Intro to Cell Balancing

Cell balancing refers to the methods used to maintain all cells at the same state of charge in a battery pack within a margin throughout its lifetime of operation.

‍

*Note: this transcript was done by an AI with minimal human oversight in the interest of time. Please let us know if anything is unclear and we’ll work on fixing it.

Would you help describe what cell balancing is?

Luke: Each cell has a capacity and a state of charge and, in a balanced situation, you would hope that they all are sharing a similar state of charge with each other.‍

Bryan: And maybe even to start farther back. This is assuming that you have a battery pack where all the cells are identical. If you made a battery pack where cells were twice the capacity of their neighbors, it would look like it's balanced but then they are imbalanced later on. So from step one, you have to actually make sure that your cells are matched in balance, or state of charge, from the factory. If one has 1 Amp-hour, it's paired with others that have 1 Amp-hour so that your active balancing system has a chance of leveling out the pack.

‍

Luke: If each parallel cell group has identical capacity, or as close to identical capacity as you can get them, each state of charge will represent an equivalent unit of capacity stored between each of the series groups.

‍Bryan: And to make the most effective use of the battery pack safely, at some point all of the cells need to stay matched together [at the same state of charge]. And so this is where cell balancing comes into play. Because if the battery pack gets too high and low, high voltage or low voltage on a single cell, that will destroy the pack. In the end, your pack is only as strong as your weakest cell. And in this case, balancing tries to make the other cells or the weak cell keep up with them.

Luke: Or drain the over performers to match the underperformer…

‍Bryan: Which, you bring up a point, that there are two different ways that you could balance this out, either by charging or discharging; the simple way is discharging. It's just as simple as tying a resistor on and bleeding [or letting the battery discharge slowly through a small current]. Like popping up in a hole and a tank that's overfilled and draining some of the excess out and then stopping it. And then there are more advanced ways to charge single cells and strings to balance them.

Luke: And whether or not that more advanced way has benefit for your needs is based on the amount of imbalance that that you may be expecting and in healthy cells. The rate of imbalance you should see approaches the difference in self-discharge rate that you're seeing between subgroups [of cells]. Packs with an active [balance] system may not be able to do this well; [driving a current to balance defective cells can damage the pack and lead to pack failure]. The ability to charge the low cell may not be as valuable as it seems, in light of what's causing the low cell to be…

‍Bryan: Right. To expand upon that, if you have a cell that's self-discharged due to a defect, no [active] balancing system is really going to help you [maintain that cell’s operation in a functional battery pack] at that point.

Luke: [The active balancing system] doesn't correct your defect or repair any damage done to your cell causing it to be defective. It does mask the effect of being able to have clear visibility that you had a defect though. [It can also funnel so much current into the defective cell that it makes the battery inoperable or fail at worse.]

The Importance of Cell Balancing and Pack Safety

There are many different concepts in cell balancing you mentioned: having a pack with all the same cells, having a pack with a bunch of recycled cells, or trying to repurpose cells that are out there. And so, even now that we understand what cell balancing is, why is it important or needed? Are there use cases where you definitely should not do it?And so, even now that we understand what cell balancing is, why is it important or needed? Are there use cases where you definitely should not do it?

Luke: Well, I guess the start of your question about using different types of cells as a product, if you're building something for yourself in your garage, that is only your risk involved with it. That's one thing, but if you're producing a product that you're selling to the public, it's critical that you're only using premium cells and cells that are matched. By matched, I don't mean necessarily that they had to be hand sorted. But what I mean is that they were all passing through formation and meeting the same criteria of manufacturing acceptance on quality.‍

Bryan: To expand upon that point, we are talking about balancing and there are two types of balance. There's not balancing but actual balance, there's a static balance, and that's when the battery is at rest. And that balance shows you a capacity difference, usually between the cells. And then there's a dynamic balance where, even if your battery pack was fully balanced during a high dynamic load and high current pulse, if there's any sort of impedance or resistance in balance of your cells, it'll show up as voltage and balance that then disappears when you let off the throttle. And the current comes back normal. And so the matching of the cells was alluding to not just capacity but also cells’ internal impedance. And so this way, while you may have no imbalance while sitting in static air on a cold day, you may show a massive imbalance while under load that disappears, and that's just showing the tolerance spread on the cells. So, as Erika brought up, Second Life in or mixing and matching cells gets very complicated if you are taking cells that don't match up via capacity and impedance and trying to use them in our overall larger pack design. Not to say it can't be done. It's just non-trivial and it's very risky to play with cells that may become overcharged in the middle of the pack.

Luke: Well, I guess the start of your question about using different types of cells as a product, if you're building something for yourself in your garage, that is only your risk involved with it. That's one thing, but if you're producing a product that you're selling to the public, it's critical that you're only using premium cells and cells that are matched. By matched, I don't mean necessarily that they had to be hand sorted. But what I mean is that they were all passing through formation and meeting the same criteria of manufacturing acceptance on quality.Also, you can get recirculating currents in parallel groups if you have different impedances. Like Brian mentioned, under those dynamic conditions where some [cell] groups sag [or exhibit a lower voltage due to capacity fade or environmental influence] further than others. In a parallel group, if some cells sag further than others, then you end up recirculating power from your low impedance cells to your high impedance cells during rest periods in the pack. All reasons why you really should be starting from all similar cells. Differences in impedance, age of cells, and capacity, all of these things make it a very difficult pack to manage safely.

Thank you. I think that puts a nice warning label on the kind of conversation we're having today, which is cell balancing. Considering you're getting your cells from the same source and then they're of the same type and have the same lifespan impedance and resistance…

Luke: In those situations, there will still be some differences because no pack has perfect temperature uniformity. Even in the situation where [the pack has] been sitting out over the morning or over the night, and now the sun warms it on one side preferentially than the other side so that you see these temperature gradients which will appear as dynamic imbalance. However, it may be a temporary dynamic imbalance once the system reaches an isothermal state.‍

Bryan: Static balance. This is why when you're flying some sort of RC you need to monitor each cell individually because at the moment when one cell reaches a critical voltage you have to shut down. This is technically a form of balancing because you're not allowing one cell to go completely out of balance and destroy itself.

That is a great way to segment the other part of my question about cell balancing. If you didn't balance this pack, now that you've monitored the cells and you're putting them into [a new] pack, if you don't have active balancing [on this new pack] what happens?

Bryan: Also, some battery packs get away without balancing because they don't end destructively. So for a lead acid battery, you can equalize charge and overcharge it to balance it and it doesn't risk destroying someone's house, and so it's not as important. Now, in our newer chemistries, it's much more important from a safety aspect. That balance [must be] done properly.

Luke: The safety aspect is obviously the biggest aspect, but from a electrical performance aspect, if whatever percentage out of balance you have in capacity is a direct percentage loss that your pack is able to make use of. So, if we have a cell that's 5% lower capacity than the other cells in the string, or lower state of charge, then, when we go to discharge, that cell reaches its minimum safe voltage 5% sooner than the rest until we have to stop the full string being discharged. Same thing if you're charging and one cell is 5% higher. Then, you would have to stop while the rest of the cells are only at 95% charge, for this reason, hurts you by the amount of the imbalance from an electrical storage perspective. ‍

The safety perspective is really the most critical thing though. I won’t get into it yet, but I actually like when I see the BMS start to show an imbalance that has a problem rather than forcibly correcting balance.

Bryan: Normally a balancing circuit is just there to make up for minor manufacturing defects in the cells. They should not try to hide problems; it should not make up for problems if you have a cell that has self discharge. A balancing circuit will hide it visibly from you unless you have something that monitors how much cell balancing is happening. But it's more improper in designs for battery packs. It's a danger to know that you have a self discharging cell and at that point the battery pack should be flagged for shut down. That's a lot different than the hobby packs where we just try to use our RC packs until they die. But they're also different scales of size.

That is a great way to segment the other part of my question about cell balancing. If you didn't balance this pack, now that you've monitored the cells and you're putting them into [a new] pack, if you don't have active balancing [on this new pack] what happens?

Luke: In a perfect world? You know, if you've had all good cells, you'd never need your cell balancing circuit to do a thing for the whole life of the product. So in a perfect world, your amount of balancing you'd have is zero and you'd never need it in any of that circuitry. In practice though, the self discharge rate, while we try to make it zero and accelerate the actual electron mobility from anode to cathode - there's the impurities in the solvent. There's nucleation sites on cut foil edges in surface of active material. And these things all have imperfections where we lose some electrons. And so this is the self-discharge effects. Another one is the gas evolution, splitting gases and electrolyte side reactions. All of these things cause self discharge [inside cells].

Bryan: So I'll go into one of my own personal applications. I have a off-grid tiny home that runs on battery packs and solar, and they are actually second life cells. What I had done with these cells was I actually measured impedance and capacity on them and matched them myself into groups since this is a very large battery pack, and it's running for a very long time, and it's my personal money. I have referred to BMS as battery management systems that fail to function as “battery murdering systems”. And the reason why is if you have circuits that are capable of discharging your cells, there's a chance they damage your pack. 

Currently on my tiny home, I check balance with a balancing board that I plug in physically and double-check manually. And this way, I'm not trusting any electronics on my battery pack. I still have things that shut down my battery pack due to LVC, low voltage cut off, or HVC, high voltage cut off, but I have no active balance in circuits to minimize risk of them not shutting off and destroying my expensive battery pack. 

Now, this is a very rare point of view. But this is my point of view from seeing many balancing circuits fail to function properly and do more damage to the cells that they were intending to save.

That's a great example. Bryan is obviously an expert and knows what to look for in the battery cells. What I'm hearing is, basically, when you are installing the cell balancing monitoring system that it needs to have a lot of considerations that you're not going to kill [the battery].

Luke: Even if [the battery] has no cells corrupt, let's say that the product was designed with no electrical or hardware fault. You could still have the harnessing of that system have corrosion related faults. These faults may cause fires or other risks.

But we're seeing it's kind of a necessary evil because you're not having a Bryan look at every cell, right?

Bryan: That's it! I have to manually go and check my battery packs, which means that it's then my fault if it breaks. So the other thing with my tiny home battery pack is that I have to sleep with it. They're also very safe chemistry cells. So there are a lot of other safety aspects on top of that, so I wouldn't suggest anyone just run out and run no cell balancing because, again, I do have LVC HVC, and those are the bare minimum requirements for safety, but to get the most use out of your pack, as Luke was saying, if you want to actually charge and discharge it to full power and full capacity, then your cells are going to have to be balanced. So, currently, on my teeny home pack, I started to lose some watt-hours of capable capacity because I'm not balancing it actively. But that's okay as far as a trade-off to the peace of mind. If I develop my own BMS and slap it on it, I might trust it, maybe.

Wow. So that kind of leads us into the next section you wanted to talk about. For the purposes of [cell balancing], we're talking about a production kind of environment with using the same kind of cells. How does one perform cell balancing?

Bryan: Hire someone.

Luke: Since we don't need it, we've already talked about that, we don't need a lot of current, even if you have many hundreds of Amp-hours of capacity in each of your parallel cell groups. You don't need a lot of current to balance this because you're only balancing to keep up with a difference in healthy cells’ self discharge rates, which should be approaching zero. So we've seen examples where 100 milliamps of balance current was over-kill for 120 Amp-hour capacity cell group, in that by over-kill I mean that you could mask the issue [with the cell]. It would have been better for the BMS to let that cell group die from from self-discharge [than try to maintain its balance with other cells in the pack].

A frequently asked question we get is: is there a minimum or maximum value specified for cell balancing or is this product specific?

Bryan: It’s mostly product specific and related to the internal self-discharge of the cell.

Luke: And the variance in it between cell groups.

Bryan: Lithium-ion batteries tend to have very low self-discharge, and so they are not as needed. But I do want to bring up one aspect that maybe a lot of our viewers have seen and I want to explain it. And that's in RCs you may find battery packs are really imbalanced, and you may find that you need to balance them all of the time. And the reason for this is those batteries were being abused. So when you touch a pouch cell you cause minor microscopic damage that starts self discharge. So even by handling the cell, you've now created self-discharge that your balancing circuit is going to have to make up for when it's charging. And so in a lot of RC packs you find that you always have to balance because your cells are dying. Those cells also don't last 1000 cycles. There's a line between the hobby cells and EV-grade class of cells. The ones that we're talking about that don't require balancing are the EV/grid-class cells. They're not your hobby packs. You're flying a drone.

Luke: And no matter how balanced or unbalanced your pack is, you still have to have your BMS make decisions ultimately off, you know, cell voltages, regardless of what your calculation may think the capacity is remaining in that cell group. HVC and LVC have to be king on all BMS decision matrices for safety reasons.

That was another frequently asked question we get is: what metrics must the BMS make in regard to operating decisions in respect to balancing?

Luke: You know, the best decision that BMS can make is when it finds there's some cell group that it never needs to turns the bleed resistor on for, and there's all the other cell groups it has to bleed frequently to keep them as low as that group that's never getting bled [or bleed current]. That's the signature that you have a defective cell or a corrosion issue external to the cell. Either way, though, that's a sign that the battery should let that BMS system flag it to stop charging and let it finish and end its life if it has a problem.

Bryan: And this is also why there are some battery packs that you may find in the field where you open them up and there's only a dead cell or two. The management system may have let it die for safety. Or, the other flip of the coin, is this management system may have murdered [the battery]. It's hard to tell at the end of it.

Is it possible to balance a pack too much, and what happens?

Luke: Definitely possible.

Bryan: You get one scenario that's just inefficient. If you were to balance at top and bottom and you were only using resistor-based balancing, then all that heat that you produce is not energy that can be used in your system. And so there's extravagant active balancing systems that use little switching power supplies to try to save this efficiency. But as we came back to it, you know, use good cells and you shouldn't have to balance.

Luke: Your deviation [between cells’ self-discharge properties] in a healthy pack should be extremely tight.

What is the maximum balance rate for a EV-grade pack that you usually consider for, let's say, an NMC cell?

Luke: The cells that I'm used to working with, both in cylindrical and pouch type you know are received at voltages tighter than 1 millivolt to each other. And after many years of use without any balancing at all, they usually are within a millivolt or two of each other. But that's an exceptional quality of cell packaged in systems that keep them at a uniform temperature without abnormal interconnect heating and other effects depending on the quality of design of the pack. You can see significant degradation differences between subgroups sometimes [if cells are not thermally and electrically packaged well].

Bryan: I mean, the only experience I have with that is a personal battery pack that I had ran for six years, cycling quite often. And it was an EV motorcycle and the battery packs at the end of life appeared to have only balanced a couple Amp-hours across the entire 25 Amp power pack in the entire lifetime. And that's because the management system actually kept logs of every time a balancing circuit was active. And I can read that out. Some management systems don't have that kind of logging. And so it's kind of hard to tell at the end of life how effective the balancing circuit really was.

That's good feedback and it leads into our next question, which is how do we design battery packs for efficient cell balancing?

Luke: Well, you start with good cells that you keep the same, uniform temperature. You don't have asymmetric heating sources and there’s bussing and interconnects and that's your best approach to maintaining cell balance because everything you do that's, that's not that way to solve it, unfortunately. It’s a band aid because you're causing asymmetrical self discharge damage in your cells or you're accepting cell defects, right so either way…

Bryan: Maybe for viewers the asymmetrical temp cell damage… how would that occur in a pack?

Luke: If you have cooled places and hot places in the cell, there will be this asymmetry of use in your active materials. Some effects impact by a factor of 2x for every 10 degrees Celsius of total imbalance…

Bryan: [The battery starts with] all the same matched cells and they can become unmatched.

Luke: Right, because once you know some area was located where it was a heat island, and some face was located where it was cooled effectively. And now the cells that started matched have taken on different rates of aging, different rates of impedance climb.

Are there times not to balance the cells, for example, when you see these impedance variations or temperature variations?

Luke: So I don't like to balance below around 90%.

Bryan: And then you also don't want to balance while in an active situation. If you have an active or dynamic imbalance you don't want to react; you're balancing circuits on voltage drop due to load and impedance.

Luke: …or voltage rise due to charging currents. Either you've got to render a no current situation when you measure balance and then engage your balance…

Bryan: And even then, the battery pack should be at rest for quite a while because all of your voltages will be changing at, maybe even hours depending on the chemistry of the cell after a high current situation.

Is there a certain kind of circuitry or design you prefer when performing self balancing?

Luke: I like resistive bleed where a small MOSFET is used to switch in and out a low or appropriate value ohm resistance to cause some value under 0.1% of the capacity of that cell string and in current, so perhaps even 0.01% is more like a realistic current value for the entire capacity. Rarely would any sized pack need more than 10mA bleed, most need under 1mA bleed.

This is a bit of a billion dollar question and we've kind of covered it a bit here: is mixing cells or using recycled cells to make refurbished batteries viable through [many amp active shuttle balancing] between cell groups?

Bryan: I personally would never suggest it, but in the past I've tried it's just a lot easier to start with good cells. If you're going to put a lot of time and effort into a battery pack just to have it die on you, it's really sad.

Luke: I gotta say, it's the kind of thing that an expert might do for themselves if they had a free supply of something suitable. At minimum you'd have to start with high-grade cells in their second life. Second Life in a low grade consumer cell might be riskier, but then are some of the new chemistries that are coming out for electric vehicles. They're quite dangerous to deal with by themselves and so creating secondary packs without really understanding the exact chemistries could be risky.

And just for clarity sake, if I were to come to you and I said I have an unlimited supply of the best cells in the world, and I have a BMS designer but I have so many questions about cell balancing. How would you walk me through implementing that into a design?

Luke: You're gonna have to sweep your self discharge curves and find out what's representative on your cell stock, every cell has some [incredibly small and difficult to measure] self discharge. So once you quantify that balance of what healthy cells look like, it's going to help you understand what a healthy BMS balance current would look like for your pack. That will still allow you to flag but not so much balance current that it'll mask when you have a defect because the BMS shouldn't be there to cover your defective cell issues.

There are some situations where maybe you shouldn't do [cell balancing]. And let's say now I overdo it. What did I do wrong? What happened?

Luke: It comes down to gas production. If you have that defective cell, it's been discussed for multiple years. Now I'm trying to get an answer if there's a method that doesn't involve gas production when cells have self discharge, but it appears to date we only know of types that do cause associated gas production. This means whether you're a cylindrical, whether you're a pouch, whether you're a prismatic, you have a finite volume to store gas evolution and you'll increase in pressure until at some point you breach containment of electrolyte. At which point your cell can ruptures its sealing. Yeah. In a nice situation, maybe you just spurted electrolyte into the assembly, but either way, this will cause a corrosion induced fire hazard risk, which is, unfortunately, fairly easy to induce in the presence of electrolyte and electricity outside of itself, especially with exposure to atmospheric humidity. For these reasons, it's best if the cell electrolyte stays sealed in its casing. And this is the reason why we mentioned not over-balancing is to keep the electrolyte, keep that corrosion fire hazard risk minimized by keeping liquid inside the cell rather than rupturing it. Even if the BMS sustains the pack until it ruptures, this also doesn’t fix the self-discharge problem until the point that it's developed so much gas it's now breached this corrosive liquid into the pack assembly generating a fire hazard.

Bryan: I’ll just add on that from the pack assembly that is in large EV battery packs, it's easy to have 200 Amp-hours on a string, in that sense, too, is if you have a balanced wire and it's only going to carry 100 milliamps you need to fuse out the battery. Because if something happens with that balance tap wire and it shorts out. It has a 200 or plus Amp-hour battery pack pushing behind it and that wire will evaporate if it's not fused and it's not something you find on a lot of battery pack designs at the moment but you want to fuse it. Every wire coming out in the battery pack right there, and that includes your balance wires because the balance current is so small. It can be something very teeny, a thermal fuse or even a fusible link of some sort, just something that needs to protect because you're balanced circuit can only pull 100 milliamps. If your MOSFET shorts or a wire gets pinched. It will flow more than that, and the higher voltage of the battery pack the higher the risk you have of a mid-pack string balanced wire touching an end, and currents will get on the handle real quick.

Do you have any opinions on battery safety sensor technology to detect early signs, like recording dewpoints humidity pressure change temperature and pack? You talked about corrosion and electrolyte spurting. Do those sensor technologies seem to be a promising way?

Bryan: Ah, I personally have not looked too far into it. But the gases that are given off or kind of interesting. So hydrogen is one, and unfortunately, lots of things give up hydrogen. So if you're trying to detect hydrogen, that's probably not the route. Other things like carbon monoxide is another gas produced internally as a byproduct. And that's another one that can just happen to be in the environment by itself. From non-battery, clean gas, like blow torches is another gas formed within there and that also breaks down very rapidly in the air and decomposes.

Luke: The VOC sensors pop for the electrolyte solvent base pretty readily, but they also pop for many many things readily.

Bryan: A paint pen, like a shakeup paint pen, actually has a solvent that's almost exactly the same as what is used in some lithium ion batteries. So if somebody used a paint pen around some of these sensors, it may trigger it. If you're in a grid-tied storage situation and you have a sealed container and a controlled environment, maybe some of these gas detectors might be useful. I personally have not used any. Just because the environments that I'm trying to implement battery packs are not clean enough to try to take a sampling of the air.

That's good input and I totally understand the niche of technology the sensors want to fit in, and I'm excited to see that grow and learn more about it. Is it possible to predict thermal runaway or problems prior to this off-gassing, or like not even need a sensor prior?

Bryan: If thermal runaway is somebody stabbing it with a knife, you probably won't detect it.

I mean, from a cell health perspective, are there any red flags that would give a way to know if a pack is on its way out?

Luke: The signature of having all the other balance resistors need to turn on except one group always. And that group always leads discharge. They'll always be some cell in a natural distribution of healthy cells. There'll be some cell that has a healthy but greater than the other healthy cells discharge. So this this would be your noise risk of false flagging. And once you understand that noise distribution among healthy cells, you can understand what bounds definitely represent an unhealthy cell. And whether it's corrosion, whether it's internal defects, whether it's a separator problem, a freelance problem, a coding problem, broken SEI problem, formation problem, whether the envelope got breached and humidity is intruding into it, all of these things will cause the observance of voltage suppression in that cell as it has increased, just self discharge. This is a signature that you can watch for to try to preventatively stop your pack from the BMS has just told at some point no longer: Where are we going to not permit recharging this pack? And that's really the way you have to safely bound the life of that system. Let them continue to discharge it, but it won't charge again. This ends up leaving it in the safest condition for its decommissioned state.

I like that you always say the best death for battery is a graceful one. Let it die in a way where it does not have thermal runaway.

Luke: No matter what fancy balanced circuit you got, it can’t fix your foil defect. It won't fix your water in the cell. It won't fix your corrosion external, you know, it won't fix any of the problems, won't fix your separator. All it can do is mask those real safety concerns from being able to be flagged and have that pack die gracefully.

Bryan: The energy has to go somewhere, you know. And if you're losing energy that you have to balance it’s more than likely that energy is going to destroy the cell.

Luke: That energy is not going into some pathway that doesn't escalate, right? It's causing degradation that leads to more degradation. It is an accelerating spiral and it will end in fire if you misuse it by just forcing balance to be numerically correct at your voltage taps while you have a serious problem happening.

I like that you always say the best death for battery is a graceful one. Let it die in a way where it does not have thermal runaway.

Bryan: Always refer to the operating manual from the cells from your supplier. Step one, read the instructions.

Luke: They’ll probably have good advice on what bounds they like to see voltage in for your type of application.

Bryan: And who knows? There might be a new chemistry around the corner that we may not know of that requires a different form of balancing. So you know, make sure you just understand what you're working with.

Luke: Your cell manufacturer will help you understand what healthy self discharge should look like too. Your cell maker knows very well if they see excessive self discharge it means they made a defect.

Bryan: And if you can't find a datasheet for your cell, it may not be a high quality [cell].

Fair enough. And it's a bit more of an advanced concept at let's say there's a dual chemistry cell out there. That's kind of been a hot topic in the industry. What are some of the challenges of cell balancing either a new chemistry or a dual chemistry?

Luke: It sounds awful.

Bryan: We haven’t looked into this. It would increase complexity, increasing ways of failure.

Luke: I would need to sweep the cell many times at different temperatures and see rates and understand the nature of that cell’s relationships. I haven't tested a single dual chemistry cell yet to date. And I think there's good reasons for that too, because most dual chemistry cells only actually get to operate on one of the two halves of their cathode material at a time [except for a small middle ground window]. So, I would say the dual chemistry from a functional concept maybe needs reconsideration.