Li-ion battery charging/BMS confusion

[newb123]

I have been given 16 cells (2 packs, 8 each) of Panasonic Ebike batteries.

The individual cells information is as follows:

CGR 26650B Li-ion MH12210

Each pack came with its dedicated BMS BLA4AE0

So, at 4.2 volts each x 8 gives total charge at about 32 - 33 volts. It could be less than that if its a series-parrallel connection, will have to check on the config

My ebike motor is 36 volts > so, I broke open the 2 panasonic battery packs and

charged each individual cell with an ebay buck CC/CV converter. One cell was flat but the

rest charged to 4.1 volts (CGR 26650B data sheet gave around this value at CC 3 amps)

so my question is this: can the 32 volt BMS be used to monitor, for ex. a bank of 5

li-ion cells instead of 8? I would be using 2 bms, as I dont have a single BMS for 36 volts.

and each BMS would monitor 5 cells (20 volts) for a total of 40 volts. i could just

get an ebay 36 volt bms I guess, but I noticed that these ebay Chinese BMS have

different voltage options, so could a 32 volt BMS monitor 20 volts? Just leave some battery connections floating?

 

[amberwolf]

First, questions:

What is this going to power?

How much current does the device require at peak?

How much current does the device require continuously?

What is the peak current capability of the cell model you have?

What is the continous current capability of the cell model you have?

What were the starting voltages of each of the cells you have?


The first five questions help you determine how many parallel cells you will need in each group, to be able to run the device without damaging the cells, or heating htem up or stressing them, or causing more than a little voltage sag under load.

The sixth question helps you determine which of the cells you have is safe to actually use. If any were below 3v, I would not use them at all. If they were really dead, close to 0V, they're really unsafe to reuse. You can actually cause a fire by recharging (or discharging after the recharge) a cell that has discharged too far, depending on the internal condition of the cell. Since you can't see that condition, there's no way to determine if it's safe, so it's only safe to not use such cells at all.


If you need more info on how to build a pack, test cells, etc., there are literally hundreds of threads with 18650 in the title that have that. Some of the titles will show you they do or don't have what you need, but other threads without good titles may still have what you' want in them, so it may be tedious going thru the threads to find all the good info. :/ Even though your cells are 26650, there's much more info about 18650 packs, which all still applies to your cells as they're still cylindrical and designed basically the same way.


Some BMSs can be used with less cells than designed, some can't. You may have to "disable" the inputs to teh unused channels one way or another, but how that's done on any specific BMS I don't know. YOu'll likely ahve to experiment.

 

[newb123]

What is this going to power? 36 volt brushless DC Ebike motor

How much current does the device require at peak? About 20 Amps (measured with amp meter, I dont have a spec sheet on this motor/controller, its quite old, but clamp meter while starting, see about 20 amps)

How much current does the device require continuously? around 10 - 15 Amps (again, measured with amp meter when riding)

What is the peak current capability of the cell model you have? Pulse Current = 31 Amps

What is the continous current capability of the cell model you have? 20 Amps continuous discharge

What were the starting voltages of each of the cells you have? Nominal Voltage 3.6 volts Cut Off Voltage 2.5 volts

I had discharged these batteries with intent to dispose of them, but recharged them with rapid CC/CV charge at 4.1 volts 3 amps, they did get warm, not hot and now holding charge at 4 volts. Only one cell did not charge at all, I suspect after several charge cycles it was the weakest link and failed...? So, when you say parallel, you mean to get the min. peak amp requirement? I was going to series connect 10 of them for 40 volts, but for more amperage I should parrallel them, like 10 in series, paralleled to another 10 in series?

 

[amberwolf]

How much current does the device require at peak? About 20 Amps (measured with amp meter, I dont have a spec sheet on this motor/controller, its quite old, but clamp meter while starting, see about 20 amps)

How much current does the device require continuously? around 10 - 15 Amps (again, measured with amp meter when riding)

What is the peak current capability of the cell model you have? Pulse Current = 31 Amps

What is the continous current capability of the cell model you have? 20 Amps continuous discharge

Based on that, and that these are used cells (possibly well-used), I'd recommend at least 3 or 4 cells in parallel for each group, to reduce the strain on them. At a guess they are probably no longer capable of their full delivery of either current or capacity.

You can test them for how much voltage sag each has at the load they'll see on your bike, at different states of charge. If they don't sag much then maybe they can be used with less cells in parallel--but if they sag a lot, then more parallel will help.



Now the next thing is, how much range do you need?

Which leads to, how fast do you go, and what is the terrain and wind like?

These will help you figure out how many cells in parallel you need to get the range you want (a separate thing from how many are needed to reduce cell strain).

If you're riding at 20MPH max, then it'll probably take around 15-20Wh/mile depending on how much pedalling you do, how many stops/starts you make, headwinds, hills, etc.

Each cell has about 11wh when new, let's round that down to 10wh (might actually be less). So if you have 10 cells, that's 100wh available. If you need 20wh/mile, then the most you'd get out of it, running it all the way down to nothing (which will age the pack faster), is about 5 miles. If you run into headwinds, hills, lots of starts/stops, that increase the usage, then it'll decrease teh range.

If you have 20 cells, that potentially doubles the range.

What were the starting voltages of each of the cells you have? Nominal Voltage 3.6 volts Cut Off Voltage 2.5 volts

Not sure, but I don't think that's the answer to the question.

It sounds like you're giving the spec sheet value, which is not what's needed. (though, so you know, that LVC of 2.5v is the ABSOLUTE lowest the cell can be allowed to go to without definite permanent damage to it. You don't ever want to really run them that low in practice; it will wear them out much faster than necessary. 3v would be the worst-case LVC I'd ever use, and I'd really keep it up over 3.2-3.3v or more.)

The question is, what were the *actual voltages* of each cell, before you started doing anything with them?

And how low is the lowest any of them got during your discharge of them?

This is important, because cells that were too low are not safe to use, regardless of how they appear to perform right now--you can't predict what will happen later.

I had discharged these batteries with intent to dispose of them, but recharged them with rapid CC/CV charge at 4.1 volts 3 amps, they did get warm, not hot and now holding charge at 4 volts. Only one cell did not charge at all, I suspect after several charge cycles it was the weakest link and failed...?

Could be. Keep in mind that rapid charging cells is harder on them, especially when they are older, and most especially when they are at a very low state of charge, near or at LVC--if any were below that to start with, it's even harder on them.

So, when you say parallel, you mean to get the min. peak amp requirement? I was going to series connect 10 of them for 40 volts, but for more amperage I should parrallel them, like 10 in series, paralleled to another 10 in series?

Yes, though most poeple put the parallel groups first, then series those groups. Makes it a lot easier to hook up to a BMS (or balance leads if using an RC charger, etc).

There's a few threads about "parallle first or series first", if you want to read up on reasoning behind each way of doing it.

 

[newb123]

I'd recommend at least 3 or 4 cells in parallel for each group What would such a config look like? 10 cells in series, and then another bank of 10 in series and then another, for a total cell count of 30 cells, with cross connections for the 3 cells in parallel?

oh I see you answered it here:

Yes, though most poeple put the parallel groups first, then series those groups. Makes it a lot easier to hook up to a BMS (or balance leads if using an RC charger, etc).

so I still dont have an image of how that would work...will have to look it up. your saying like have a group of 5 in parallel series to another group of 5 in parallel and then series that to other groups of 5 in parallel, for ex.?

You can test them for how much voltage sag each has at the load they'll see on your bike, at different states of charge.
What is sag, and how to I perform such a test?

Now the next thing is, how much range do you need?

Which leads to, how fast do you go, and what is the terrain and wind like? Ride scape is relatively flat, controller limits to about 28 mph max, would like 15 miles range but according to your data I would need some serious parralleling done to get that. My ebike motor is rated at 400 watts / 36 volts so that gives me 11 amps of cont. current I guess. Ive seen it jump to twice that on start up or hills, though.

And how low is the lowest any of them got during your discharge of them?
0 volts was going to dispose them so I discharged them through a incandescent lamp. Interestingly, the undamaged ones charged to 4 volts and I discharged one through a 100 ohm resistor and led and its still discharging ( need to find another filament lamp or parallel a bunch of leds). I had shorted one of the nickel plates to the battery casing when I cut it and some of the battery wrap had come off. some heat, sparks but no explosion, lesson learned. So I will dispose of it because since I shorted it Im sure its not safe.

It seems??? possibly one benefit of getting these used, if they are still good, is the nickel tabs still on them, I just have to cut them as close as I can to the poles and not leave an overhang. Dont need a spot welder, just jumper wires.

the other pack of 8, have not discharged them, their total voltage is 25 volts. there are 8 @ 3.1 volts each, seem to be in a series config to get that sum

Also, on the BMS there are 4 FETs and 2 thick Red wires 2 Thick Black wires, for the charging and output, but there are 2 small wires, a Yellow and White wire going to a small connector, and labeled T and S. Any idea what those might be?

Thanks very much for your detailed explanation; still learning

 

[amberwolf]

What would such a config look like? 10 cells in series, and then another bank of 10 in series and then another, for a total cell count of 30 cells, with cross connections for the 3 cells in parallel?

Yes, although I"d wire up 10 sets of 3 cells in parallel, then series those 10 sets (as if they were a single cell each, that was 3x the capacity/capability of the original).

Look up thread with 18650 in the title, and you'llf ind a lot of pack builds that explain and show what you need.

What is sag, and how to I perform such a test?

Sag is how much voltage drop there is under a load. (load being something that draws current from the battery; more current - more sag).

With a single cell it's hard to do a test because it's hard to get something low enough resistance to usefully do a load test at those voltages. But you can put them all in series and test wth a load across the whole pack, and then measure voltage across each cell with and without that load. The difference in voltage is the sag.

More sag = worse performance.

You can look up various methods of load testing under those words as search terms Load* test* ; light bulbs, heater elements, etc. it just has to be something that draws teh same kind of current as your bike will under it's varied loading conditions.

Ride scape is relatively flat, controller limits to about 28 mph max, would like 15 miles range but according to your data I would need some serious parralleling done to get that.

Depends on the performance you actually get out of the system--the Wh/mile.

If you have a bike computer that measures speed and mileage, and a wattmeter (like the RC wattmeters), you can determe the actual wh/ mile for your setup and usage. Otherwise, you have to guesstimate by searching for similar systems and usages, and borrowing their performance stats. Or use Grin Tech's trip simulator with a system that is approximately like yours if yours isn't listed in their dropdowns.

My ebike motor is rated at 400 watts / 36 volts so that gives me 11 amps of cont. current I guess. Ive seen it jump to twice that on start up or hills, though.

Sounds normal. Keep in mind that 400 watts by itself won't get you to 28mph, it takes around 750watts to do that under normal condtions (no tailwinds, not downhill). So if your system will really reach 28MPH, it's likely drawing more than 400watts, closer to 750 watts, around 20A+. So the battery will need to be able to sustain that.

ALso, since ti takes more power, it takes mroe wh/mile, to go that fast. IIRC, power required just about doubles for each 10MPH above 20. You can check this out at http://ebikes.ca/simulator

0 volts

If they were really discharged to 0v (anything below the minimum on the spec sheet, really), they are no longer safe to use, even if they successfully recharged (*especially* if they were recharged quickly from that state, rather than just trickling at a few mA).

They have been physically damaged internally, and it's potentially asking for a fire at some random point.

If you're riding it, it could burn your bike, and you.

If it's just leaning up against a building, even not in use, it could set fire to the building, and kill anyone that doesnt' get out.

If you're charging it in your house, it could burn your house down, and kill anyone inside.

Battery fires are serious stuff. To get an idea of the problem, you can look up
batter* fire*
or
batter* burn*
here on ES for posts about it.


It might not happen--they could be fine. But you won't know until the fire starts that it's going to. :/

the other pack of 8, have not discharged them, their total voltage is 25 volts. there are 8 @ 3.1 volts each, seem to be in a series config to get that sum

If they are all equal voltages and never were below that 3.1v, they are still safe to use. But it's only 8 cells...so you'd need to get more of the same cells to run your bike (assuming your controller's LVC is the usual 30v-ish for a 36v pack).

Also, on the BMS there are 4 FETs and 2 thick Red wires 2 Thick Black wires, for the charging and output, but there are 2 small wires, a Yellow and White wire going to a small connector, and labeled T and S. Any idea what those might be?

They're probably proprietary communications to the bike they came from. Most OEM bikes have closed systems whose parts will not work outside the system, and will not work in the system unless all the system parts are present adn working. You'd have to know what bike they were from, and look around to see if anyone has done a "dissection" or rebuild or customization of that bike, to find out if it's really proprietary or not.

At a guess, the T is probably a thermal sensor (likely a 10K NTC, they're really common), and the S is probably serial data, and each of those probably uses battery negative as ground. But without knowing the protocol and commands and responses expected, there's no way to know what to send or expect to receive on the serial bus.

You probably cant' really use the existing BMSs without the bike they came from, but you can try.

 

[newb123]

ok the discharged batteries arent worth the effort if they are that dangerous.

I will discharge them to 0 volts and dispose of them

For the sag and other topics, I will look into it

You are correct, it was a 750 watt motor, dont know where I got the 400 from

 

[newb123]

I have opened the other 7 cell battery pack and each one measured 3.7 volts except one, that was at 3.6
I will keep them at 3.7 until I get more