“Normal” Cell Delta at Charge Voltage

[alferz]

Is there such a thing as a normal, or acceptable range, for the cell voltage differential at top of charge? Strictly as an example, if a 230ah 8s pack stops charging at 27.8v, leaving it’s cells with a [min, avg, max] of [3.440, 3.475, 3.498] and thus a 58mV cell voltage delta/differential, would that be considered normal?

At what point would the mV delta value become abnormal and warrant increased absorption time and/or further testing? 200mV? 500mV?

 

[740GLE]

I think the delta isn’t that big of a deal but how that delta changes over time. If it stays at 58mv over months on end I wouldn’t worry about it

Is the balancer keeping up or are the cells diverging and it becomes 100mv in short order.

 

[alferz]

So far mine holds at 60mV between cycles, the balancer doesnt make much of a dent beyond that (20mv diff to balance, 3.4v or above only, and whilst charging and discharging). Prior to manually in place top-balancing to 28v (ie while in series formation, not on the bench and in parallel) I was regularly seeing 200-300.

I guess what Im wondering is, what is the magic number I should look for and at what point does the value become a problem?

 

[Steve_S]

You will find that when you go above the 3.400 mark the deviations increase. That is because you are above the working range which ends at 3.400Vpc.

As this is a 24V system set the following params.
Bulk/Absorb : 27.6V (3.450Vpc)
Float : 27.5V (3.437Vpc)
Your EndAmps/TailCurrent is 11.5A. when charge is only taking 11.5A switch to float mode
* EndAmps is when the cells are actually "full" at which time switching to FLOAT aka CV (Constant Voltage Variable Current) to top off and level the packs.

A healthy cell differential is <20mv when static (not charging or discharging). Differentials can increase during heavy loads or heavy charging (at 0.5C for example). Differentials vary depending on load/charge, properly Matched & Batched cells do not deviate much within the working range (that is the purpose of matching cells, so that the IR is consistent will all cells through the whole working range of 3.000-3.400)

DO REMEMBER TO CALIBRATE !
Using a DVOM/DMM Read the voltage at the battery terminals and jot that down, then check at the Inverter/Charger Terminals and jot that down, also if you have a separate SCC get the voltage there too. There will most likely be a small difference between the battery & inverter, so you have to correct for that. Ultimately the Voltage @ Battery Terminals is the Critical Number.

Depending on BMS, some can be calibrated internally as well like the JKBMS (which sadly most don't do).

Hope it helps, Good Luck

 

[alferz]

One thing I’ve noticed about both of my 8s packs, is the JBD BMS really doesnt seem to balance at all when the charge current is zero. I can let them sit in absorb at 27.8v for 2 hours without the 44mV and 42mV cell delta moving more than 1 or 2 mV.

So thinking balancing under current was the key, I wrote an algorithm to control my Victron Multiplus-II (running a custom ESS Mode 3 implementation I wrote with node-red) during the latter part of the charge cycle. I call it “Slow Bulk” or like a pre-absorb phase. From 27.2v to 27.8v it slows the charge way down, to 5a and tapering off to 1a by the time it gets to 27.8v. At the 70a max the MP-II would normally charge in this range, it screams through it in about 13 minutes and doesnt give the tiny passive 150ma balancer on the JBD BMS a fighting chance at balancing those cells while they are actually taking current. With Slow Bulk, I slowed the time down from 13 to 60 minutes.

Unfortunately I dont notice any real difference in balancer performance when charging in this manner. I still reliably get 42-44mV delts at the end of the cycle. Im not going to lose any sleep over this small delta which disappears as soon as the packs are back in service and the voltage drops. Maybe 150ma just isnt enough current to make a real difference on the passive balancer with a bigger pack like my 248ah’s?

 

[Dynoman]

One thing I’ve noticed about both of my 8s packs, is the JBD BMS really doesnt seem to balance at all when the charge current is zero. I can let them sit in absorb at 27.8v for 2 hours without the 44mV and 42mV cell delta moving more than 1 or 2 mV.

So thinking balancing under current was the key, I wrote an algorithm to control my Victron Multiplus-II (running a custom ESS Mode 3 implementation I wrote with node-red) during the latter part of the charge cycle. I call it “Slow Bulk” or like a pre-absorb phase. From 27.2v to 27.8v it slows the charge way down, to 5a and tapering off to 1a by the time it gets to 27.8v. At the 70a max the MP-II would normally charge in this range, it screams through it in about 13 minutes and doesnt give the tiny passive 150ma balancer on the JBD BMS a fighting chance at balancing those cells while they are actually taking current. With Slow Bulk, I slowed the time down from 13 to 60 minutes.

Unfortunately I dont notice any real difference in balancer performance when charging in this manner. I still reliably get 42-44mV delts at the end of the cycle. Im not going to lose any sleep over this small delta which disappears as soon as the packs are back in service and the voltage drops. Maybe 150ma just isnt enough current to make a real difference on the passive balancer with a bigger pack like my 248ah’s?

My system uses the Overkill Solar 8s 100a LifePo4 BMS which is a JBD. It is a very good BMS, but not very good at keeping my 2 packs of 8s non grade A 271 ah Lifepo4 battery banks balanced at the upper knee of the voltage curve. I needed to go with the generic Heltec 5A active balancer to assist the Overkill BMS in keeping a good balance of all the cells. By using this approach the cells only needed the initial top balance.

Found it best in my system to only enable the active balancer when the the cells are above 3.4 volts. This can be accomplished by simply plugging & unplugging the Heltec active balancer or solder a switch on the Run solder pad. Some automate the process by switching the Run solder pad by a remote device.

On my system I use the Victron BMV-712 Smart Shunt relay to switch On the the Run pad of Heltec 5A active balancer(s) at 27.4 volts and Off at 27.2 volts (24 volt system). There is a 0.2 volt dead band with a 5 minute delay by the BMV-712 so no switching on/off when sun goes behind clouds for a few minutes.

When there is enough sun the cells balance at 28.5 volts for 20 minutes in absorb with no problem (no high voltage disconnects with setting of 3.65 cell over voltage). This works even if it has been a couple of months since the batteries were in the upper knee while using a high charge current.

The MPPT SCC float voltage is also set slightly above 27.2 volts so the cells balance while in float also.

Andy at Off-Grid Garage recently did an interesting video on the subject using a circuit board to turn on/off a Heltec active balancer.

Heltec 5A active balancer available on Amazon

https://www.amazon.com/Equalizer-Balancer-Transfer-Equalization-Capacitor/dp/B09MS9JPDZ

See attached Heltec 5A active balancer manual.

 

[venquessa]

As this is a 24V system set the following params.
Bulk/Absorb : 27.6V (3.450Vpc)
Float : 27.5V (3.437Vpc)
Your EndAmps/TailCurrent is 11.5A. when charge is only taking 11.5A switch to float mode
* EndAmps is when the cells are actually "full" at which time switching to FLOAT aka CV (Constant Voltage Variable Current) to top off and level the packs.

While this sounds good it has a few practical issues outside of an ideal setup.

27.6V is enough to top of charge at 3.450V yes, but it's not enough to leverage the balancer and come anywhere near solving top of charge imbalances. If all cells are at 3.450V you can call them charged, but I don't believe 27.6V and 11.5A will achieve that.

11.5A taper will never allow an 8S pack to balance without high voltage disconnecting a cell.

If you have perfectly matched cells and your daily cycle is shallow and always top down, then it might work out.

If you have poorly matched cells and your daily cycle is entirely weather dependant, often spending days and days in the lower 30% then it most likely will not work out. Even at that it comes down to your balancer and how you set it up.

Given an extreme case, I have a "weak cell". One of the 8 (no. 6) appears to have about 5% less capacity to it's peers (and maybe getting worse). Without longer absorb times and far, far, far lighter taper currents I will never get a top of charge in balance and I will be limited to 8 times the capacity of that weak cell -minus other balance loss. With longer absorb times and more dynamic top of charge current rating (with cell awareness) allows me to top balance the pack in normal operation and then continually balance it all the way down as well, such that when it spends a bunch of days between 10% and 30% it balances itself there and I gain more access to the low end capacity i would otherwise miss.

As soon as a single cell goes "vertical" on the chart you have two options. You terminate charge, shrug and call it good, leaving a mess behind.... or you limit the charge current to that which the balancer can fully control. If you have a 5A balancer, for example and a cell nears HVC you would need to limit the current to 2.5-5A. With a 5A Heltec I found that 1 Amp was needed to stop the runaway cell. The "across the pack current" is equal to the "across each cell" current. so at 5A, you WILL be pushing 5A into a full charged cell. It WILL 100% overvolt. Unless your balancer is capable of discharging that cell consistently at 5A you will need to go to a lower current.

Balance is more about the cell delta. it's also about how many cells are high/low. If it's a single cell that can lead to the worst case. A single cell being 0.5V higher than the others terminating charge. This is a different situation to having 4 out of the 8 high and 4 out of the 8 low. The later is easier to deal as you have 4 high cells sharing the current.

 

[Craig]

It's also important to remember that these BMS' do not have all that great of voltage sensors. I bet that the margin of error is somewhere around .05 v. So my answer is yes .06 or less is probably OK.

 

[octal_ip]

Another thing to keep in mind is that the current most active balancers can move is entirely dependent on the delta voltage between adjacent cells. A 5A balancer might be able to move 5A current under ideal conditions with a very large, unrealistic cell voltage delta (>1V), however if your cells are only a fraction of a volt apart, this current won't be anything near 5A.
For large capacity cells, it's not really possible for a balancer to make large corrections, they're just enough to provide a bit of maintenance due to natural variations between reasonably well matched and charged cells.

The only exception would be the few active balancers that have a charge pump that forces higher currents at low voltage deltas between any cell, and not just adjacent cells (like the JK BMS).

 

[venquessa]

A technique I have seen on some BMS's (NEEEY maybe) is a secondary charge circuit.

A true balance charger has the balancer and cell monitor built into the feedback loop which limits the current to keep the high cells within limits while the low cells continue to charge. It's very, very, very slow charging an inbalanced pack. Most will attempt to pre-balance before going full current at the start.

The trouble is... these work with small cells and small currents. it is a lot more difficult to current limit 100Amp than it is to current limit 10A. This is why we just don't see them in bulk storage lithium.

However, the interesting technique used, which matches my balancing technique nicely, is when a cell (or pack) reaches a certain limit, the charge mosfets open, cutting the main direct charge link to the pack. However, a second smaller buck converter is then used to current limit charge the pack at only <10A. So on a sunny day with an inbalanced pack the looming HVC would cause the charge to stop, but this little 10A charge controller (basically a similar topology to an MPPT), which has information about all the cells continues allowing the balancer plenty of time with low current to balance the pack for the rest of the day while it's sunny. You can of course implement the tapering when all cells come to 3.450V or terminate it.

I believe this should be retro-fittable. I'll look into it. One big concern is how it would handle being on the bus with multiple active MPPTs and loads.

 

[venquessa]

Also. The solution is not to constantly "slow charge" to top balance.

The solution is to capacity balance the cells with shims or by paralleling them into equal pairs.

Apart from the fact they are turning into hens teeth, I want to get a few dozen small cylindrical cells which can be used to increase the capacity of ... that weak cell.