First ebike build. I have some battery questions.

[transposon]

I plan on building my first ebike with a 1000W rear hub motor. I have 2 6s 10Ah multistar lipos but I want to build another pack for additional range.

I have access to lots of used 18650s from my IT department. Since I already have the equipment to test individual cell capacity and resistance, and I've already tested ~70 cells, I figured I would make a 14s8p pack from salvaged cells.

Since these cells are of varying age, resistance, and capacity, I thought that rather than use a BMS (with a relatively low balancing current) I should split this pack into 2 7s8p packs and use a hobby balance charger to charge them. I'm leaning toward the icharger x8 since it is portable and has a 2A/parallel pack balance current. All of the BMSs I have seen have 200mA or less balancing currents.

I plan on checking the corresponding parallel group voltages and if they are w/in 0.2V of each other, I will parallel charge the two 7s packs with a hobby balance board. If they are too far away in voltage, I will charge each separately. In either scenario, both packs should be at the same voltage when they are connected in series.

I was thinking of using low voltage alarms and skipping the BMSs all together.

Does this sound feasible or I am I just going to be riding on top of a ticking time bomb? Is this charging scheme better than just using a single, homemade pack with a 14s BMS? Any suggestions would be welcome.

Also, I realize that if my time is valuable, it is better to buy brand new cells, but I actually enjoy salvaging them.

Thanks

 

[BlueSeas]

Using a good hobby charger to recharge reduces the charging risk. I've used the Cell Lab products, and they work really well.

On discharge. Without a BMS, the risk is pushing a depleted cell negative. This will ruin the cell and make the pack unusable until that cell group is fixed. But risk...it's minimized because the cell has little potential remaining. Heat transfer to other cells a potential risk, but I don't see the low cell making a fire on its own. If you can run a few monitored cycle tests determining the right pack voltage to cutoff, it will reduce the potential issues.

The question is economic to some degree...is it worth the time to manage the risk or just buy a good pack, or build with new cells?

 

[john61ct]

A BMS is a collection of functionalities.

The balancing part is the least critical.

HVC / LVC, cell/group level vs pack

current OCP, thermal protections.

To use scrapped cells safely is a much bigger challenge than new ones.

 

[amberwolf]

On discharge. Without a BMS, the risk is pushing a depleted cell negative. This will ruin the cell and make the pack unusable until that cell group is fixed. But risk...it's minimized because the cell has little potential remaining. Heat transfer to other cells a potential risk, but I don't see the low cell making a fire on its own.

It shouldn't be a fire risk *during discharge*, but if you don't know that this has occured, and then you recharge it normally, the current thru that damaged cell *could* cause a fire.

Or it could recharge normally, but because that cell is basically dead, 0v, and potentially internally shorted or otherwise damaged so it doesn't take a charge, the extra voltage that should've gone into it will be distributed among the other cells, forcing them to overcharge, since the BMS wouldn't be there to stop charging and rebalance the cells. This overcharge will damage the other cells, to some degree, increasing risk on *those* cells.


Those are just two of the various possible risk situations.


As far as the overall risk of the pack doing the method the OP wants to do, it isn't really any more risky than the many people that do exactly that with RC LiPo, which has over the years been notoriously variable in quality even within a single 6s/etc pack, with little trouble (though there *are* fires, I suspect they're not caused by that process).

 

[BlueSeas]

The proposal by the original poster was to use a balance RC Charger to charge the pack one at a time in halves. I've only used one brand of RC charger, but if the balance wires are connected, it will not overcharge a cell. Further it displays cell voltage, presumably you would look at that beginning the charge cycle, and if one cell showed below minimum discharge voltage that would bear investigation before charging.

Think most top RC chargers are similar, basically both a BMS and charger in one unit. But they work together. For example, at top of charge, the current is lowered so the high cells can't overrun the balance shunts while the low cells continue to charge. They calculate internal resistance where wiring or cell anomalies show up as an outlier compared to other cells. This is way more advanced than typical Ebike battery charging.

Not mentioned, but a good idea would be to use one of the $50 cell charger/dischargers on stripped cells to determine capacity and function before building a pack. Any cell below say 60-70% of original capacity should be discarded.

Is it worth it? Over new cells. Depends on how much money you have, the quality of the recycled cells, and how much time you have to tinker with it. But I think this approach has a reasonable degree of safety if you know how to identify dysfunctional cells.

 

[BlueSeas]

Above...I said reasonable, not absolute safety. Until the chemists come up with a better battery, we all need to remember that these cells are intrinsically unsafe. No matter what we do to mitigate the problem.

Interesting article on just that....but I found it entertaining that the crux of the story was about a multi million (maybe billion) dollar company, and one of the conclusions was DIY was bad.

https://www.bicycling.com/bikes-gear/a28778383/electric-bike-explosion/

 

[john61ct]

these cells are intrinsically unsafe. No matter what we do to mitigate the problem.

Safe vs not is nonsense, no such b&w binary

% risk of a catastrophe, all greyscale.

Quality cells, brand new, good protective / monitoring / testing gear + knowledge to use it

can eliminate 99% of the risk compared to the usual carelessness with scrapped cells.

"Safe enough" is a judgement call.

Still not as safe as LFP or LTO, but their lower density makes them unusable for many use cases.

 

[john61ct]

Think most top RC chargers are similar, basically both a BMS and charger in one unit.

No

A BMS provides monitoring and protection of the pack while in use.

Balancing is not necessarily even included in their list of functions.

 

[transposon]

I will test each individual cell for capacity and resistance. I plan on making parallel groups so that they each have roughly the same capacity.


I would happily use a 14s pack with BMS if I could find a BMS with a decent balancing current.

 

[transposon]

Would it be possible to have a 14s battery with a BMS but have 2 external sets of 7s balance leads for balancing on a hobby charger if needed?

 

[john61ct]

I will test each individual cell for capacity and resistance. I plan on making parallel groups so that they each have roughly the same capacity.


I would happily use a 14s pack with BMS if I could find a BMS with a decent balancing current.

They exist but may cost more than your entire build.

Again, in your case, best to use BMS for **protective** functions only, not balancing.

Many cheap ones are sold with no balancing at all, but as long as you just keep your charge-stop setpoint below their start-balance voltage, same difference.

Then use one of the dozens of other options available to get your desired balancing functionality.

Some balancers go up to 10A,

current level does not depend on voltage delta, and

can be used

1. only during/after charging (top balancing), or

actively during usage, discharging

2. only as the bank approaches LVC (bottom balancing), or even

keep balancing at all SoC from top to bottom, in some cases getting a bit more range out of grossly mismatched cells.

Personally I recommend approach #1, but repeat, none of this balancing talk has to have anything to do with BMSs.

 

[john61ct]

Would it be possible to have a 14s battery with a BMS but have 2 external sets of 7s balance leads for balancing on a hobby charger if needed?

Only if the two 7S charge sources were completely isolated from each other as well as from their upstream input.

Which I've never seen in a hobby charger.

So, the solution is, just connect two 7S packs in series for discharge / use.

 

[BlueSeas]

I don't see any reason you could not bring out 2 sets of balance leads and a center tap from the pack and use a single RC charger to recharge the pack, 1/2 at a time.

Simultaneously charging with 2 RC chargers could be problematic. I believe John is right that the DC input and DC output of an RC charger commonly share a ground. However if powered by two unconnected sources of DC input, it would probably work. The best example, each charger connected to its own isolated car battery. Using 2 power supplies might work too, but they would require the AC side fully isolated from the DC side.

 

[BlueSeas]

I would happily use a 14s pack with BMS if I could find a BMS with a decent balancing current.

Don't get too hung up on balancing current.

There are 2 types of balancing. Active balancing is where high cells charge low cells during operation. This theoretically permits maximum utilization of a pack with cells that have mismatched capacity to better consume all the energy in the pack. Consider the extreme case where you have 14S, with 7 cells having 10Ah and 7 cells having 5Ah. During use, the 10Ah cells shuttle 2.5Ah to the 5Ah cells, and the whole pack has 7.5Ah of capacity. This *does* require higher balancing current. However, I don't know of any Ebike BMS doing this. Further, there are losses in the shuttling process and it must work again in reverse during charging. I've never had the chance to use this technique, and a bit skeptical of how well it works in real life.

The other type of balancing is called Passive Balancing. This technique only balances at top of charge. It works by shunting (discharging) high cells allowing low cells to reach charge termination voltage. So all cells end charge reach full capacity. In my previous example, using passive balancing the pack would be empty after 5Ah of use, the capacity of the lowest cell. However since there was no shuttling between cells, 5Ah of charge returns the pack to full charge.

Actually building a pack with a combination of 5Ah and 10Ah cell groups would be a poor design. Even using recycled batteries, you should be able to match cells much closer. If you find the cells measuring in groups of 1.75, 2 and 2.25 Ah, put one of each in a 3P string. The combined cell is going to provide almost 6Ah, and the "active" balancing is free. Tie the 3P cell groups into 14S and you have a pretty usable pack.

Some may argue this...but I will tell you that once in "balance", lithium cells tend to stay in balance. Very little adjustment is required by the passive balancing during charge. If this isn't the case, there is something wrong with a cell somewhere, usually one with with a high self discharge rate. This cell is failing and must be replaced. Other cases to cause rapid imbalance is a faulty BMS via the cell taps or someone having the bright idea to tap 12V off the pack for accessories. Don't just take my word...have a look at this:

https://www.orionbms.com/features/intelligent-cell-balancing/

These BMS systems are used usually in much larger packs/banks and get by with 200ma.

A couple more thoughts...make sure you fully charge all cells prior to building the pack. It isn't the job of a balancer to initially balance a pack. And leaving the cells at 4.2V balancing for a long period of time is a bad idea. In addition to the capacity test on scavenged cells, check for self discharge. After a full charge to an established voltage, wait say 48 hours and check cell voltage. Most cells should group into a tight range. Any low outliers should be discarded.

 

[transposon]

I would happily use a 14s pack with BMS if I could find a BMS with a decent balancing current.

Don't get too hung up on balancing current.

There are 2 types of balancing. Active balancing is where high cells charge low cells during operation. This theoretically permits maximum utilization of a pack with cells that have mismatched capacity to better consume all the energy in the pack. Consider the extreme case where you have 14S, with 7 cells having 10Ah and 7 cells having 5Ah. During use, the 10Ah cells shuttle 2.5Ah to the 5Ah cells, and the whole pack has 7.5Ah of capacity. This *does* require higher balancing current. However, I don't know of any Ebike BMS doing this. Further, there are losses in the shuttling process and it must work again in reverse during charging. I've never had the chance to use this technique, and a bit skeptical of how well it works in real life.

The other type of balancing is called Passive Balancing. This technique only balances at top of charge. It works by shunting (discharging) high cells allowing low cells to reach charge termination voltage. So all cells end charge reach full capacity. In my previous example, using passive balancing the pack would be empty after 5Ah of use, the capacity of the lowest cell. However since there was no shuttling between cells, 5Ah of charge returns the pack to full charge.

Actually building a pack with a combination of 5Ah and 10Ah cell groups would be a poor design. Even using recycled batteries, you should be able to match cells much closer. If you find the cells measuring in groups of 1.75, 2 and 2.25 Ah, put one of each in a 3P string. The combined cell is going to provide almost 6Ah, and the "active" balancing is free. Tie the 3P cell groups into 14S and you have a pretty usable pack.

Some may argue this...but I will tell you that once in "balance", lithium cells tend to stay in balance. Very little adjustment is required by the passive balancing during charge. If this isn't the case, there is something wrong with a cell somewhere, usually one with with a high self discharge rate. This cell is failing and must be replaced. Other cases to cause rapid imbalance is a faulty BMS via the cell taps or someone having the bright idea to tap 12V off the pack for accessories. Don't just take my word...have a look at this:

https://www.orionbms.com/features/intelligent-cell-balancing/

These BMS systems are used usually in much larger packs/banks and get by with 200ma.

A couple more thoughts...make sure you fully charge all cells prior to building the pack. It isn't the job of a balancer to initially balance a pack. And leaving the cells at 4.2V balancing for a long period of time is a bad idea. In addition to the capacity test on scavenged cells, check for self discharge. After a full charge to an established voltage, wait say 48 hours and check cell voltage. Most cells should group into a tight range. Any low outliers should be discarded.

Thanks so much for all of this insight. Most of my cell scavenging techniques I've learned from the DIY powerwall people. I use a 4 bay charger that test capacity. I also test for self dishcharge. What do you think is an acceptable voltage drop for a week in storage, I was thinking anything >= 4.1V.

 

[BlueSeas]

Offhand 4.1V sounds pretty good. But until I get my eBike and battery built, my experience is with large prismatic LFP cells with a nominal 3.2V rating and a charged voltage of 3.6V. A .1V drop would be reasonable. Still...I would focus on outliers. If you had 10 at 4.15 and 1 at 4.1V, would only use the 4.1V cell as a last resort. Likewise if you had 10 at 4.05 and one at 4.00V, then the 4.1V might be too restrictive.

 

[john61ct]

once in "balance", lithium cells tend to stay in balance.

No, not at all.

Certainly not with scavenged cells.

Only true with new cells from high quality manufacturers.

Which are very very rare, 99% of LI batteries sold to consumers through established retail is meh, at best.

Your experience is colored by your background - let me guess, quality LFP makers like Winston/Thundersky/Voltronix, CALB, GBS, Sinopoly and A123 (now Lithium Werks / Valence / Super B) ?

And in a low C-rate context, perhaps even coddling care, expensive protective gear like that Orion, etc?

DIYers doing propulsion at multiple C-rates both charge and discharge, get very different, higher imbalanced results.

And finally, after 10+ years of daily cycling even your top-notch coddled House bank will start getting increasingly imbalanced, and eventually require the more capable balancing gear and care that everyone else here requires from the start.

 

[BlueSeas]

once in "balance", lithium cells tend to stay in balance.

No, not at all.

I don't have experience with the 18650's. I have observed balance maintained in charge/discharge cycles of abused Thundersky cells, bloated 180Ah cells with capacity ranging from 40% - 60% of original new capacity. These were removed from a customer's vessel and relegated to the test bench.

Since I didn't know specifically about eBike batteries, I asked in this thread earlier:

https://endless-sphere.com/forums/viewtopic.php?f=14&t=102249&start=25

Presently the next to last post on page 2. Now that wasn't specifically recycled cells, but it was well used high demand cells. And the comment was..."I rebalance once a year".

So it seems like is a difference of opinion here on the forum.

Having worked with the Orion BMS engineers, if higher balance current was needed in older banks, I believe they would have that in their BMS, not make excuses in the discussion I linked to in this thread.

At this point...I will stick with the thought that any pack becoming rapidly unbalanced has other problems. A fast rebalance at the end of every cycle is just masking those problems.

 

[999zip999]

Look at DrAngle and read his post on building packs with recycled cells.

 

[john61ct]

any pack becoming rapidly unbalanced has other problems

Yes of course, if the owner can afford to, the preferred solution is to scrap the cells and buy new ones.

But in this forum's context, many if not most hobbyists are fooling around with free or cheap scrapped cells, and

with the right balancing gear, can get perhaps years more service out of them.

There is not just "one right way" to accomplish the widely varying objectives in completely different use case scenarios.

 

[BlueSeas]

Look at DrAngle and read his post on building packs with recycled cells.

Do you mean DrkAngel? I found the index thread and read a little. But it would take days to read it all. Back in 2015, he didn't believe in a need to have a BMS to rebalance at top of charge. Did his opinion change? If so I'll read some more......

 

[transposon]

OK, thanks to all of the replies here. After thinking about this for a while, I think I am going to build a 14s8p pack with one of the bluetooth BMSs and the salvaged cells. This will give me visual feedback on pack behavior. If I have trouble getting this pack to balance, I'll bite the bullet and buy new cells. Then the old cells will get delegated to the 2 7s split pack charging scenario. Worst case, I end up with 2 large 14s packs.

 

[999zip999]

The trick to using used cells is to check internal resistance and capacity. Get them to all the same balance before hooking up bms. Always make a bigger pack then you need and never find lvc never.