When are cells 'in balance'?

[Bikes4two]

• This is not a question about the whys and wherefores of cell balancing as I'm well versed in the necessity of this.
• What I am looking for is in two parts:
  1. I've got a bunch of new cells ready for my first 10s 2p build and will be measuring the voltage of each cell and 'pairing up' the cells such that each cell is near enough the same as it's partner. So here's the first question: what is an acceptable (milli) voltage between the two cells?
  2. So the battery is built and used for a number of charge/discharge cycles but a voltage check of the individual cell groups shows that there is a difference of some millivolts across the 10 groups despite the BMS. So the next question is, how much voltage difference in cell groups is OK to live with.

I'd appreciated it if any answers to the above can include a reference to a reputable paper/website rather than 'my granny who watches YT a lot says it's xxx mV' I ask this as try as I might, I've not found such a source for myself. I know there will be plenty of opinions from memebers on here, but a reference rather than a 'granny says' answer is preferable.

Thanks in advance, B4t

 

[Sturmey]

• This is not a question about the whys and wherefores of cell balancing as I'm well versed in the necessity of this.
• What I am looking for is in two parts:
  1. I've got a bunch of new cells ready for my first 10s 2p build and will be measuring the voltage of each cell and 'pairing up' the cells such that each cell is near enough the same as it's partner. So here's the first question: what is an acceptable (milli) voltage between the two cells?
  2. So the battery is built and used for a number of charge/discharge cycles but a voltage check of the individual cell groups shows that there is a difference of some millivolts across the 10 groups despite the BMS. So the next question is, how much voltage difference in cell groups is OK to live with.

I'd appreciated it if any answers to the above can include a reference to a reputable paper/website rather than 'my granny who watches YT a lot says it's xxx mV' I ask this as try as I might, I've not found such a source for myself. I know there will be plenty of opinions from memebers on here, but a reference rather than a 'granny says' answer is preferable.

Thanks in advance, B4t

1.2. SOC 0%
- Cell voltage in SOC 0% of Pack should be Min. 3.0V.
(Discharging Cut-off Voltage)
1.3. Operating Charging Voltage of a cell.
- Normal operating voltage of a cell is 4.20V
- Max operating voltage of a cell is 4.25V.

Above quote for Samsung 29E spec sheet.
So taking above as an example for a 10s 36 volt battery, if a cell is out of balance and is .05 volt higher than rest of pack at 42v, it runs the risk of going above the the 4.25 volt threshold.
At the other end, if the controller low voltage cutout is set for 31 volts and a cell is .1 volt low, it runs the risk of dropping below the 3 volts threshold. (Note however that some controllers by default may be set for only 30 volts in LCD) Thing may also be worst if you get a combination of high and low cells
So to allow some headroom and variables, cell would ideally need to be within 10-20 millivolts balance as in the example above.

PS Also note that cells that appear to be in balance at 4.2 volt may not necessarily be in balance at 3.1 volts. However, if the cells are of good quality and from the same batch, this will normally be the case.
If in doubt, set your controller LVC higher if this is possible.

 

[guerney]

@Bikes4two - I'm wondering if you used reclaimed cells from cellsuply.co.uk, which you mentioned in another thread?

 

[matthewslack]

1. You can remove any worries there by temporarily connecting in parallel via a low ohms value, high wattage resistor. Voltages will equalise at modest current via the resistor. But for differences of a few mV, the few milliohms of cell internal resistance will keep the equalizing current under an amp anyway, which is probably fine given the cells are not at top or bottom of charge.
2. Those differences are likely just tolerance within the BMS spec. These are high volume, low cost electronics, so precision is limited by cost. My recently purchased 21700 charger only claims 4.20 +/- 0.05V. I would not be surprised to find similar variability between BMS channels.

 

[Bikes4two]

@Bikes4two - I'm wondering if you used reclaimed cells from cellsuply.co.uk, which you mentioned in another thread?

Well remember on the reclaimed cell bit - as it happened I ended up with LG MG50LT from Fogstar wholesale at a great price of £2.99 each. (£4.14 at Cellsupply for new or £3.48 for reclaimed Sam 40T so Fogstar was a no-brainer).

 

[Bikes4two]

• @matthewslack - thanks for your posts and all very useful in expanding my understanding of ebike batteries.
• I've only been into ebikes for around 18 months (but have an old skool background in electronics so hopefully not a complete numpty ) , and the BMS for the 36v 10Ah bottle battery on my TSDZ2 build has been problematic (a replacement BMS for this battery costs, $5, yes $5 - so top quality eh!) with it giving intermittent power cut outs, with sometimes not even recovering with a power off/on, so a bit of a pain.
• The first thing I did was to run in BMS discharge mode but a short while after that the BMS wouldn't charge properly either so for the past 3000km I've been charging through the discharge port (shock horror I hear many cry but probably not if you're from the E-scooter world?) whilst manually checking cell balance with a volt meter about every few weeks.
• The latest check gives the difference in cell groups at full charge between 4.11v and 4.17v, so 60mV. The cell group voltage variation at around 60% SoC is better at around 40mV but much more varied at discharge (for me that's LVC = 31.5v) but the lowest group is nowhere near 3.1v let alone the spec'd end voltage of 2.5v.
• I will be fitting a diode in series with the charging lead to give around a 0.6v volt drop from the charger to further insure no cell group gets too high.
• The cell group difference at discharge (for me that's LVC = 31.5v) is much worse but the lowest group is nowhere near 3.1v let alone the spec'd end voltage of 2.5v

I asked the question in the original post to see at which point I needed to consider manually re-balancing the battery either by fitting the replacment BMS (not keen on this idea given the initial experience) or by manual means (which is my preferred method at the moment). I have a standard iMax b6, 4-button balance charger which of course is only good up to 6s but I was going to partially discharge the battery and use the iMax to discharge the higher group cell groups ones, one at a time down towars the lower group ones (but I like your idea of using a resistor across the necessary cell groups for this - I need look in my scrap box for the high wattage power resistors I have and work out the current/power disipation for something within the current capacity of the balance wires.).

 

[matthewslack]

• @matthewslack - thanks for your posts and all very useful in expanding my understanding of ebike batteries.
• I've only been into ebikes for around 18 months (but have an old skool background in electronics so hopefully not a complete numpty ) , and the BMS for the 36v 10Ah bottle battery on my TSDZ2 build has been problematic (a replacement BMS for this battery costs, $5, yes $5 - so top quality eh!) with it giving intermittent power cut outs, with sometimes not even recovering with a power off/on, so a bit of a pain.
• The first thing I did was to run in BMS discharge mode but a short while after that the BMS wouldn't charge properly either so for the past 3000km I've been charging through the discharge port (shock horror I hear many cry but probably not if you're from the E-scooter world?) whilst manually checking cell balance with a volt meter about every few weeks.
• The latest check gives the difference in cell groups at full charge between 4.11v and 4.17v, so 60mV. The cell group voltage variation at around 60% SoC is better at around 40mV but much more varied at discharge (for me that's LVC = 31.5v) but the lowest group is nowhere near 3.1v let alone the spec'd end voltage of 2.5v.
• I will be fitting a diode in series with the charging lead to give around a 0.6v volt drop from the charger to further insure no cell group gets too high.
• The cell group difference at discharge (for me that's LVC = 31.5v) is much worse but the lowest group is nowhere near 3.1v let alone the spec'd end voltage of 2.5v

I asked the question in the original post to see at which point I needed to consider manually re-balancing the battery either by fitting the replacment BMS (not keen on this idea given the initial experience) or by manual means (which is my preferred method at the moment). I have a standard iMax b6, 4-button balance charger which of course is only good up to 6s but I was going to partially discharge the battery and use the iMax to discharge the higher group cell groups ones, one at a time down towars the lower group ones (but I like your idea of using a resistor across the necessary cell groups for this - I need look in my scrap box for the high wattage power resistors I have and work out the current/power disipation for something within the current capacity of the balance wires.).

'For the avoidance of doubt' as they say, my answers were in relation to your new battery build, not an existing battery with issues. In particular, my (1) is talking about cells not yet spot welded together into a parallel cell group, not connecting a resistor between cell groups within a fully assembled battery.

Take care!

 

[Bikes4two]

Thanks for clarifying that point @matthewslack - I will indeed take care - I've worked on many high voltage systems (RADAR) and I'm here to tell the tale although the high energy density of lithium cells and what causes them to go wrong is something not to be complacent about.

The chemistry and behaviour of lithium cells is quite intriguing - every day is a school day it, even after all my years.