Hi all,
After testing a few inexpensive Li-Ion chargers that turned out to be using some strange and incorrect ways to charge the batteries, I decided to try to build a charger for myself that would use a dedicated IC. I chose to go with the Microchip's MCP73861. I think no one described it previously on CPF so I thought I would share the experience and use it as an opportunity to ask some questions.
The MCP73861 is an integrated charge management controller. It is relatively inexpensive (~$3), capable of charging with any current between 100mA and 1.2A (programmable), has programmable safety timers and thermistor input. It is specified as a true CC/CV charger with the voltage regulation at 4.1V or 4.2V (selectable) and ability to precondition (i.e. start with a trickle charge) deeply discharged cells. The MCP73861 offers acceptable voltage regulation accuracy of 0.5% (so ~0.021V) and has two LED outputs to inform about the status of the charge. As it is usually the case with the charge management ICs, the voltage regulation accuracy doesn't depend on the accuracy of the external components so cheap, low accuracy resistors and capacitors can be safely used. Finally, the IC accepts input voltage from 4.5V to 12V (13.5V absolute max.) so it can be used with a wide variety of power sources like USB ports (with appropriately low charge current), car adapters and different power supplies.
The circuit has everything that's needed to charge a single Li-Ion battery already built in. One has only to add two filter capacitors, one timer capacitor, temperature output resistors, current programing resistor (or a few of them, if user selectable current is required), a status LED and two resistors for voltage regulation selection and logic enable pins to have a fully functional Li-Ion charger. The MCP73861 is available in a SOIC16 package so it's relatively easy to solder.
As you can see from the description above I was quite satisfied with the IC specification so I bought it and built a prototype to test how it really behaves. I started from checking the safety timers and thermistor functionality, which behaved as expected. Then I proceeded to looking at the voltage and the current during the charge:
I made other tests with different batteries at different states of charge and the charger behavior is consistent. It starts at the preconditioning current level of 10% of the selected CC current (63mA at the plot above) immediately after this it goes to the CC current level (630mA). So far so good, but then it drops down fast to ~75% of the selected current and then slowly raises the current until the end on the CC phase. The CC phase ends at around 3.14V and then the charger proceeds to charging with the lowering current and the voltage raising up to 4.18V-4.19V.
This charging algorithm looks quite good to me but it is not exactly the CC/CV as described in the IC datasheet. I made this prototype on a breadboard but the behavior seems too consistent and thought out to be just a result of using too long cables. Do you have an idea why the charger behaves this way?
I couldn't find precise measurements of other CC/CV chargers - does anyone have them or know where to find them for the Pila IBC, Schulze or other respected chargers?
The last question I have is about the 4.1V level - is there a point in selecting it instead of the usual 4.2V level for LiCoO2 cells? Would I gain cell longevity at the expense of slightly undercharging the cells?
After testing a few inexpensive Li-Ion chargers that turned out to be using some strange and incorrect ways to charge the batteries, I decided to try to build a charger for myself that would use a dedicated IC. I chose to go with the Microchip's MCP73861. I think no one described it previously on CPF so I thought I would share the experience and use it as an opportunity to ask some questions.
The MCP73861 is an integrated charge management controller. It is relatively inexpensive (~$3), capable of charging with any current between 100mA and 1.2A (programmable), has programmable safety timers and thermistor input. It is specified as a true CC/CV charger with the voltage regulation at 4.1V or 4.2V (selectable) and ability to precondition (i.e. start with a trickle charge) deeply discharged cells. The MCP73861 offers acceptable voltage regulation accuracy of 0.5% (so ~0.021V) and has two LED outputs to inform about the status of the charge. As it is usually the case with the charge management ICs, the voltage regulation accuracy doesn't depend on the accuracy of the external components so cheap, low accuracy resistors and capacitors can be safely used. Finally, the IC accepts input voltage from 4.5V to 12V (13.5V absolute max.) so it can be used with a wide variety of power sources like USB ports (with appropriately low charge current), car adapters and different power supplies.
The circuit has everything that's needed to charge a single Li-Ion battery already built in. One has only to add two filter capacitors, one timer capacitor, temperature output resistors, current programing resistor (or a few of them, if user selectable current is required), a status LED and two resistors for voltage regulation selection and logic enable pins to have a fully functional Li-Ion charger. The MCP73861 is available in a SOIC16 package so it's relatively easy to solder.
As you can see from the description above I was quite satisfied with the IC specification so I bought it and built a prototype to test how it really behaves. I started from checking the safety timers and thermistor functionality, which behaved as expected. Then I proceeded to looking at the voltage and the current during the charge:
I made other tests with different batteries at different states of charge and the charger behavior is consistent. It starts at the preconditioning current level of 10% of the selected CC current (63mA at the plot above) immediately after this it goes to the CC current level (630mA). So far so good, but then it drops down fast to ~75% of the selected current and then slowly raises the current until the end on the CC phase. The CC phase ends at around 3.14V and then the charger proceeds to charging with the lowering current and the voltage raising up to 4.18V-4.19V.
This charging algorithm looks quite good to me but it is not exactly the CC/CV as described in the IC datasheet. I made this prototype on a breadboard but the behavior seems too consistent and thought out to be just a result of using too long cables. Do you have an idea why the charger behaves this way?
I couldn't find precise measurements of other CC/CV chargers - does anyone have them or know where to find them for the Pila IBC, Schulze or other respected chargers?
The last question I have is about the 4.1V level - is there a point in selecting it instead of the usual 4.2V level for LiCoO2 cells? Would I gain cell longevity at the expense of slightly undercharging the cells?
