Operation of simple Li-ion Battery Charge System

Question

I am testing out a cordless screw driver. When I first plugged it into the charger a red LED light was suppose to light but it didn't. After a couple hours I opened the unit up and checked for circuit problems,

The SOT89 in the center right of the circuit board is a B772 NPN transistor and it controls the charge current. Below it, the small transistor is a 2N7002 ( an N Ch MOSFET) and it is the driver that controls the B772. The U1 which I haven't IDed is probably the charge controller. When I first opened the unit I had 5 volts on the emitter of the B772 and no volts on the Collector. The Li-ion battery was sitting at 0.6 volts. In the course of my testing I eventually put a small token charge into the battery and at this point the unit started charging as normal.

My question is this: Did the extremely low charge level of the battery cause the charge circuit to default to a no charge situation or is there likely a secondary fault in the circuit that has just decided to temporarily heal itself? Any insights would be appreciated. I have minimal experience with this type of charging circuit.

John

genebren · Accepted Answer

John, I am working on a Li-ion battery charger right now (and I have developed others in the past). Most of the Li-ion chargers have a 'battery check' (injecting a small current into the battery and looking for the voltage at the battery terminal to stay below the charge termination voltage, i.e 4.2V) process that must pass before the device goes into the charging mode. The next step is usually a preconditioning phase, where again a low current (% of normal charge current) is applied to the battery until the battery voltage rises to the proper level at which point the battery enters a constant current charge. The charger chip that I am using is a Microchip MCP73831/2. This chip turns on the 'charging' LED as soon as the device enters either the preconditioning or charging stage. Not all devices have the same logic, but if my charger failed the 'battery check' the device would sit with the LED off. Having used other charging chips, I have seen a fair number of chargers fail to charge batteries. One of the reoccurring failure that I have seen has been due to a high impedance path (connectors, switches, wires, etc.) on the charger input voltage. As the charger enters 'charging', the current rises quickly and the 'high input impedance' causes a large voltage drop, causing the charger to shutdown due to insufficient voltage. What happens is the charger chip begins to oscillate and the charger never delivers charge to the battery (sometimes actually further draining the battery). As pointed out by mcb1 , operating/allowing Li-ion batteries to drop to this low of a voltage is very bad for the life/performance of the cell. Keep an eye on this drill and/or replace the battery or return it as damaged to the vendor. Good luck, Gene

mcb1 · Answer

You should know by now that anything that is broken, always works when you take it to get fixed ... It wouldn't surprise me if the low voltage caused the charging to shut down. Lithium Ion cells shouldn't go that low, so don't be surprised if the battery performance is reduced either in capability or life. The best source of information is here. http://batteryuniversity.com/ Charging information http://batteryuniversity.com/learn/article/charging_lithium_ion_batteries Mark

jw0752 · Answer

Hi Shabaz,

I actually jump started this one by bridging the charge regulator with a 10 ohm resistor for 10 seconds. That was apparently enough for the controller to sense the battery and take over. According to common wisdom the battery is probably damaged but I will use it a while to see how it goes.

John

jw0752 · Answer

Hi genebren , Your recommendation of the MCP 73831/2 Li-ion charge management chip is very appreciated. I have finally received some of these chips and I have begun to experiment with them. Since they are SMD I mounted them to adapter boards so that I could use them on the bread board as DIP 6. Here is the bread board with two individual chargers set up. I used the typical application schematic from the data sheet to wire the circuits. This chip is very user friendly. It is programmable so that the Max charging current can be controlled by changing a resistor between the Prog pin and ground. The optimal charging current is between 1/2 and 1 C. By using a 2.2K resistor as the program resistor the constant current during that phase of the charging is 450 mA. I experimented with other values and settled on using a 3.9 K resistor as my little adapter boards did not provide much of a heat sink for the chip. The chip is thermally controlled so that the chip will self protect to a lower current if the junction temperature gets above 150C. Because of this protection my chip was running at 250 mA even though I had it programmed for 500 mA. By changing the program resistor to 3.9K I just brought the charge level back to a point where the junction temp stayed below 150C. The program resistor can also be used as an enable/ disable to the charger. If the resistor is left floating the circuit shuts down. This would enable interfacing with a micro controller. The chip has many conditions that it checks and then decides how best to charge the battery. If the battery is very low it initiates with a precharge protocol to get things started and then transitions to constant current. The point when it transitions to constant current is when there is the most heat stress on the chip. The supply voltage Vin is typically between 4.5 and 6 volts. With the battery at a low voltage level and the current at the program limit there may be as much as a 3 volt drop at 500 mA across the chip at this point. This is a lot of wattage to shed from a small SOT23-5 package. By automatically lowering the charge current as needed to control the heat the chip charges the battery and as the voltage of the battery increases the heat problem ameliorates. When the voltage of the battery approaches the terminal voltage the chip switches to a constant voltage phase and eventually shuts down when 4.2 volts is reached. The chip has versions available for 3 different terminal voltages. The final digit on my chips "2" designates them as 4.2 volt chips. The chip also has a status pin that will turn an LED on or off to give a general indication whether the chip is in standby or charge modes. I want to thank you for the fun I had exploring this neat little charging tool. The great benefit for me in the forum is the inspiration and guidance that I receive from interacting with everyone here. John

mcb1 · Answer

I've seen this type of residue on some items before. My thought is that the flux hasn't been cleaned off fully, and with some moisture it has turned into this. I found it when my parents had a cordless phone that the company wouldn't repair under warranty because they claimed it had water damage. My parents asked me to look at it, I cleaned it up with meths/alcohol and it worked for many years after that. Mark

shabaz · Answer

Hi John,

I believe I've hit this issue a couple of times with new products too : (

One was a wireless keyboard, which was a model presumably in a warehouse for too long, and it refused to charge, and the replacement had the same issue. Another was a label printer, which was new but sold as faulty (at a lower price) because it wouldn't charge either. In that latter case, it had a removable battery pack inside (NiMH I think from memory).. so I removed it, carefully gave it a bit of charge, and it's been fine ever since.

genebren · Answer

John, The programming resistor sets the maximum current value (2K = 500mA). In normal operation, when the charger is applying the maximum current, the voltage on the PROG pin will read 1.0V. If the charger is limiting the current to 250mA, or 50% of the maximum value, the voltage on the PROG pin will read 0.5V. The same is true for your case, with a resistor of 2.2K (450mA max), a reading of 1.0V on PROG means the the current is 450mA, while a reading of 0.5 means that the current is 225mA (or 50% of maximum). So what ever the programmed maximum current, the voltage on PROG gives you the relative current. Gene

shabaz · Answer

Hi John, I agree, some of these batteries are pricey, worth attempting this process : )

genebren · Answer

John, Great job on the charger and an excellent description of its functionality and operation. The specification sheet for the device does a good job of describing the operation, but it also obscures some very valuable information. The PROG pin - current programming and floating the pin to disable the device. Yes, but a more useful fact that the specification does not highlight well is that you can also drive this pin to VCC to disable it. This is a bit more microprocessor friendly, as you can tri-state a connected GPIO and then switch it on and drive it high to disable it (i.e you do not need to psychically remove the programming resistor to disable) . Also, you can monitor the voltage on this pin to determine the actual charging current. A reading of 1.0V represents 100% of the programmed current value and a reading of 0.0 would represent 0% (although the device never really drive below 5%. Glad you had a little fun with this chip. Gene

jw0752 · Answer

The board you are looking at was after I cleaned it up. Here is what it looked like before: The battery was my first thought too but there was no sign of leakage on the battery itself. John

jw0752 · Answer

I reuploaded the picture in case it still wasn't coming through. John

jw0752 · Answer

Hi Gene, Thanks again for your suggestion to check out the MCP73831-2. I have ordered a few so I can experiment. John

jw0752 · Answer

Hi Gene, You are correct those are good things to know. I did not like the idea of disconnecting the resistor as this did not feel very elegant. Much better to drive the resistor closer to Vcc. A quarter watt 2 K resistor will be able to handle 5 volts easily and any program resistor greater than 2 K will have even less wattage to dissipate. I also like the idea that the charge current can be monitored by measuring the voltage on the program pin. This is much less invasive that using an ammeter. Thanks John

Operation of simple Li-ion Battery Charge System

Top Replies