Teardown: Anatomy of a Failed Aftermarket Laptop Battery

Just the other day, I had an excuse to pull out my seven year old Toshiba Satellite L30 laptop. I’ve had it since the first day of my undergraduate uni career, and it’s definitely seen better days. It’s had one screen replacement under warranty, lost a few keys, has flaky USB ports, a broken power supply (which I replaced with another) and had the RAM and Hard Drive upgraded.

But overwhelmingly, what has been replaced most was the battery! In five-day-weeks of university away from powerpoints, it’s a case of cycle life hell. The original battery lasted three years, and all the replacements have lasted two.

The replacement batteries have always been aftermarket batteries. Often available for $30-60 instead of the $200 demanded by the OEMs, they represent sensible value for a laptop which would be otherwise obsolete and not worth investing money into.

A dead battery

The laptop had sat for about a year without being charged. That, in itself, is a bit of a danger point for any lithium ion battery. The circuits within the laptop and within the battery consume a small parasitic current which works to drain the battery. Lithium ion batteries do not like to be fully discharged and can be permanently damaged by this.

That being said, properly engineered battery safety systems can cut off any load from the laptop to minimise the parasitic load when the battery has reached empty, and the parasitic load from the protection system is likely low enough to allow the battery a fair amount of time before it is permanently disabled. If you catch a battery in this time, it should be rechargeable and the system will recognize the battery just fine. If your battery isn’t detected at all by the system, it might be too late, and the battery has locked out because over discharge promotes the growth of dendrites which could cause short circuits and thermal runaway – leading to spectacular fires and explosions.

This battery seemed alright though. I plugged the laptop in and the charging indicator came right up. The battery began to charge, but then, strange things started happening. The charge stagnated at 85% and never completed, and the charge voltage bounced up and down when monitored with HWMonitor.

The battery is sick. The nail in the coffin? When the power adapter is disconnected, the laptop dies right away. It’s not working.

Taking it apart

This is one of many dead batteries – but since I never had the time or daring to take them apart before, there hasn’t been any postings about it. But now that I have a blog – I wasn’t going to let this opportunity go to waste. Laptop batteries are actually one possibly great way to salvage some workable 18650 lithium ion batteries. Almost all bulky non-Apple laptop batteries utilize 18650s. It would be a good chance to also take a chance to try and work out what the real capacity is.

The battery itself is an eight cell design in a 4s2p configuration. There is a blue protection and gas gauge PCB in the bottom, marked APOKIN APK-TS3420Y-BQ3060 V/A1 D20101023. Already, quite heartening is the connections between each series set as that implies balanced charging which keeps the batteries at the same level of charge, preventing “weaker” cells from causing the early death of the whole pack. Also rather great is the thermistor which is secured to the surface of one of the cells which will help temperature compensated charging and possibly safety.

Aftermarket Toshiba Laptop Battery Taken Apart

It claims to be a 14.4v 5200mAh battery pack, although the cells themselves are relatively anonymous light purple cells with no markings on the outside of the jackets.

Anonymous Cells in Aftermarket Laptop Battery

I should mention that, from experience, I have learnt to look three times before sliding my cutting tools into places. It’s so easy to slice through a jacket, bridge the terminals and get a good high-current short that melts tabs and or otherwise causes smoke to form or even an explosion in the worst case. It’s also very dangerous as you can easily generate large amounts of heat which can cause burns. If you’re taking batteries apart, do be extra careful with your tools and where you put them. I took the time to strip out all the tabs on each of the cells to separate them and measure the voltage – that’s where I got a surprise

Disassembled Battery Cell Voltages

Stand back – very far away. One pair was allowed to get down to 1.025v (approximately) which is definitely over-discharged. The balancing doesn’t seem to have worked properly either as the cells differ widely in voltage. And it seems the circuit allowed most of the pack to be charged, when one of the series sets was too discharged and is permanently damaged. Unfortunately, those cells post a safety hazard – especially with anonymous branded cells, it could explode spontaneously for all I know. And it could have done that on my lap. With a large number of Lithium Ion cells in people’s houses, and many cheap products incorporating them from Chinese manufacturers, it could only be a matter of time before the probability of encountering an explosion or mishap becomes high enough to scare people.

Anyway, I was still daring, so I decided to de-sleeve the two obviously failed cells to see if there were any printed markings on the cells themselves – and there was – 1826-C3L29-FX. Unfortunately, searches on that string brings us no further information on the actual maker or capacity.

Cell Marking under Jacket

In the process of de-sleeving, I decided to take off the vent cover rings to find something rather distressing – a sign of corrosion and electrolyte leakage from one of the two low-charge cells. I also saw that on some of the better charged cells as well – implying maybe problems with the vent seals, charge controller circuitry programming, or just pure construction quality problems.

Corrosion Deposits near Vent

Others were nice and clean, but I wasn’t going to risk it. All of these cells are not going to get my attention – it’s just not worth the risk. It’s fairly clear that the cells themselves will fail in short order even if they’re still somewhat functional right now.

Clean Cell Top

Quality OEM batteries often have branded cells – those from Sony, Samsung, Sanyo, Panasonic tend to be fairly common and these are of higher quality and are less likely to see this kind of thing happen to them.

PCB Top

Here’s a better view of the top of the battery protection PCB. It’s pretty bare, and the name APOKIN doesn’t really point to a big known company at all. The date is 23rd October 2010 for the PCB, and it gives away that the PCB uses the Benchmarq BQ3060 (now TI) charge controller and gas gauge chip. The TS3420Y appears to refer to the Toshiba **3420*-1B** series of batteries which this PCB intends to replace. There are connections (from left to right) for ground (B-), thermistor (two connections), tap 1, tap 2, tap 3 and battery positive (B+). Only some of the connections have epoxy on them.

Battery Protection Board Underside

On the underside, it’s clear to see the charge controller chip and the SMD resistor used as a current measurement shunt. There’s also a mounted through-hole battery protection fuse (which, normally is blown by a serious fault or on command in case of anomaly detected by the charge controller to permanently disable the pack). I’m surprised this did not happen.

It appears that there is some white silicone rubber over two SMD chips – lets take a closer look:

SMD MOSFETs

They appear to be both 4825P P-channel MOSFETs used for switching purposes – the silicone may form some insulative protection against high humidity potentially?

What’s clear is that there is no secondary protection – e.g. a BQ29415 chip for independent verification of software based over-voltage protections. This appears to be a “common sense” safety precaution in designs from OEMs in order to ensure the highest level of safety for their batteries.

Conclusion

Well, the audacious claim of 5200mAh could only be realized if each cell was 26oomAh in capacity, which I highly doubt. This is probably a 4000-4400mAh pack from my estimates and that might even be a bit generous because it definitely fell short on cycle life having been very little used.

More importantly is to remember that the cheap batteries you purchase aftermarket are cheap for one of several reasons:

  • They appear to often overstate capacity and under-deliver.
  • They often use lower quality lithium-ion cells which may not last as many cycles before failing in mysterious ways.
  • They may omit extra safety features which may cause fires and explosions (for the minority who are unlucky).

This is probably not the worst battery I’ve seen – after all, it appears to have battery balanced charging ability and software level protections with a hardware fuse and thermistor compensated charge voltage. But it doesn’t have any second level protections, and it seems that the cells themselves have “defeated” the balancing ability of the charger – i.e. they may have internal shorts developing already causing current leakage or severe reduction in capacity due to loss of electrolyte.

That being said, buying a cheap battery often makes economic sense – that is, until you become the unlucky victim of one. I count myself lucky so far

About lui_gough

I'm a bit of a nut for electronics, computing, photography, radio, satellite and other technical hobbies. Click for more about me!
This entry was posted in Computing, Electronics and tagged , . Bookmark the permalink.

4 Responses to Teardown: Anatomy of a Failed Aftermarket Laptop Battery

  1. ginbot86 says:

    Good to see I’m not the only one out there that decided to look inside one of these cheap laptop batteries!

    I had a similar 12-cell HP battery a while back, and it was also using Apokin-branded PCB, also using the bq3060. Initially the battery had more than the designed capacity (about 9000 mAh compared to the designed 8800) after 100 cycles, it held 800 mAh (so 200 mAh/cell) and 1.8 ohms internal resistance per cell.

    The fuse wasn’t blown in these packs because the fuse isn’t able to be blown like OEM packs. The ‘chemical fuse’ used in the OEM batteries has a third terminal that is wired to a heater in the fuse, and that provides the fuse’s ability to be blown by the battery-management chip.

    I have Texas Instruments’ tools to interface with this battery and curiously enough, the permanent-failure tracking feature was enabled but even with a severe imbalance (the voltages were fine except one set was about 2 volts) and the chip didn’t report a failure. Unfortunately, Apokin (or whoever made the pack) locked the bq3060 and changed the password so I wasn’t able to get all the detailed info that was logged on that chip.

    • lui_gough says:

      Many thanks for that very useful snippet of information. Some of them really are made with low-grade Chinese unbranded cells which mysteriously fail, despite quality charge controllers etc. The protection being incomplete is a common sight amongst cheaper clones. Glad to see you’ve managed to obtain the TI utilities to diagnose your battery – I didn’t even think of doing that! Maybe next time …

  2. texaspyro says:

    Take a look at this open source Arduino battery pack data dumper. It connects to the pack SMBUS (IIC) pins and dumps the data from the battery management chip:

    http://budgetlightforum.com/node/35383

  3. Pingback: Teardown: Another Dead “Compatible” {Laptop, Camera} Battery | Gough's Tech Zone

Error: Comment is Missing!