Battery life is always an issue with technological devices. At least, with many larger devices, you can still solve the issue with buying a spare battery to swap in and extend your runtime away from the powerpoint.
But of course, everything has a lifetime and eventually, the batteries will fail. Overall, I’ve had my fair share of failed batteries, and in general, low-cost compatible batteries fail more often than genuine OEM batteries. Every failure is an opportunity for a teardown and post-mortem, and that’s what this post is about.
HP PH06 Probook 4525s 9-cell Compatible Battery
Having bought a run-out clearance HP Probook 4525s back in 2011, I needed some more battery life than the two-hours offered by the stock LG-cell based 10.8v 4200mAh 6-cell battery. I ferreted these out from eBay, offered for sale at US$39.99 each back in December 2011. I purchased two, opting for the higher 7800mAh rating. Interestingly, this one was rated as 11.1v, but such differences are just due to the nominal voltage per cell being taken as 3.7v instead of 3.6v.
The battery was quite chunky, and had a pretty poor fit, meaning it didn’t like to clip into the laptop. It ended up breaking one of the battery latches on the laptop which was quite a disappointment, but it was used several times. It was very handy, but it hardly gave me the three-and-a-half hours I expected – it only really gave me about three, which wasn’t much for a battery of the size. I suppose that’s what you’d expect from an AMD-based laptop.
A while back, I noticed one of the two batteries had a highly non-linear discharge characteristic, which meant that as the cell approached the bottom 50%, it depleted much quicker than the first 50% even after recalibration. Charging it up and storing it charged seem to have it “self discharge” fairly quickly.
That’s the sign of a bad battery, and I marked it as such. When I came back to revisit it today, it was bad. It would not charge at all, and strangely, the battery management tool said it was only measuring a measily 5.2v! I suppose it’s old enough to retire, but the original battery is still going, for the record.
I had previously looked at a dead laptop battery a while back, so it might be interesting to compare. As a result, I decided to take it apart. The battery seams are very unusual, made out of three segments, all superglued together at the seams. Prying at it for over ten minutes with flat-head screwdrivers and by using side-cutters and pliers to crack the plastic apart, I was eventually able to get access. But this was not without some collateral damage, in the form of a pinched palm.
Already, a lie is visible. The battery is composed on a 3s3p configuration, but each cell is only rated at 2200mAh according to the writing, making the battery a 6600mAh rather than a 7800mAh. That’s about a 15% discrepancy. The cells themselves aren’t particularly quality cells either, marked ICR18650 3.7V 2200mAh PPK, manufacturer unknown.
I took the time to peel back the casing, and extricate the cells from the packaging and “silicone” goop it was glued in with, then untab all the cells. I noticed the tabs featured corrosion, which suggests cells may have vented their electrolyte at some point in their past. After that, I measured the cell voltages with my Agilent U1241B and labelled each cell.
A very interesting result – six of nine cells are permanently damaged by discharge quite far beyond their operating window. It’s likely that at least two cells (one in each bad set of three) may have gone bad by internally shorting, and as sets of three cells are hooked in parallel, the other two cells in the set may have dumped into the bad cell, resulting in the whole set being permanently damaged.
The last set of three seems to be healthy though, but I wonder for how much longer. Similarly, I wonder how much longer my other 9-cell battery will last, or whether it’s a safer choice just to retire it ahead of time given the “statistical” failure we have here.
Anyway, lets take a look at the circuit board that accompanies the cells:
The board has the blade style connector on it, but interestingly, not all pins are used. The board is “taped over” for insulation, and the battery has thermistor protection.
The cells are catered for with “balancing” connections at each cell. The PCB is marked SLN615(3) Rev. 2 8530. The top side appears to have a voltage reference zener, and two MOSFETs for protection.
The underside features the 0.1 ohm shunt measurement resistor, a smattering of capacitors, resistors and transistors and the “regular” Texas Instruments/BenchmarQ bq3060 gas gauge controller. The PCB seems to be configured for up to a 4-series configuration for batteries up to 14.4/14.8v.
Again, it seems the manufacturer has skimped on the recommended secondary protection and permanent fuse which provides additional safety and is almost always found on properly engineered Li-Ion battery systems.
Nikon EN-EL14 D3200 Compatible Battery
As it turns out, this is not my first bad compatible battery and it’s not my first design of compatible battery either. It doesn’t bode well for the supply of compatible batteries.
This particular battery is probably still on sale today, and was bought off eBay back in November 2013 for US$14.99 each. Foolishly, I bought a total of four of these cells – this battery is only a little older than one year old! Not a very good result.
Unlike the other failed battery which had partial capacity, this one was entirely dead and resulted in fast flashing on the charger almost immediately on putting the battery on the charger.
To disassemble, I carefully pried along the edges to separate the glued halves. It wasn’t as easy to break apart as I expected because of the use of double-sided adhesive tape to glue the batteries to the inside of the shells.
Internally, two metal-shelled prismatic lithium-ion cells were double-sided taped together and configured in a 2s1p configuration. The tabbing for the “centre tap” looks very interesting with its “wrapped back” design.
The sides of the cells had some stampings or laser etchings on it and they seemed to be mismatched for font. Again, it seems mixed batches of cells may have been used which may contribute to the early demise of the batteries. Careful peeling apart of the sandwich and removal of the green insulating cardboard revealed the two cells.
The cells are marked with RF053040AL/500 ADHQ. The 053040AL is a cell size designation, and the /500 implies 500mAh per cell, as is common for cells of this size. As a result, this battery is less than half the capacity that’s stated on the outside – 500mAh rather than 1030mAh.
Measurement of the voltages gives us a similar story to the laptop battery above – one of the cells has clearly failed and has somehow over-discharged or developed an internal short.
This is rather surprising given the connection to the PCB features balancing connection, meaning the active balancing by the protection circuitry couldn’t protect the battery from failing. Lets take a closer look at the PCB.
The PCB is marked ZYT-ENEL14 V1.4 RoHS.
The other side houses all of the ICs, including MOSFET, voltage references, etc. It seems the DRM protection is handled by the LatticeSemi’s SiliconBlue iCE65L04F-T FPGA. No proper protection in the form of a permanent fuse or secondary monitoring chip is visible.
While older portable devices using removable batteries have DRM systems which are generally well-emulated by third-party “compatible” batteries, these batteries still don’t meet the mark when it comes to implementing proper safety protection mechanisms.
The main let-down with the compatible batteries is the quality of the Li-Ion cells, which seem to come from mysterious no-name Chinese manufacturers which seems to have a unexpectedly high failure rate. The rest of the battery, in the form of the controller, is probably still perfectly healthy.