As a result of visiting Hamfest, I ended up with a laptop to take apart – a fairly new Toshiba Satellite C675D with a broken screen. It’s not a Hamfest if you don’t bring home something to take apart of course! Today we’ll be testing the battery it came with to see if it’s salvageable.The date code says it was made in 11/2011
Label data indicates we are dealing with a 10.8V 4.4Ah pack. Since a typical Li-Ion cell is around 2-3Ah and 3.6-4.2V, that tells us the pack has 6 cells arranged as 3Series2Parallel (or 3S2P). The battery was fully depleted at the time of purchase and would not power the laptop. I could also not see any voltage present on any of its pins. Since we are dealing with a parts laptop here, an easy way to figure out full pinout is to look at the motherboard end of things:
It is pretty obvious that the larger blades are power and ground, and by measuring resistance to case ground we can tell that the leftmost two pins are ground. The next two pins (3,4) are routed through a SOT23 device that measures as a dual diode- most likely a dual Zener/TVS for ESD protection of comm lines. Pin 5 has a single channel TVS/Diode looking thing. That most likely means we have SMbus (a typical comm channel for notebook batteries) on pins 3 and 4, and a battery temperature on pin 5. Pins 6 and 7 remain mystery for now and pins 8 and 9 are the battery positive.
Next step is to try waking the battery up. We apply normal for it voltage (in the range of 3-4.2V per cell) and see if it “takes”any current. Nothing happens, no current is going in. Plan B- put the battery back in laptop and observe signals on the pin. With the laptop powered by a bench power supply, we can see a notable surrent spike when battery is inserted, so something does happen. Voltage on the pins is now 12.6V, meaning the charger is trying to do its job but no current is actually making it in. Poking around with a scope, we can see activity on pins 3 and 4. And pin 3 definitely looks like an I2C clock, pulsing nonstop during the comm intervals, while pin 4 looks like an I2C data. Ok we’ve just confirmed 3 and 4 are indeed the SMbus pins and the battery is talking to the charger.
That seems to indicate it is alive, but too discharged to allow for charging- most likely staying in precharge mode with a large series resistor limiting input current to very low values. Leaving things alone for a bit and checking back in a few hours, the power supply current went up drastically from about 0.5A at 19V to 1.5A. That may be a good sign- battery is now taking current. Voltage check confirms that- we are now at 11V and rising, meaning the fast charge constant current phase is on, and the battery is finally charging. Well, might as well let it finish. Once the laptop’s charge indicator goes from yellow to green and power supply current drops, we can start the tests.
First check is battery voltage and once again it is 0V. At the same time, I can plug into the laptop and it stays powered. Something else is going on- time to check if we are missing anything. It then occurs to me that some batteries have a signal called System Present, and only allow charging/discharging if they see that signal. Well, we still have pins 6 and 7 to figure out. Measuring the pin to ground resistance in and out of system we see that the laptop pulls pin 6 to ground when battery is inserted. A quick test with a resistor between pins 6 and 1 confirmed the guess- battery voltage appears on the positive terminals.
Now we can run tests on the bench. First we connect a TI EV2300 interface
and wire its SMB Clk, Data and ground to the battery. We then use TI’s evaluation software to look at the reported data. It is not known if the chipset inside the battery is indeed the part this program was written for or if it’s even a TI chipset in the first place, but luckily SMBus standard defines the base set of registers everybody has to support:
From SMBUs data we can tell that the battery cells are healthy, well balanced and fully charged. The pack saw 6 discharge cycles and thinks its full capacity is 4400mAh, just as indicate don the label. So far so good. Next we apply constant current load and run a discharge cycle to see how much capacity we can get out of the battery. Running at 1.5A with a 4.4 AH capacity, we should see close to 3 hours of runtime.
After 2.5 hours or running at 1.5A, the battery reported 0% capacity. It also adjusted full charge capacity down a bit to 4Ah.
That’s the amount of charge internal gas gauge counted going out during the last discharge cycle. So we have a 4Ah pack, out of 4.4Ah initial design capacity, or about 91% state of health. Seems like a reasonably healthy battery that should provide service for a while.
Links and resources:
For those interested, here are the things mentioned:
Pinout: 1,2 GND, 3 SCL, 4SDA, 5 Temp?, 6 SysPresent, 7-??, 8,9-BAT+
TI EV2300 http://www.ti.com/tool/ev2300
TI bqEasy software http://www.ti.com/tool/bq20z45-r1_bqeasy-sw