage drop across them, and thus differences in the voltages on the resistors, just as you would with cells of different characteristics instead of different pure resistances.


It comes down to the resistances of the interconnects--if they're high enough, sure you'll see cell voltage differences under current flow that's high enough. Otherwise, you won't.



FWIW, a lower capacity battery (or cell) doesn't have a different voltage than a higher capacity battery *unless* it is *also* at a different state of charge, all other conditions the same.

It's impossible for paralleled cells to be at different voltages, because they are directly wired in parallel with such a low-resistance connection it is essentially none, under normal conditions.

I believe you missed the point and your blanket statement above just isn't true as the battery pack experiences load.
Further if you have a badly shorted failed cell within a parallel group, that cell will not have the same voltage as the other cells in the same parallel group.

The following is from an electrical engineering forum addressing the dynamic of inequivalent parallel cells discharging at a different rate under load and how cell voltages vary 'during discharge':

The circuit obeys Ohm's law at all times, so during any two batteries connected together, when initially both batteries are charged to same voltage, there is no current between them. When connected to a load, they both discharge to load via their ESR. As the current through ESR depends on voltage at the load and voltage at the cell, the cell can never discharge to lower than voltage at the load, so during the time when load is connected, no cell charges the other one. When disconnected from load, there are two different cell voltages connected via ESRs, so in this case higher voltage will charge the lower voltage. Feel free to simulate or calculate which battery charges which, depending on their capacity and ESR, as Q=It=CU. If ESRs match, the lower capacity battery has less voltage and gets charged after removing load. If batteries have same capacity, the one with lower ESR has less voltage after removing load as it gave out more charge. In real world though, the battery chemistries and other behavior of the batteries won't match so it depends on other things than capacity and ESR as well.

 

[eMark]

Can you explain why adjacent cell parallel groups connected in series to the 'targeted parallel group' being charged, why they don't derive some residual voltage from the target parallel group being charged with the hobby charger?

In answer to your question i once got distracted a couple days ago while charging one p-group suffering from self-discharge (30Q). The variance of the other 9 p-groups (10s3p - 30Q cells) was only 14mV (3.511V to 3.525V). The one i was charging at 1.0amp rate was at 3.487V at the beginning. After remembering and stopping the charge it was over 3.80V. This had no noticeable affect on the voltages of other 9 p-groups. Using my discharger i discharged it in line with the other 9 p-groups ... no harm done. This is my handy-dandy discharger for running capacity checks, etc ...

No ...the other p-group voltages apparently aren't affected, at least mine weren't. In a recent experiment 7 of the 10 parallel groups (10s3p 30Q cells) stayed within 5mV of each other after discharge. The other 3 p-groups experience self-discharge within 24-48 hours for the past few months. So i always bottom balance before bulk charging at 2amp rate to 41.0V. About 300 c/d cycles on my experimental 30Q 10s3p Vruzend V.2 kit going on 3 years of good use.

I always bottom balance these 3 p-groups suffering from self-discharge to a higher voltage than the others say +0.005V. Than wait a day with self-discahrge again below that of the other 7 stabilized p-groups. I do this a couple times until the 3 p-groups also stablize before bulk charging the 10s3p battery to 4.10 volts.

What i've recently tried is charging these 3 p-groups 30mV higher than the other 7 p-groups and then immediately bulk charging the battery to 41.0V. Then a discharging ride from 5 bars to just 1 bar of casual riding. Would you believe that the 3 problem p-groups still have higher voltage (after discharge) than the other 7 p-groups with the self-discharge of these 3 p-groups not as noticeable after 24-48 hours.

 

[m3vuv]

I will wait until my litokalla lii500 gets here then pull the pack remove the cells on the iffy groups and test them,i have a 150 watt electronic load to discharge them,i want to open the pack anyway to fit the wires for a uart interface i have,the case is made of 2mm folded steel held together with ss allen bolts so easy to do,heres my bike,bit heath robbinson but covered over 5000 miles so far on it ignoor the crutches strapped to it,i was recovering from a foot operation at the time!,the tourq reaction arm is a cheap adjustable spanner hose clipped to the frame,bit of a bodge but cheap and works lol.

 

[m3vuv]

incidentaly when the charger is cycling on( red off green) say on for 2 seconds then off 2 seconds than back on for 2 seconds i assume its trying to balance the cells?,if so i wonder if i should leave it on charge for a few days?

 

[eMark]

It would be worth your time (DIY with used cells of questionable quality) to add sensing/balance leads to each parallel group so you can occasionally bottom balance the p-groups (as needed) before bulk charging.

 

[amberwolf]

A-To-B, if you owned this battery could you please provide a roadmap aka numbered steps of how you would go about 'manually balancing' all cells in the battery that aren't deemed to be failed...failed cells can't be balanced of course and need to be replaced. But curious what procedure you would use to balance all the cells in this pack when you have an assembled pack interconnected with spot welded nickel strips.

Replying not just to the above, but generally for the many current and future that don't already know all this stuff:

Balancing is not done externally between cells within a parallel group. That's done by the parallel interconnects which force all cells within the group to the same voltage. (if that isn't happening, there's something wrong iwith the interconnects that needs to be repaired).

Balancing is only done between different parallel groups to bring them to the same voltage. (and it only brings them to the same voltage, not the same capacity).

Below are some basic methods of doing this, details vary between specific systems and people using them. This doesn't include much about the bad side of doing certain things, or what can happen if voltages are too different, etc etc. Just the basic idea:

Most common is top balancing, which is what almost all balancing BMSs do, where they cause the cells to end up equal in voltage at full charge. There are a few methods to do it, but the most common in ebike battery BMSs is to resistively shunt aside some charge current on the first cell group to reach a trigger voltage, while the rest continue to charge, and every subsequent group to reach that trigger voltage also will have current shunted around it, until all groups are equal in voltage, and charge current is cut off at the charge port of the BMS. If voltage continues to rise on any of the groups past a second trigger point, while charge current is still flowing, the charge port is turned off to allow the shunts to drain down those groups until they reach the shunt-turn-off trigger voltage. The process is repeated automatically by the BMS until they all equal the same voltage (which is all that "balanced" means in context of typical battery packs you will see used here), and the charge and balance process ends.

This process can be performed manually just as well, if not as easily, using an external resistive load to connect directly across any parallel group higher in voltage than the others, until it drops down to match the voltages all other groups are at. This can be done with or without the charger and/or BMS connected to the pack. When all groups are equal in voltage and at full voltage, the charge and balance process is complete.

Another variation of top-balancing is to use individual chargers for each parallel group, set to the full voltage of that group. No group can be imbalanced at end of charge unless the charger itself is set wrong or failed, or the group itself is so badly degraded or damaged that it can't reach full voltage anymore.


Less common but still used is bottom balancing, typically done manually rather than by BMS (but could be done that way), where every cell is drained down to the same point once the first group has reached the lowest state of charge it is allowed to.


Another method is charge shuffling, where a system of "bucket brigade" type devices capacitively move charge from a higher voltage cell group to a lower voltage cell group. This may be done only during charge, or it may operate continuously the entire time the BMS is connected to the pack. (While I haven't yet seen this happen, the latter method, if no safety shutoffs exist in the hardware or firmware controlling the process, could kill an entire pack of cells, if there is a leaky cell in any group, as it will keep dropping in voltage and drag it's group with it, and the BMS will keep trying to fill that group from other groups until they are all drained down to empty, and potentially beyond).

A variation of charge-shuffling that's sometimes done manually is to disconnect the series connections between every parallel group, and manually connect every parallel group in parallel with every other. Depending on the interconnect resistance (which ideally would be high enough to limit current flow significantly), this can take some time, or be very quick (assuming the cells are designed for high charge and discharge rates and the interconnects are all low enough resistance to not heat up during the process).

A variation of that one that's often used on RC LiPo packs is to disconnect the series connections between every separate pack (often 6s in ebike packs built this way) and connect the main and balance lead connections of every pack in parallel with every other pack, to let them equalize. This is also often done while connected to an RC charger (sometimes in balancing mode), to complete the rest of charging.


Note that during charging, the higher voltage cell groups are the ones with *less* capacity and/or higher internal resistance. (the opposite of what you see during discharge).



Also note that to actually fix any pack with an imbalance caused by cells and not interconnect or BMS failures, cell replacement is required, because the imbalance is caused by cells with differing capacities and/or other characteristics, and rebalance (to whatever degree) will be required after every use of the pack.

Assuming well-matched cells used well within their limits, then:

An imbalance caused by an interconnect failure (such as a group with a disconnected cell or cells that no longer contribute to that group's capacity and other capabilities) requires repair of the interconnect, and possibly replacement of the cells that were still connected if they were stressed beyond their limits enough during the time the problem existed to become different from the rest of the cells in the pack. After that the pack would remain balanced, until the cells degrade enough to become mismatched

An imbalance caused by a BMS failure (such as a stuck-on resistive balancer continually draining a whole parallel group) requires replacement or repair of the BMS, and then rebalance of the groups, either manually or using the new BMS. After that, the pack would remain balanced until the cells degrade enough to become mismatched.


Assuming the much more common scenario of unmatched cells used at or even beyond their limits, the cells will never remain balanced and will always require rebalancing to some degree after every use.

 

[amberwolf]

Thank you for your explanation and perhaps you can answer a general question about the above procedures.

All parallel groups of cells are connected by nickel strips in series for an aggregate battery voltage.

If I go inside the pack and strip the shrink and attempt to connect two leads via aligator clips to each side of a parallel cluster, one side of a single parallel pack being positive and other side of the single target parallel groups being negative in an effort to balance this cluster to a target voltage, why isn't this target voltage 'tainted by being connected in series to other parallel clusters within the battery? Isn't there bleed off of voltage to interconnected parallel groups in series within the battery?

Can you explain why adjacent cell parallel groups connected in series to the 'targeted parallel group' being charged, why they don't derive some residual voltage from the target parallel group being charged with the hobby charger?

If you connect a voltage source (charger) across just the one parallel group, and there is no current flow thru the pack as a whole via other separate connections to the charger, then there is no current flow from the charger thru anything other than the one parallel group, so there will be no voltage change in any other group either.

To cause a voltage change, there has to be a current path, but the low resistances of the parallel interconnects will cause the current to flow thru the paralleled cell groups.

There isn't a way I can imagine to cause a failure of cells and interconnects that could cause a charger connected that way to charge other groups. The pack would have to be so damaged that it would not operate, because it would have to short those other groups to this one to be able to charge them.

 

[eMark]

Top balancing is most common only because of a BMS whose balancing ability is questionable. You could call a BMS a lazy man's way to "hopefully'' balance a pack of good cells. Even though we all know that it's inefficient taking a long time to balance a pack (if ever) of used cells of questionable quality. We have no way of knowing how good of a balancing job it's really doing or what it's suppose to do? However, that's not the case with bottom balancing.

The best way to balance 18650 and 21700 DIY builds with suspect unbalanced parallel groups is via bottom balancing. The downside is ii requires a RC Charger, patch cord, and some manual labor, but is far, far better than top balancing with a BMS. Balancing with a BMS that could take forever and there's no way of knowing how effective it is. Whereas bottom balancing should only take 15 to 30 minutes if done regularily after every discharge with a pack of used cells

 

[eMark]

Don't know if you even had time to read my previous reply to your question (with pace of reples) about a parallel group being affected by the voltage of another parallel group so will post again in cas other didn't read it ...

Once got distracted a couple days ago while charging one p-group suffering from self-discharge (30Q). The variance of the other 9 p-groups (10s3p - 30Q cells) was only 14mV (3.511V to 3.525V). The one with high self-discharge i was charging was at 3.487V at the beginning. By the time i remembered (78 and forgetful) and stopped the charge it was over 3.80V. This had no affect on the voltages of other 9 p-groups. Using my discharger i discharged it in line with the other 9 p-groups ... no harm done.

This is why bottom balancing of the parallel groups is far better than trusting a BMS to top balance p-groups of questionable quality/reliability. Will answer you post later with a few photos so OP gets a better idea how to bottom balance his DIY build before bulk charging.

Won't be able to post until tomorrow as i've got another preventiive maintenance repair project to finish today. My bottom balancing technique (w/photos) should be of some help to those with DIY builds using used cells or older(new) cells in storage for a few years (battery hookup) whether or not one decides to use a cheap or so-called 'smart' BMS.

 

[amberwolf]

This to me, is the foundation of why it is imprudent to mix different capacity cells within the same parallel group within the same battery enclosure. Perhaps you disagree and believe its ok to mix different amp-hour cells but that would belie your starting point of 'a correctly built battery pack'.

I don't disagree--I think it's a bad idea to use mismatched cells. You can find my posts around the forum suggesting to use well-matched cells used well-within their capabilities.

(unfortunately for most DIY people it's apparently just not practical / not within their budget of time/money, and some just don't care enough to do these things or don't believe it's necessary).


I suppose it depends on one's definition of "correctly", but for safety, longevity, etc, rather than cost or space/weight vs capacity/capability, a truly correctly built pack wouldn't be much like most of the ebike packs out there (which are generally built to optimize the other factors vs safety or longevity). Some OEM systems do it "correctly", but few packs you can buy anywhere do, and almost none of the DIY ones.

A "correctly" designed pack by the above definition (which is what I prefer) would use well-matched cells (QCd to be as close to the same characteristics as possible for the entire pack, so every cell is virtually identical in every characteristic and so discharges and charges identically). The cells would be protected by a system (BMS, etc) that limits usage to well within what their capabilities will be even when degraded significantly by age, etc, so that you may only be using 50-80% of their capability (capacity, current delivery, etc) when new, and not close to 100% of that even when degraded (especially when degraded as that's when cells begin to drift apart in characteristics more).

At the point the cells are being pushed near their limits, it would be time to retire that entire pack entirely from that usage, and apply it to a less strenuous one. (or re-cell it entirely if that's more cost-effective or desirable for whatever reason).

Large-EV packs appear to do these things in general. Small-EV (ebike & scooter) type packs are much less likely to especially the cheap ones--the less the battery costs, the less likely it is to be built this way, if for no other reason than the cost of matching every cell in a pack to every other, and the cost of warrantying the pack based on expected performance over time.

But also because the smaller the EV, the less space it has for a pack, and the more performance that is required of whatever pack it can carry, so the more that has to be asked of it. (higher performance cells can be used, but they cost more, and for some use cases just don't exist yet).

Some small-EV packs (by bigger / better OEMs) will be better made, but there are still limits to how far they can limit performance and still sell them for some particular price point.


FWIW, if I hadn't been lucky enough to get the EV-grade EIG cells I use on my trike and bike, I'd almost certainly be using well-oversized packs of recycled used mismatched cells myself, for budget reasons. (oversized so they still have more than enough capability to never be stressed even as they age further)

Do you think this battery pack is correctly built by mixing different capacity cells? Isn't that the subject we are discussing? Did you miss the OP's description of his cell selection?

No, I didn't miss it--the OP didn't actually *do* any cell selection (there is implication that some testing may have been done in order to use a "repacker" program (not linked) to do some form of "matching" (process otherwise undescribed by the OP).

I don't think it's correctly built, but neither are pretty much any of the packs people use around here, with very very few exceptions.

But what I said about how the groups will behave is still true, given the assumptions I stated within my posts. Whether those apply to this pack no one here (including the OP, AFAICT) can state--but it *can* be tested, even with this pack, given the test described in my post.

It pretty much isn't done in the industry. Homogeneity aka equivalent cell capacity and discharge is good. Heterogeneity is bad because cells that discharge at different rates will arrive at different voltages either singularly or in aggregation when underload and then after load they will charge one another to derive equivalent voltage.

If you mean that correct pack building isn't done in the industry, well, in general for cheap ebike packs, it's not. Neither is it for most of the DIY packs.

If you mean that mismatched cell pack building isn't done (by battery makers or DIY, either one), you haven't read enough of the pack building / rebuilding / troubleshooting threads around here. :lol:

It's certainly problematic to the end user (and those nearby when it causes a fire) for anyone to do it, but that doesnt' stop it from being done all the time, usually for cost reasons.

 

[amberwolf]

Top balancing is most common only because of a BMS whose balancing ability is questionable. You could call a BMS a lazy man's way to "hopefully'' balance a pack of good cells. Even though we all know that it's inefficient taking a long time to balance a pack (if ever) of used cells of questionable quality. We have no way of knowing how good of a balancing job it's really doing or what it's suppose to do? However, that's not the case with bottom balancing.

You're correct; I was just listing the methods as-used, for those that read this now or in future that have none of the information yet.

As noted previously, balancing a pack only brings cells to the same voltage, not the same capacity or capabilities. If they're not already matched, there's no way to fix them (without replacing them) to make them matched.

I have no direct testing experience with bottom balancing, but it may work better for unmatched cells to keep problems at bay, based on what it does.

The primary problem I'm aware of for the general public is there's no commonly-available cheap automated device to do it, so it depends on the end-user correctly performing the steps necessary every time. My experience helping people here on this forum is that a significant number of people (probably most) are unwilling to do this, and some are incapable of doing it, for whatever personal reasons they each have.

So for the typical end-user, we tend to be stuck with the common top-balancing BMS, of which many are poorly designed and/or manufactured.

Some can be convinced to do manual testing / fixing, and walked thru it, and to continue doing this as needed for the life of the pack, but not all.