Li-Ion charging algorithms theory. How flexible is it?

[linterno]

Li-Ion algorithm theory for charging Li-Ion batteries states that the batteries must be charged using Constant Current (CC) until the battery reaches 4.2 volts. After that point a Constant Voltage (CV), which I suppose must be 4.2 volts, must be applied to the battery until the charging current is 0.05C to 0.10C and then terminate the charge.

Some time ago I used a cell phone to charge a 18650 Li-Ion and noticed (doing some measurements with a Fluke 179 DMM) that the charging algorithm was far from the theory. This cell phone was constantly (all the time) delivering/providing 4.2 volts to the battery for 3 seconds stop charging for half a second or less and started again for other 3 seconds and repeating this until the battery voltage was closed to 4.2 volts (closed to 4.194 - 4.196) and then stopped the charge. I guess the charger stops for half a second to measure the battery voltage.

Well, I have been tinkering on a couple of charging circuits and I have found that the behavior is the same if I provide 4.2 constant volts to the battery. When the battery voltage is 4.1 or less then the current is constant (I am working with 750 ma). After that point the current starts to decrease as the battery voltage approaches to 4.2 volts. At 4.19 volts the current is closed to 70 ma, which is 4% of the battery capacity (Panasonic 18650 1800 mAh).

Using the simplest charging circuit (based on LM317T regulator) it stops the charge - it doesn't really stops the charge, but the voltage doesn't increase any more, even if the current is ~40- ma (trickle charge = bad thing, I guess) - when the battery voltage is 4.197 volts. After removing the battery from the "charger" and allowing it to rest for 12 hours the battery voltage remains at ~4.183 volts, something not bad considering that most Pila IBC charger users have written in the forum that their battery voltage remains ~4.17 volts after allowing it to rest off the charger for a while.

This is my current Li-Ion most used charger.

My questions are:

1. Is this 4.2 volts Constant Voltage (Current controlled by the battery voltage) a valid algorithm? If invalid, why?
2. What could happen to the battery using this procedure?

 

[Mr Happy]

As I understand it, the charging theory is fairly simple, but there are many circuit variations that can be used to approximate the design goal.

The theory says two things:

1. You must not exceed the maximum cell voltage (4.2 V)
2. You must not exceed the maximum charging current (e.g. 0.5C)

If you connected a 4.2 V regulated supply to the cell you would achieve constraint 1. But when the cell was discharged the initial current would be too high and would violate constraint 2.

Therefore, you can put a current limiting circuit on the 4.2 V supply so that the current cannot exceed the maximum permitted. This will then satisfy constraints 1 and 2.

Now comes the fuzzy stuff. Are the current and voltage constraints absolute and constant, or averages over time? For instance, if you pulsed the charge on/off with a 1C current and a 50% duty cycle, would that still count as a 0.5C charge limit? And if you pulsed the voltage between 4.1 and 4.3 V every half second with the current pulses, would that satisfy the max voltage constraint of 4.2 V?

If you look at real world chargers, you will probably observe all sorts of variations and "short cuts" like this in use. Within reasonable bounds, many of these are probably OK since they have been used so much in practice.

Lastly, the constant trickle when the cell is fully charged is a bad thing. So there should be a switch that positively cuts off the charge current once it reaches a certain minimum.

 

[SilverFox]

Hello Juan,

There are many adaptations of the Li-Ion charging algorithm. In the purest sense you want to regulate the current, as Mr Happy has indicated, and clamp the voltage at 4.2 volts. At the end of the charge, you need to shut the charge current off.

The charge rate is selected with the idea that the more time spent at 4.2 volts the more possible damage is done to the cell simply by spending time at 4.2 volts. If you charge at 0.5C, it takes longer to reach the CV stage and less time is spent at 4.2 volts. On the other hand, if you charge at 2C, you quickly reach CV and spend a long time there as the current drops off to the cut off value.

Many cell manufacturers recommend 0.7C as an "ideal" charge rate, but the RC people are interested in fast charging and they are using 2C charge rates. I don't think this is best for the cells that we use, but it seems to work well for the Li-Po cells that the RC people use.

The cell phone people are also interested in fast charging, but they are trying to balance that with long life. I think the realize that most people charge their phones while they sleep and they have optimized their charging algorithm to incorporate the needs for a quick top off with the overnight charge.

I think the key component of charging Li-Ion cells is voltage control. As long as your charger never exceeds 4.2 volts, your cells will last a long time. Charge current is also important, but since it quickly tapers off when reaching the CV portion of the charge, it is not as important as the voltage limit.

As far as the amount of time spent at 4.2 volts goes, I think this makes a good topic of discussion over a round of beers at the Pub, but in actual use I am not sure it makes all that much difference.

Tom

 

[linterno]

Thank you Tom and Mr Happy for your replies.

Mr Happy. Constraint 2 is being achieved by me not by the charger. I always check the battery voltage with a DMM before inserting them in the charger. The current limit of 750 ma, which in fact is 0.42C is limited by a current regulator.

One of my prevoious circuits used an initial 4.25 volts charge, but, I noticed that using that algorithm the battery reached 4.2 volts pretty soon. Then, when the voltage raised to 4.2 volts, the microcontroller dropped the charging voltage to 4.2, but, as soon as the charging voltage was reduced to 4.2 volts the battery voltage dropped to 4.17. Since it didn't work as expected I removed the microcontroller portion from the circuit. Now I am going to use the microcontroller to disconect the battery when the charging current is 0.05C or when the battery voltage is 4.19 volts.

Juan C.