Another Top Balance Screwup

[Luthj]

Not my cells. A friend was top balancing, and got in a rush. Turned the voltage up on a batch of cells, and walked away for a quick drink. Forgot about the cells until the next morning.

The cells were over 5V the next morning.



I had him pull the cells down to 3.6V right away.

They went pretty flat once put in a fixture. Still a noticeably bulged though.



I had him do a test discharge/charge cycle outside just in case. I will report back with the capacity test results. I am curious what kind of degradation the cells experienced. Anyone wanna take bets?

 

[Bob B]

I think we need to save all these links in one thread and put it in a sticky

 

[noenegdod]

I so badly want to say "how the hell do people do this?" and then I remember I once walked out to the garage while I was cooking bacon and came back in over an hour later. Thank god the fan was on....

 

[Bob B]

I so badly want to say "how the hell do people do this?" and then I remember I once walked out to the garage while I was cooking bacon and came back in over an hour later. Thank god the fan was on....

We gotta draw a line somewhere ..... and bacon abuse may be where it is ?

 

[Luthj]

Pack just tested over 250AH (near to 265 I think). So capacity loss doesn't seem to be severe initially. The pack got warmer than I would expect. Its 70F ambient and the pack is around 100F. Though this is in direct sun, so that accounts for some of the rise. As expected the internal resistance has increased a fair bit.

 

[Bob B]

It'll be kind of amazing if it got to 5V/cell and not much damage.

 

[Luthj]

At 5V with LFP the main issue is the electrolyte starts to break down. This causes voids in the sandwich/jellyroll. If you catch it soon enough, and If you can compress the roll down without venting the cell, some of that damage can be mitigated. Though the internal resistance will never return to normal.

I think 265/272 = 97% isn't too bad. Though these cells were probably closer to 280AH, so the capacity loss seems to initially be 5% or so. I am being conservative, its possible the cells actually delivered 275AH, but thats at the limit of the accuracy of the gear.

However I have no idea what the cycle life of these cells will be. Of course even if its only 1000 cycles, thats still a good deal.

 

[toms]

I never recommend this method of top balancing - it is so much safer to always use cell level voltage protection.

The vast majority of cell damage i have seen is due to top balancing errors, and it is so unnecessary.

If your cells are very unbalanced when you get them, charge them one at a time to 3.65V 0.05C. (i use a modified EVworks 3.65V balancer board as a backup disconnect for my charger).

Then connect all cells in series with cell level BMS protection.

The last few packs i have set up i have connected straight up in series and used an active balancer on the first few charge cycles - this has worked fine. (this is with Sinopoly and Winston cells, i haven’t used EVE/Lishen)

 

[Luthj]

This guy did 5 or 6 other packs with the top balance method and was fine. He just got impatient and complacent. There is not excuse for turning a power supply up to 5V.

If I was building a number of these packs I would rig up a disconnect board.

so unnecessary.

Cells need to be synced somehow, and passive balancing doesn't work with any speed. If you want to do a detailed writeup telling folks what to buy, how to wire it up, explain how your method works, please do so.

 

[toms]

I thought my write up was detailed, which part don’t you understand?

 

[noenegdod]

I never recommend this method of top balancing - it is so much safer to always use cell level voltage protection.

The vast majority of cell damage i have seen is due to top balancing errors, and it is so unnecessary.

If your cells are very unbalanced when you get them, charge them one at a time to 3.65V 0.05C. (i use a modified EVworks 3.65V balancer board as a backup disconnect for my charger).

Then connect all cells in series with cell level BMS protection.

The last few packs i have set up i have connected straight up in series and used an active balancer on the first few charge cycles - this has worked fine. (this is with Sinopoly and Winston cells, i haven’t used EVE/Lishen)

Have to disagree. If you take a power supply and set its open circuit voltage to 3.650 volts, connect the cells and walk away, you will never have an issue, ever. It is about as simple and idiot proof as you can get. I did 16 cells like this and never once did I have an issue. The problems arise when people get impatient. When you see someone posting threads like "top balancing taking forever" I cringe because I anticipate another "ruined my cells" thread.

 

[Luthj]

which part don’t you understand?

I can't tell if you are being intentionally condescending or not, so I will give the bennefit of the doubt.

I understand your proposed solution, but that's not the problem. I have 20 years of electrical and battery experience. Most folks don't. So saying you have the perfect method compared the detailed guides written here and elsewhere, is disingenuous. Your approach requires a fair bit of knowledge and decision making based on observations of the specific cells state of charge, balance/matching. This is knowledge many/most don't have. So unless you can break it down in a detailed tutorial, saying your method is superior is purely academic. The fact that you either don't see this, or won't acknowledge it, is silly at best. You also skip over that your approach requires several weeks of monitoring the pack to make sure its staying in balance, as well as adjusting the charge parameters as you go along. I repeat, this is a non starter for many people. Not everyone can or wants to do that.

On the other hand, the power supply top balance approach works for all cells regardless of initial balance state. It requires minimal skill/knowledge. In this case the user got impatient, and acknowledges what they did was stupid, and they knew the risk when they cranked that knob up.

In the future I think replacing the suggestion to use these cheaper bench power supplies, and using a better high output single voltage unit is a better approach.

For every fool that screws up a top balance, there are probably hundreds of others who never post cause it just works.

 

[blutow]

I'm new to building batteries, but I can't understand why anyone would risk turning the power supply up past 3.65v. Is it really that much faster? I don't understand why. I have a cheap 10a supply and set it to 7a. It pushes at 7a until the cell is almost full whether I have the voltage set to 3.65v or 5v. It's hitting it's current limit of 7a (constant current) regardless of voltage setting, what is the point of setting it higher and risking going over.

I get that setting the voltage higher near the end will push more amps when it's in constant voltage mode, but you should be near the end at that point and you won't be saving much time.

I have been setting mine to 3.60v as a first step and then turn it to 3.65v to finish them off. I figure if I forget, I'd rather they be held at 3.60 for an extended period rather than 3.65. I also have a way to send an alarm when they hit 3.6, so I know they are basically done.

Again, I'm pretty new to this, but I just don't get it. It seems that a lot of folks struggling to push amps early in the charge need to get better leads. Trying to overcome a poor connection by upping the voltage doesn't seem like a good path.

 

[Luthj]

Is it really that much faster? I don't understand why

The issue is cheap power supplies and small wires/leads have a lot of voltage drop. So while the power supply might read 3.65V with no load, when connected to the batteries the drop can be 50 or 100mV. This limits the current the power supply will output. By turning up the voltage a bit, you can overcome the drop in the wires and speed charging. However, if you wait long enough the current will taper down, and the cells will charge to whatever the power supply is now set too.

Good quality power supplies with properly sized leads don't have that issue.

 

[blutow]

The issue is cheap power supplies and small wires/leads have a lot of voltage drop. So while the power supply might read 3.65V with no load, when connected to the batteries the drop can be 50 or 100mV. This limits the current the power supply will output. By turning up the voltage a bit, you can overcome the drop in the wires and speed charging. However, if you wait long enough the current will taper down, and the cells will charge to whatever the power supply is now set too.

Good quality power supplies with properly sized leads don't have that issue.

Yeah, I guess I get it if people are doing it to overcome crappy leads, but that seems like a really poor and risky solution to a simple problem. Just make some proper leads and don’t risk cooking your cells.

 

[toms]

I understand your proposed solution, but that's not the problem. I have 20 years of electrical and battery experience. Most folks don't. So saying you have the perfect method compared the detailed guides written here and elsewhere, is disingenuous.

I wasn’t being condescending, sorry if it came across that way - and i’m not saying my method is perfect.

I am saying that some of the battery systems i have commissioned consist of over $30k (aud) worth of cells, and i don’t think it’s wise to risk a power supply malfunction on that.

Hence always using cell level protection - even when initially charging.

There is no more skill required to individually charge cells and then assemble the pack than there is to parallel top balance cells.

Just because a detailed guide has been written, and many people have followed it - doesn’t make it the best way.

One of the biggest killers of LiFePO4 is being held at high charge, especially when the SEI is in its infancy. With highly mismatched cells i’d be concerned at the length of time some cells are being held at 3.65 while already fully charged.

Nothing i’m saying is new, when LiFePO4 became readily available to the public over a decade ago the same discussions (and unfortunately the same mistakes) were made as they are now that LiFePO4 has become very cheaply available.

I’m just glad that the bottom balance disciples have vanished!

 

[Lt.Dan]

I'm new to building batteries, but I can't understand why anyone would risk turning the power supply up past 3.65v. Is it really that much faster? I don't understand why. I have a cheap 10a supply and set it to 7a. It pushes at 7a until the cell is almost full whether I have the voltage set to 3.65v or 5v. It's hitting it's current limit of 7a (constant current) regardless of voltage setting, what is the point of setting it higher and risking going over.

I get that setting the voltage higher near the end will push more amps when it's in constant voltage mode, but you should be near the end at that point and you won't be saving much time.

I have been setting mine to 3.60v as a first step and then turn it to 3.65v to finish them off. I figure if I forget, I'd rather they be held at 3.60 for an extended period rather than 3.65. I also have a way to send an alarm when they hit 3.6, so I know they are basically done.

Again, I'm pretty new to this, but I just don't get it. It seems that a lot of folks struggling to push amps early in the charge need to get better leads. Trying to overcome a poor connection by upping the voltage doesn't seem like a good path.

I'm still relatively new to this too, but I think guys are setting to 5v at 10 amps to push 50 watts into the batteries, Compared to 3.65v at 10 amps is only 36.5 watts. So you do increase the speed at which its charged by almost 37%. I don't think its strictly to overcome the wire loss.

 

[noenegdod]

One of the biggest killers of LiFePO4 is being held at high charge, especially when the SEI is in its infancy. With highly mismatched cells i’d be concerned at the length of time some cells are being held at 3.65 while already fully charged.

If cells are "highly mismatched" they dont stay at 3.65. When connected in parallel, they bleed down into the others until they all average out and then all come up together.

 

[blutow]

I'm still relatively new to this too, but I think guys are setting to 5v at 10 amps to push 50 watts into the batteries, Compared to 3.65v at 10 amps is only 36.5 watts. So you do increase the speed at which its charged by almost 37%. I don't think its strictly to overcome the wire loss.

maybe other power supplies work different from mine, I honestly don't know (and maybe I'm confused about how mine works) .

If I start with a cell at ~3.4v and set my power supply at 3.65v and 10a (my max), the display would show 10a and ~3.45v and "CC" mode for constant current. If I reset my power supply to 10a and 5v, the display still shows 10a and ~3.45v because it's still limited by current. I assumed those numbers on the screen represent the volts/amps it's putting out.

The way I understand it is that as long as you are in CC mode (the PS is pushing the set amps), there is nothing you can do to increase the voltage being pushed, regardless of how high you turn up the voltage. I don't know if I'm right, but that's what I thought based on looking the display. Once I cross about 3.5v on the cells, my amps start to drop and I could increase the amps by bumping up voltage, but that's where it gets dangerous. It also doesn't take that long to go from 3.5 to 3.65.

 

[Luthj]

SEI formation essentially stabilizes after the first couple cycles, and prohibits the further formation (at least rapid formation) of capacity blocking SEI deposits.

The hours (at most a day or two) it takes to get to full (from say 3.4Vpc) is not likely to cause any significant damage.