Li-Ion State of Charge and Voltage Measurements

[SilverFox]

There has been a lot of discussion on how to figure out the state of charge on Li-Ion cells by measuring their resting voltage.

I picked up some information on high current draws that gives the following values:

4.2V – 100%
4.1V – 87%
4.0V – 75%
3.9V – 55%
3.8V – 30%
3.5V – 0%

Please note that resting voltage means the cell has stabilized at room temperature and the voltage has also stabilized.

I decided to check a brand new 18650 cell at a defined current draw. This cell is a Pila 600P rated at 2200 mAh. The test current was 2 amps with a low voltage cut off of 2.8 volts.

At a 2 amp current draw, this is what I observed:

4.20 volts – 100%
3.97 volts – 80%
3.85 volts – 60%
3.77 volts – 40%
3.72 volts – 20%
3.58 volts – 0%

This cell tested at 2000 mAh capacity at 2 amps. I ran 2 amps for 400 mAh, then stopped the test to let the cell and voltage stabilize. I then continued to do this 5 times to come up with the values listed.

Luna brought up a question involving the state of health of the cell. I got to thinking and realized that I have a Sony 18650 that has been abused by over charging, over discharging, and has been cycled to near death. I charged it up and ran a similar test on it.

On major difference was that I discharged at 1 amp.

Here are the results for a "damaged" cell, that should be thrown away, at a 1 amp current draw down to 2.8 volt cut off:

4.20 volts - 100%
3.95 volts - 80%
3.85 volts - 60%
3.68 volts - 40%
3.45 volts - 20%
3.31 volts - 0%

It is interesting to note that at this lower discharge rate, there was less voltage bounce back during the last half of the discharge.

Tom

 

[goldserve]

I find this hard to believe. I always thought that cells from manufacturers were 80% changed, 3.8V.

Also, why is LVC sometimes at 3V and as low as 2.7V as well...

Cheers!

 

[Luna]

Also, why is LVC sometimes at 3V and as low as 2.7V as well...
Cheers!

Coke vs Graphite cells. The older coke units (like sony) required going as low as 2.5v to provide the speced mah.


I still belive the SOH is a pivotal variable in the the determination of SOC but Tom has provided data to me in the past that seems to not be the case.

 

[wptski]

I find this hard to believe. I always thought that cells from manufacturers were 80% changed, 3.8V.

Also, why is LVC sometimes at 3V and as low as 2.7V as well...

Cheers!

I wish LVC was 3.0V! At one time it was thought that the proper LVC was 2.5V and there are many PCB out there that still use that, I think that it's the chip that they use. RC type Li-Poly battery manufactures found out that 2.5V was too low and raised it as time went along. It's at 3.0V at least if not higher now.

Thanks Tom for posting the list!

 

[SilverFox]

Hello Goldserve,

Manufacturers ship their cells at a 40% state of charge, or around 3.8 volts.

Over discharging a cell can do damage to it and make it unstable during the next charge cycle. The low voltage cut off that a cell can be safely discharged to depends on the current draw. For low current draw applications, 3.0 volts is used. For higher current draws, you can drop it down a bit. The protection circuit usually kicks in at around 2.3 volts, but this is a last stop before destroying the cell. I do not recommend running to the protection circuit cut off all the time. It is there to protect you if you miscalculate your run time.

Li-Ion cells thrive on frequent top offs. If you treat your flashlight batteries similar to how you treat your cell phone batteries (toping them up after each day of use) you will get very long cycle life from them.

Tom

 

[SilverFox]

Hello Luna,

Please note that I specified that this data was taken from a brand new cell...

Tom

 

[cy]

Thanks Tom, this jives with what I've experienced too.

 

[wptski]

Luna:

I believe the coke/graphite thing has to do with the 3.6v/3.7V per cell value.

 

[wptski]

Tom:

How about two JSB RCR123 3.7V first generation Li-Ion in a 3W DB970 light which has repeatedly self discharged to the point of not working? This is the third of fourth time with different cells which I have used in other configurations with no self discharge problems as yet. This time one is 4.05V and the other is 1.7V!!!

 

[SilverFox]

Hello Bill,

1.7 volts seems a bit low...

Does the cell self discharge when not in a light?

Tom

 

[Luna]

Luna:

I believe the coke/graphite thing has to do with the 3.6v/3.7V per cell value.

Has to do with both

Here is the first item I clicked just now from Google:

http://www.buchmann.ca/Article5-Page1.asp

"As a result, the useful energy of the graphite system can be retrieved by discharging only to 3.0 volts per cell, whereas Sony's coke version must be discharged to 2.5 volts to get the same performance. "

 

[Luna]

Hello Luna,

Please note that I specified that this data was taken from a brand new cell...

Tom

Understood. I just wanted to mention SOH (State of Health) for the sake of this thread.

 

[wptski]

Hello Bill,

1.7 volts seems a bit low...

Does the cell self discharge when not in a light?

Tom

Tom:

Well, that's the odd thing. I've taken those same cells, put them in a different light and there's no problem. I put different cells in the problem light and the problem comes back although this time it was a long time before it happened again. I thought that I had it licked by cleanig all the threads up and since it has a McE2S switch I used NyGel non-conductive grease. There is one type which is used on switches, etc. which was used on the threads which is a electrical path. There is one "O" ring is used as a insulator in the switch , so I made sure that the grease used was of the non-electrical type.

So much for that theory! I posted this when it happened before in the McGizmo thread and it stumped them there too!!

Oh Yeah! All my other JSB RCR123 cells stored at around 3.8V have held up. I haven't checked all my RCR123 cells yet.

Funny, I had another light configured the same way with two Li primary cells in my truck go dead too but that may have been a fluke since they are still good now.

 

[wptski]

Has to do with both

Here is the first item I clicked just now from Google:

http://www.buchmann.ca/Article5-Page1.asp

"As a result, the useful energy of the graphite system can be retrieved by discharging only to 3.0 volts per cell, whereas Sony's coke version must be discharged to 2.5 volts to get the same performance. "

Luna:

Being half right is darn good for me!

 

[SilverFox]

Update:

Luna's question about the state of health got me thinking...

I have added the results from a damaged cell that I have that is near the end of its life.

Tom

 

[LuxLuthor]

Thanks for linking this thread in the other recent topic. It sounds like a reliable way to make sure your Li-Ion cells are at the ideal 40% level for storage is to check their resting voltage, and have it be about 3.7-3.8V

That is something I was wondering how to figure out when storing.

 

[Muse]

Thanks for linking this thread in the other recent topic. It sounds like a reliable way to make sure your Li-Ion cells are at the ideal 40% level for storage is to check their resting voltage, and have it be about 3.7-3.8V

That is something I was wondering how to figure out when storing.

That was my reason for starting the other thread.

 

[VidPro]

There has been a lot of discussion on how to figure out the state of charge on Li-Ion cells by measuring their resting voltage.



It is interesting to note that at this lower discharge rate, there was less voltage bounce back during the last half of the discharge.

Tom

ok so its an old thread, its a good one
Hi SilverFox, i have a question, or clarification.
are ALL of the voltages listed in the original post in this thread, RESTING voltages? and this is ALL about the resting voltages?
SO even though you have different discharge Rates, these are not under any type of LOAD when tested for voltage.

so this does not represent how the li-ion batteries tested will look under a load, JUST when RESTED right?
and
(just for clarification) these are Not discharge graphs, like the wonderfull actual discharge graphs as shown in this thread.
https://www.candlepowerforums.com/threads/117117
(Discharge Graphs)

(Only because i see this great data, which is really usefull, can/has been confused with how they work under a load.)

 

[SilverFox]

Hello VidPro,

Correct. Measuring the stable resting voltage can give you an estimate of the state of charge with a Li-Ion cell, or pack.

The discharge graphs show the voltage under load, for various loads. This is voltage under load, and not resting voltage.

Here is a formula that I have been playing with that is suppose to give you a reasonable estimate for cells in the 30 - 100% state of charge range.

State of Charge % = 100 (0.1966 + SquareRoot(0.0387 - 1.4523 (3.7835 - Vresting))

I believe the formula may give a closer approximation with cells used at lower loads.

Tom

 

[dfred]

For those of you who spend their time in a Unix shell, here's a simple little script to do this. Will return math errors if voltage is less than about 3.76, which corresponds to approximately 30%.

#!/bin/sh

if [ "X${1}" = "X" ] ; then
  echo usage: li-ion voltage  
  exit 1 
fi

echo "scale=2 ; (0.1966+sqrt(0.0387-(1.4523*(3.7835-$1))))*100" | bc