Can I run lithium ion batteries in parallel safely?

Nazo

Newly Enlightened
Joined
Oct 27, 2015
Messages
5
So for various tasks (more on that in a moment) I've been wanting to be able to run lithium-ion batteries in parallel for a time. However, it has been my understanding that ideally you need to impedance match cells when doing this. Certainly real companies (obviously excluding cheap Chinese generics) can do this as they have a large scale with a lot of stock and thus can just check them all and put together matches. A DIY home user like myself who finds it to be exceptionally costly to buy lithium ion batteries (I generally just buy two at a time) obviously can't do this. In fact, the older ones have to be treated as not matching even if the impedance comes up the same as they've been used and thus have more cycles on them. Now, one thing I would swear I've read in the past is that the primary reason for this matching is to keep one from simply outputting its own excess voltage into the other one. If I have since read correctly in various searches (basically running across things by accident here while looking for other things) it really sounds like once they're voltage matched, essentially it's less of an issue since they will charge and discharge together.

I presume this would mean it would be more wear on the weaker of the two cells though, correct? The question then is if this is a big issue here? The key thing that finally got me to break down and start looking into this after all of this time (I've been messing with lithium ion for quite some time now but always using one cell or two in series for simplicity) is I dragged out my netbook after a really long hiatus (I wouldn't even care to guess how long it has been since I last properly charged it) and discovered that the battery pack is essentially dying. I didn't feel like being bothered to rig up a proper test mechanism, so I plugged it into a cheap low end Kill-A-Watt model and removed the battery pack (so it's running a power supply unit which is obviously not 100% efficient but it's not charging a battery. Thus my wattage readings include a bit of loss due to inefficiencies, but I can at least assume they are within a reasonable percentage of the tested results.) The absolute maximum reading I ever got was 19 watts for a split second and I only saw that once. Generally speaking it hovered between 17 and 18 (this thing doesn't do decimals, so it rounds and thus may have been 17.5 or something -- I could do current, but that gets messier since it's testing on the AC end) so I treat it as being 18 at the most. Average during heavy real usage after everything settled down was around 15 or so, give or take (really once it fully settles it goes down to around 13, but I assume that if I've pulled it out I'll be at least doing a bit with harddrive access and such going on. It's probably 14 or less realistically speaking, but I caught it showing 15 a lot during a quick test and I'm going with that.) This means absolute worst case scenario for any sustained usage is roughly (18/6)/2 = 1.5a per cell (absolute worst case scenario being when the battery is critically low and that chip is just about to clamp down and shut it off forcibly, so 3.0V x 2. It may not even turn on at that low of a voltage though, I'm not sure.) More realistically I'd really be looking at a much more typical range averaging around (15/7.2)/2 = 1.04a per cell. In each case well below 1C for even some really cheapo batteries. However, I don't plan to go cheapo here.

I pulled the battery pack apart out of curiosity and I guess one of the cells (or one of the packs of cells in parallel) must have been damaged due to a too low voltage when sitting around for too long (in particular, probably when I forgot to pull the pack out for a long time. It seems the netbook drains a trace amount of power from the battery while turned off. It's very little, but it probably had months to do this and then even once the protection chip presumably shut off the circuit it had plenty of time to slowly self-drain that last tiny bit after the cutoff point before damage occurs.) I noticed of the two sets of cells, one set was of a considerably lower voltage even after charging for a while. Assuming this thing isn't horribly made (I don't know my chips enough to really tell just from looking, but I believe that there are enough chips for this) it should be balance charging (are there any laptop batteries made in the last like ten years or so which do not balance charge? Lol.) Basically I think what happened is it sat around so long that eventually even after the voltage cutoff point the cells self-drained very very slowly over time until eventually at least one of them was dangerously low and was damaged. Maybe two weeks ago (three at the most) I had actually completely charged the battery pack overnight in the system and then taken it out of the system and left it out this time, then I found that when I tried to use it last night it wouldn't even turn on until I plugged it in (at which point it indicated that the battery was extremely low,) so one way or another the battery pack does appear to be damaged.

This comes to my biggest reason for wanting to do this. I'm basically obsessed with Panasonic's NCR18650 cells. These are designed to drain down to 2.5V instead of 3.0V. Now, of course, few things can handle that. In most cases everything below 3V should be considered useless. Especially when using protection circuits designed to stop at 3V anyway. Looking at their discharge charts though, the NCR18650s are still quite impressive when stopping at 3V. (It's difficult to exactly visualize, but I'd guesstimate it is roughly in the area of 2800mAh. It's definitely more than 2600mAh as far as I can tell at the very least, so at a bare minimum this means they're at least equal to the competition. As I'm sure you people in particular will know, it's rare to squeeze more than 2600mAh out of a 4.2V cell and to get more you need to go up to 4.35V with one of the newer tech higher voltage cells that cellphones do love so very much.) The real point here though is that the NCR18650 has more tolerance for applications where one may not necessarily be able to watch them as carefully. For example, a netbook sitting around for a very long time without charging. Once the protection circuit shuts it off around 3V or so, a normal lithium ion only has to self-drain down to 2.8V or so before damage begins occurring (and at that voltage range they drop much more quickly.) These would have to drop to around 2.4V or lower (I forget where the actual point of damage was on these) and they have a considerable capacity from 3V down to 2.5V as well. Thus even if I screw up again and forget to charge it for a long time, the actual cells shouldn't be damaged unless I basically just never use the thing again (or have other issues like the harddrive dying from old age.) Anyway, my point here is that you just can't do this no matter what battery packs you find made for any specific device on Amazon or eBay or whatever.

Now, I want to be clear that I'm 100% aware that I can just buy another battery pack. In fact, I did buy an old used one as a backup in case I screw things up. However, my goal here is just to make one that is more durable in general. This should produce a higher capacity along the way as a nice little bonus though. (Like I said, they're probably closer to 2800mAh or so. Maybe slightly more even. That means my 5200mAh battery pack could theoretically become a 5600mAh one. Not a huge amount, but on this thing that really would make a difference. Just a bonus though, nothing more. I need that low voltage safety net more than anything else I think.)
 

SilverFox

Flashaholic
Joined
Jan 19, 2003
Messages
12,449
Location
Bellingham WA
Hello Nazo,

Welcome to CPF.

Running cells and charging cells in parallel is not a problem even if they have different capacities and impedance. The problems arise when you hook up cells in series.

If your battery pack has a maximum voltage of 4.2 volts you can just use it. If the voltage is higher, then you need to pay attention to cell matching.

Tom
 

Nazo

Newly Enlightened
Joined
Oct 27, 2015
Messages
5
Running cells and charging cells in parallel is not a problem even if they have different capacities and impedance. The problems arise when you hook up cells in series.
Are you sure that's not backwards? Why would running them in series be an issue? And surely you need to at least semi-match batteries to run in parallel or one will end up being strained far harder than the other then wouldn't it? Surely they need to at least be within a decent percentage tolerance (10%? 20%? Just something reasonable anyway.) Right?

If your battery pack has a maximum voltage of 4.2 volts you can just use it. If the voltage is higher, then you need to pay attention to cell matching.
Well yeah, obviously if you try to charge a 4.2V battery to 4.35V it would be... bad... But if you charge a 4.35V battery to 4.2V it's pretty harmless. (Probably actually good for it since it has much less voltage-related stress, but obviously you lose a very significant amount of capacity since a 4.35V cell only has all that extra capacity because of that higher voltage.)
 

hiuintahs

Flashlight Enthusiast
Joined
Sep 12, 2006
Messages
1,840
Location
Utah
Are you sure that's not backwards? Why would running them in series be an issue? And surely you need to at least semi-match batteries to run in parallel or one will end up being strained far harder than the other then wouldn't it? Surely they need to at least be within a decent percentage tolerance (10%? 20%? Just something reasonable anyway.) Right?
Initially, you'd want to match up two parallel lithium ion batteries when connecting them in parallel. But after they are connected they will equalize. From then on they should be the same. I don't think you need to worry about cell impedance when you are talking batteries of the same part number. Single layer lithium ion in parallel is 4.2v and charging them shouldn't be a problem.

.......Running cells and charging cells in parallel is not a problem even if they have different capacities and impedance. The problems arise when you hook up cells in series.
If your battery pack has a maximum voltage of 4.2 volts you can just use it. If the voltage is higher, then you need to pay attention to cell matching.
Tom

As far as charging batteries that are in series, I'm no expert, but I do believe you may have to have a balance type of charger so as to get equalization. From what I've seen SilverFox knows his stuff with batteries. I think what SilverFox was eluding to in a series battery pack is that the termination voltage would be 8.4v or 12.6v depending on how many were in series. That is where the charger has to have an input of the voltage tap between the batteries.
 
Last edited:

SilverFox

Flashaholic
Joined
Jan 19, 2003
Messages
12,449
Location
Bellingham WA
Hello Nazo,

Parallel use is very forgiving. For example look at jump starting a car. I doubt that the jumper battery has the same impedance or capacity as the car battery that is low. You hook them together and as long as the car battery is not shorted you start the car and go.

You can also use cells in series that are mismatched, but run the risk of over discharging the weaker cell(s) and pushing it into reverse polarity. Charging a battery made up of mismatching cells is another matter. A lot of the problems occur as a result of mismatched capacities. One or more of the cells in series get full first, then are driven to overcharge as the rest of the cells get filled up.

Everything works best with matched cells.

Tom
 

sidecross

Flashlight Enthusiast
Joined
Jul 29, 2012
Messages
1,369
Hello Nazo,

Parallel use is very forgiving. For example look at jump starting a car. I doubt that the jumper battery has the same impedance or capacity as the car battery that is low. You hook them together and as long as the car battery is not shorted you start the car and go.

You can also use cells in series that are mismatched, but run the risk of over discharging the weaker cell(s) and pushing it into reverse polarity. Charging a battery made up of mismatching cells is another matter. A lot of the problems occur as a result of mismatched capacities. One or more of the cells in series get full first, then are driven to overcharge as the rest of the cells get filled up.

Everything works best with matched cells.

Tom

Bingo! :)
 

TinderBox (UK)

Flashlight Enthusiast
Joined
Jan 14, 2006
Messages
3,488
Location
England, United Kingdom
I asked this the other day, I bought an empty diy usb power bank, I just need to put 8x 18650 cells inside, but i only have 4x2200mah and 4x2400mah but they are connected in parallel so i though it would be safe, Anyway SilverFox gave me the thumbs-up :)

John.
 

Nazo

Newly Enlightened
Joined
Oct 27, 2015
Messages
5
Initially, you'd want to match up two parallel lithium ion batteries when connecting them in parallel. But after they are connected they will equalize. From then on they should be the same. I don't think you need to worry about cell impedance when you are talking batteries of the same part number. Single layer lithium ion in parallel is 4.2v and charging them shouldn't be a problem.
Darn, I meant to ask. Should I do something like connecting the two positives to each other with a resistor for a while or something to help them equalize slowly? Anyway, I get that they'll equalize voltage-wise, I just worried about the stress of two batteries taking things at potentially slightly different rates. I'm guessing people have been doing this for a long time though. From the sound of it, perhaps the problem here is that they've been doing it so much with so little thought that no one even thinks to mention it?

As far as charging batteries that are in series, I'm no expert, but I do believe you may have to have a balance type of charger so as to get equalization. From what I've seen SilverFox knows his stuff with batteries. I think what SilverFox was eluding to in a series battery pack is that the termination voltage would be 8.4v or 12.6v depending on how many were in series. That is where the charger has to have an input of the voltage tap between the batteries.
Balance chargers just charge each cell (or cell pack) individually since they aren't all the same and thus would be worn significantly by the charge process itself if charging in series since it would otherwise essentially be treating them all as being exactly the same and thus could run too much voltage in too fast sometimes.

Parallel use is very forgiving. For example look at jump starting a car. I doubt that the jumper battery has the same impedance or capacity as the car battery that is low. You hook them together and as long as the car battery is not shorted you start the car and go.
Well, that's not really very indicative of very much. In the case of jump starting, you want to get that voltage up a little bit anyway. I'll admit it's more about capacity, but that too is kind of part of it since it's still going to be able to provide a decent number of amps still anyway but the voltage will drop too low under load when the battery is too low to handle things anyway (so it's kind of all of it together I guess, but anyway, the point is that discharge from the good battery into the weak one will happen and is expected and necessary.)

You can also use cells in series that are mismatched, but run the risk of over discharging the weaker cell(s) and pushing it into reverse polarity. Charging a battery made up of mismatching cells is another matter. A lot of the problems occur as a result of mismatched capacities. One or more of the cells in series get full first, then are driven to overcharge as the rest of the cells get filled up.
Your saying this is the part that is confusing me. How could this happen in series then? In series it should essentially not know or matter what capacity each one has except insofar as the total voltage drops as each voltage drops. Do you have anything that goes into more detail about what you're talking about here and what I need to look out for? The thing is, a lot of things really do need to run in series since you can obviously only get up to 4.2V (or 4.35V for the batteries that use that obviously) to get a higher voltage and there are a lot of times this is necessary. I already have one application that requires this in fact and was thinking of a few others I may someday need.
 
Last edited:

more_vampires

Flashlight Enthusiast
Joined
Nov 20, 2014
Messages
3,475
You can also use cells in series that are mismatched, but run the risk of over discharging the weaker cell(s) and pushing it into reverse polarity. Charging a battery made up of mismatching cells is another matter. A lot of the problems occur as a result of mismatched capacities. One or more of the cells in series get full first, then are driven to overcharge as the rest of the cells get filled up.

Your saying this is the part that is confusing me. How could this happen in series then? In series it should essentially not know or matter what capacity each one has except insofar as the total voltage drops as each voltage drops. Do you have anything that goes into more detail about what you're talking about here and what I need to look out for?

In series, when one battery runs out before the other(s) that's when you get fireworks. This is called a "reverse charge" scenario, resulting in a "vent with flame" event. This is why you really shouldn't run say an 18650 and a 18350 in series. The 18350 will always run out first and it'll blow. Basically.

In parallel, I guess you could say that 18650x4 will keep each other in check.
 

SilverFox

Flashaholic
Joined
Jan 19, 2003
Messages
12,449
Location
Bellingham WA
Hello Nazo,

Let's look at a couple of examples...

Suppose you have 3 cells. One has a capacity of 2500 mAh, one has 2000 mAh, and the third has 1500 mAh. With Li-Ion chemistry there is a reasonable correlation between voltage and remaining capacity although it isn't a perfect correlation. All cells start at 4.2 volts.

If you hook up the cells in parallel you end up with a battery that has 4.2 volts and a capacity of 6000 mAh. As the capacity is used up, the voltage will remain similar between the cells because they are hooked up in parallel. The 2500 mAh cell will carry most of the load and the 1000 mAh cell will carry the least amount, but after 3000 mAh is used up all the cells will show similar voltages.

Now we hook the cells in series. We now end up with a battery that has 12.6 volts and a realistic capacity of 1000 mAh. Unfortunately when the 1000 mAh is used up the other two cells will still have 50% or over of their capacity left. This gives a higher voltage for the battery pack and as a result the pack is continued to be used driving the 1000 mAh cell into reverse polarity.

If your device had the ability to count the mAh used and shut off after 1000 mAh were used, you would minimize the risk using these cells in series.

So far we have been discussing using the cells. Charging is another matter.

Tom
 

Nazo

Newly Enlightened
Joined
Oct 27, 2015
Messages
5
Ok, so I got the batteries in today and decided to rig up a quick test. I hooked up just one on each side of the battery pack's circuit board after charging them completely. I had trouble at first with the wires coming loose (yeah, shouldn't have used cheap alligator clips, but it was meant to be a quick test) and the charging light on the netbook would start steady then go to blinking after a few moments. I connected them a bit more directly so nothing could come loose and it never stops going steady now (no blinking.) However, it will not turn on even if I leave it a while. I wanted to be sure nothing was loose (cheap battery clips) so I used a voltage tester on the wires leading to the circuit board. While charging, one side looked very normal, but the other side read 4.25V. Kind of suspiciously high really... Meanwhile, the batteries I pulled out of it originally haven't dropped significantly in charge. Perhaps I misjudged.


Or just buy protected cells as long as they can handle your current requirement.

John.
What does that have to do with any of this? For the record though, I have a few protection circuits I use with things like cheap lanterns where there is no circuit in the device itself to prevent it from over-discharging the battery otherwise. I don't think I have any applications that use really high currents (and I've already described the currents the netbook uses. That's actually the highest of anything I have.) The battery pack has its own protection circuit and more though. (I suspect it actually handles the charging even.)

Now we hook the cells in series. We now end up with a battery that has 12.6 volts and a realistic capacity of 1000 mAh. Unfortunately when the 1000 mAh is used up the other two cells will still have 50% or over of their capacity left. This gives a higher voltage for the battery pack and as a result the pack is continued to be used driving the 1000 mAh cell into reverse polarity.
I'm not really getting this reverse polarity part though. As others in series drop, they should just drop and be lower. Even if one set drops to, say, 3.0V, it should just be, for example, 3.0 + 3.6 = 6.6, not -3.0 + 3.6 = 0.6V. Why do you say it reverses polarity? At what point does this polarity suddenly reverse?

I still think it sounds like you're getting the two switched. If by "reverse polarity" you mean that the higher voltage one would discharge into the lower voltage one, that is what should happen in parallel, not in series. You keep saying it that way, so I assume I'm missing something here, but I really haven't a clue what it must apparently be and it's this which I'm asking here.
 
Last edited:

Rick NJ

Newly Enlightened
Joined
Feb 8, 2013
Messages
154
...
...
I still think it sounds like you're getting the two switched. If by "reverse polarity" you mean that the higher voltage one would discharge into the lower voltage one, that is what should happen in parallel, not in series. You keep saying it that way, so I assume I'm missing something here, but I really haven't a clue what it must apparently be and it's this which I'm asking here.

SilverFox is right. To make it easier to see, picture just two in series and assume Cell 1 is less able than Cell 2

[look at below pictorially]

(wire to load ground) -[Cell 1]+ -[Cell 2]+ (wire to load Vcc)

As you can see, Cell2's positive is connected to Cell1's minus via the load. Mentally take the load out for the time being, Cell1 and cell2 are connected in reverse: + to - and - to +. Now consider the load as just a "poor conducting wire" since it chew up a lot of the power (ie:loose some voltage), the connections between the two cells are + to - at both ends.

Once cell 1 is sufficiently depleted, if cell 2 has sufficient power (Voltage), its positive is connected to cell 1's minus, so charge is getting put into cell 1 in reverse polarity. If the load drops the voltage significantly, it may be ok, but if the load doesn't drop it by so much, problem is there.


Now expand that to three cells:

(wire to load ground) -[Cell 1]+ -[Cell 2]+ -[Cell 3]+ (wire to load Vcc)


Now if you have a third cell in the picture, and cell1 is still the worst one: mentally combine cell2 and cell3 as a single composite cell.

(wire to load ground) -[Cell 1]+ -[ComboCell 2&3]+ (wire to load Vcc)

This composite cell has the combined voltage of cell2 and cell3. You can almost be certain that sum of cell2 and cell3 will be a higher voltage than the lone bad cell (cell1) so negative charging is almost sure to happen.

Such arrangement could makes a nice boom if the cells are unprotected. Once a cell gone sufficiently off balance from others... kaboom is now a possibility.
 
Last edited:

Gauss163

Flashlight Enthusiast
Joined
Oct 20, 2013
Messages
1,604
Location
USA
If your goal is to rebuild the laptop battery then that will probably not work, for a variety of reasons. First, most modern BMSs have firmware sealed with passwords, and you cannot even reinitialize it without the password. Second, many have programmed tables that depend on the cell type (e.g. chemistry). If you change the cells without updating the tables then that may greatly confuse the BMS, e.g. the fuel gauge may be wildly inaccurate.
 

uralshina

Newly Enlightened
Joined
Dec 7, 2017
Messages
1
I have hooked up a 4000mah cell(Samsung phone battery).to a GPS 900mah battery on a GPS.unit seems to charge and run just fine
 

vicv

Flashlight Enthusiast
Joined
Mar 22, 2013
Messages
2,790
Location
Southern Ontario
I usually charge my cells in parallel. I'll even charge a 3200mah cell with an 1100mah one. They're will always charge to full equally. You can put a 16340 in there and it will make no difference. They will always charge and discharge together as one cell basically
 

Gauss163

Flashlight Enthusiast
Joined
Oct 20, 2013
Messages
1,604
Location
USA
[...] Do you have anything that goes into more detail about what you're talking about here and what I need to look out for? [...]

You can find much enlightening information in the excerpts I posted here from an interesting study of the effects of severe overdischarging and reverse charging of Li-ion cells.
 

Overclocker

Flashlight Enthusiast
Joined
Aug 13, 2005
Messages
1,585
Location
Philippines
xWEWAun.jpg


that's 13-series 4-parallel Samsung 30Q

you could even parallel another 13-series pack with this one, even with different cells i.e. different discharge curves, different IR, etc. but you'll get some inter-charging going on
 

JoeRodge

Newly Enlightened
Joined
Oct 8, 2017
Messages
133
I just learned a lot. I need to go wrap my head around all of this new knowledge lol.
 
Top