When to replace the battery cells?

[Just]

Hello All,

How would I know when to replace my battery cells?

Thank you!

[Bikemad]

I would suggest they would need replacing if your battery capacity (or your typical range) is greatly reduced because failing cells are causing the BMS to cut off battery power, even though the remaining good cells may still contain a large percentage of unused capacity.

For example, if a well used 10Ah 16 cell pack has 14 reasonable cells that can still deliver between 8-9Ah capacity and two failing cells, which can only manage 3Ah and 4Ah capacity, you will only be able to use 3Ah from the the fully charged pack before the BMS senses a dangerously low cell voltage and automatically cuts off all power.

Replacing the two worst cells with new 10Ah cells should bring the packs usable capacity back up to somewhere around 8Ah.

Alan
 

[Just]

Thank you Alan for your comment!

What about measuring an internal resistance of the cells? Some guys recommend to measure an internal resistance of the cells periodically and keep a history of these measurements for each cell. Once an internal resistance increases by about 20%-25%, the cell should be replaced.

What do you say? Should this maintenance be redundant and too much complicated? What device would you recommend for measuring an internal impedance (a cheap one  ?

Thank you!

[Bikemad]

Measuring a cells internal resistance is not something I've ever tried, but I've just had a quick look on Google and came across This:

I've seen here and there posts about people having trouble with their batteries, some from DX (I've had issues, too) and other places, so now's a good time to post a simple test you can do to find the health of your rechargeable battery.

For an example I'll use a common 18650 li-ion cell. When fully charged these carry a potential of 4.2-4.25V and when discharged 3V. This is true of any cell that hasn't shut down whether it's new or old and tired. So what is the difference between a new cell and an old cell? Internally a battery consists of an electrolyte/electrode interface where the potential is generated by chemical reactions. As the reactions progress forward and are reversed during charging, the interface slowly degrades. This degradation results in not only a loss of capacity but also an increase in what is known as the batteries internal resistance.

High battery internal resistance is a BAD THING for a number of reasons. First, it makes the battery output voltage drop when the battery is being used. As soon as you take the battery out of the circuit to measure its voltage, the voltage rises and it seems OK again. Secondly, high internal resistance saps output power from the battery's already diminished capacity. This sapped power winds up as HEAT in the battery and can cause ugly issues and further degradation.

Measuring the battery's internal resistance is simple. All you need is a 4 ohm 5W power resistor or similar and a multimeter. Five common 1W 1 Ohm resistors in series would work. If you aren't measuring an 18650, choose a load resistance that will load the cell but not overload it. Capacity/3 should be ok.

1) First, measure the battery's voltage when charged. This is V1
2) Next, connect your multimeter leads to each side of the resistor and briefly connect the resistor across the battery. Note the voltage reading, this is V2.
3) Measure your resistor to get its precise resistance, this is R

4) The battery's internal resistance (Ri) is calculated with the formula:

Ri = (V1-V2)*R/V2

Example: cell measures 4.2V unloaded, 4.0V when connected to a 4 ohm resistor has an internal resistance of 200 mOhms (0.2 ohms)

A new high quality 18650 battery will have an internal resistance under 100mOhms.

A used up 18650 battery will have a resistance 400mOhms or more.

Please note that the above procedure is for checking individual cells only, so don't put the 4 Ohm resistor directly across the battery terminals:

Current = Volts / Resistance, therefore 48 Volts / 4 Ohms = 12 Amps.

Watts = Volts x Amps, therefore 48 Volts x 12 Amps = 576 Watts, which might be slightly too much for the 5 Watt resistor.

You have been warned!

Alan
 

[Just]

Alan, thank you for your post!

Actually there are special Internal Resistance Meters on the market... The cheapest one, which I found, costs about $50. It has a resolution of 0.1mOhm on measurements up to 200mOhm -> should be accurate enough for measuring of battery internal resistance, which is usually less than 10mOhm for LiFePo4 cells.

But I see nobody does these measurements...

[MAGICPIE3FOCUSPOWER]

I do!
I was into RC cars hobby.
Sometimes it's very helpfull to check the internal resistance of the lipo's.

[Just]

And what about LiFePo4 cells? How much a battery resistance should be increased so that a decision "replace it!" be made?

Thank you!

[MAGICPIE3FOCUSPOWER]

I don't know for lifepo4.
But more resistance: more heat and less output (current and capacity).

Just using them till they can't power up your setup

[Just]

I just tried to measure an internal resistance of my LiFePo4 cells with an external register. I have to say, it's not an easy job... My fully charged cells (balanced by BMS) have 3.33V unloaded (no load connected). But, with the 1.2(Ohm) external register, the cell voltages drops to about 2.8V (not stable value) and then continuously vary from 2.54V to 2.9V.
So, if I'm taking 2.8V as a cell voltage under the load, then its internal resistance should be equal to about 228(mOhm) - too big! According to the cell spec, it should be less than 10(mOhm)!!!
I'm frustrated... 
Bad cells or bad measurements? Hard to say... Probably both of them?

[Bikemad]

Dmitryl, something is not quite right with either your cell or your figures.

If you take a look at the attached data taken from my cell voltage monitor & logger, you should be able to see that my LiPo cells have a typical initial voltage drop of less than 0.10V with up to a 32 Amp load , yet your LiFePO₄ cell appears to have a 0.5V drop with less than a 2.5 Amp load.

I would suggest that your cell is extremely weak, unless your readings are totally wrong due to poor connections between the voltmeter and the cell or between the resistor and the cell (or possibly both).

Just out of interest, I calculated my cell's internal resistance to be approx 2.5 mOhms for a maximum 0.8V drop with a 32.11 amp load:

V1 = 3.974V
V2 = 3.894V
R = V/I = 3.894/32.11 = 0.1212706

Ri = (V1-V2)*R/V2 = (3.974-3.894)*0.1212706/3.894 = 0.08*0.0311429 = 0.00249 Ohms (2.49 mOhms)

Alan
 

[Just]

Alan, thank you for your post!

As for the Cell Voltage Monitor & Logger graph, it only represents Voltage variations over the time. But, how did you measured the Load Current and correlated it with Voltage Drops? Honestly I'm wondering to see Voltage vs Current Graph... 

Anyway, it seems that the Cell Voltage Monitor & Logger is extremely useful device and I'm going to order it as well :-)

BTW, why do you prefer LiPo cells over the LiFePo4 for your bike? isn't dangerous?

Thank you!

[Just]

As for the battery cell impedance, is it reactive or passive? I mean, does it depend on the current flowing through the cell?

Does anyone talk about induction of the cell (besides internal resistance and capacity)?

[Bikemad]

Dmitryl, I also had a pair of Turnigy watt meters connected to the battery during the test so that I could measure both the maximum current draw and the maximum regen current.

I can't say for sure which of the lows in the above graph relate to the 32.11 Amp load, but I know that the maximum load occurred within that particular time-scale of the testing, so it must have coincided with one of the low points shown.
For the calculation in my previous post I used the low point which had the largest overall voltage drop, assuming it would have resulted from the largest current draw.

I use LiPo because they are comparatively cheap, plus I'd already gained a fair bit of knowledge from R/C model use.
The pack used in the above test is made up of two of these 5000mAh Hard case LiPos connected in series:



The two batteries cost $42.76 plus delivery and the tiny frame pannier that they snugly fit into was only $6.44 delivered.
Incidentally, the frame pannier was originally purchased so I could make up a second 14 cell pack, but when it arrived it was too small for the batteries to fit into, as it was considerably smaller than the first one I purchased from the same seller, even though the auction details were still identical. 



Fortunately, the hard case LiPo packs are shorter than the standard soft pouch Lipos, so I was able to fit both batteries, an in-line 30 Amp fuse and a pair of watt meters into the bag for a neat and compact installation which only weighs 1.4 kg  (3.09 lbs) in total, which can easily be removed from the bike when necessary.

As far as being dangerous is concerned, apart from physical damage, the two main causes of explosions tend to be excessive discharge current and excessive charging current/voltage. The packs are rated for 20C continuous discharge and 30C peak (100-150 Amps) so the 30 Amp fuse should prevent this from being exceeded as the result of a shorted cable etc.
The 5C charge rate (25 Amps) is much greater than either the regen or my 7 Amp charger can supply, and I always monitor the packs carefully during the charging process, so I have tried to cover the most dangerous aspects of using these particular cells.

This 5Ah pack is used mainly for shorter runs, so I'm unlikely to use much more than around half of it's capacity, but the audible alarm function on the battery monitor would alert me if I start to drain it too much anyway.

Alan
 

[Just]

Alan, are you running on the 24V system?

Could the LiPo cells be exploded just either from strokes or kicks or falling on the ground (which may often happen during the ride)?

As far as I understand, you don't use BMS when your cells on the bike and only your charger balance the cells during their charge... Correct?

In fact, LiFePo4 cells weigh twice as LiPo once...

Thank you

[Just]

BTW, LiFePo4 10Ah 10C 3.3V cells on the eBay available just for $27USD per cell shipping included. It seems that LiPo cells/packs with the same characteristics cost the same. Am I wrong?