The IBM ThinkPad battery problem

IBM ThinkPads are generally good machines, but at least some models (including the 600 series) have defective battery designs that lead to premature battery death. The 600 is an old enough model that we've collected lots of data about the problem. For newer models we can't yet be sure, but it's probably wisest to avoid these models.

IBM's handling of this problem has been quite poor. I have tried to contact them multiple times, with no real response. This is extremely disappointing behavior, expecially for a brand which has traditionally stood for quality and support. I personally consider IBM's behavior to verge on the disreputable.

As a result, I do not recommend IBM ThinkPad laptops.

If you are stuck with an IBM laptop and its battery is dying, my sympathies. I don't have much advice other than to consider upgrading to a better brand.

31 Jan 2003: This page was linked from Slashdot. There is some interesting discussion in the comments. The consensus is that TP600 batteries die quite earlier than most (but not all) other laptop brands.

I've tried Valero's reconditioning approach with some success, but the battery seems to revert to its very short lifetime quite quickly, so I don't consider it a real solution.

25 Oct 2002: I sent a fairly strongly worded letter to IBM through the feedback mechanism. We'll see if it gets a response.

In the grand scheme of things, it's a small problem, but it bugs me nonetheless. I'm pretty sure IBM has made a defective product in this case, and they're not admitting it. I've gotten quite a bit of feedback on the issue, and have summarized it here. Finally, a bit of good news: Javier Valero seems to have figured out how to improve the battery life a bit. Read on.

Paul Hutchison reports that his 600E battery has the same problem. I wouldn't be surprised if there were other models similarly affected. Feedback is welcome: <raph@acm.org>

When your battery dies, you're most likely to buy a new one from the IBM online store, where you'll be confronted with this text:
Thanks to IBM's advanced power management systems, the ThinkPad's battery life has earned a powerhouse of a reputation. But as all travelers know, extra batteries can still be your best friend. With ThinkPad's battery and power options, you can ensure that your notebook is always ready when you are.
It all depends on what is meant by "reputation", I guess.

The battery lifetime on my ThinkPad has been getting progressively worse. I'm nearly certain that this is a software problem rather than a physical problem with the batteries. In this section, I'll write up my experiences, with the hope that it will help to resolve the problems once and for all.

The main symptom is that, while discharging, the APM battery indicator jumps immediately from some value (currently around 65%) to 5%. At that point, discharge continues at the normal rate to 0%, meaning that I have about 10 minutes max. The situation when charging is strangely symmetrical - it jumps from around 35% to a high value (I'll do some more testing to see if it's 95% or 100%).

James Corey reported a similar experience on the Linux-TP600 mailing list:

2) My battery life went down drastically. It used to be a couple of hours, now it's 15 minutes; it still reports a couple of hours remaining, until it gets down to about 85%, and then suddenly starts reporting 5% and 5 minutes remaining. This also happens in Win95. I know laptop batteries eventually wear out, but this was a very sudden drop, not a gradual one. Also, I expected better life given that I'm usually plugged in and not using the battery...

I suspect that the "sudden drop threshold" is quite strongly related to patterns of battery charging and AC operation. I also keep the computer plugged in most of the time. When I go out on the road, I usually go through both of my batteries (takes less than an hour), then get back home and recharge both, usually with the computer on.

It was suggested that a BIOS update, marked as "optimizing battery recharging", might fix the problem. So I got this update and tried it, but it didn't work. D. Sen's experience was similar:

Just recounting my experience. A couple of months ago, I got a new battery from IBM. I have also upgraded to their latest BIOS. All the latest BIOS seems to do is *not* charge the battery when its around 90% charged. The idea I guess is to charge it when the battery is fully drained as is recommended by IBM.

This however doesnt solve the problem of the battery going from 75% charge to 5% charge in a matter of seconds. I have gone through the 3 drain/recharge cycles....and it hasnt helped...

Alan Jaffray had this important information if you want to buy a new battery:

I bought a new battery from IBM, and my laptop just took a little over two hours to run down from 100% to 5%, while playing MP3s and CDs the entire time. All is well again. I'm also trying a "battery reconditioning" service offered at the local university bookstore, which costs $35 if the treatment is successful. I'll report results later on.

Important note: If you have to buy a new battery, be sure to get it as a "part" rather than as an "accessory"! They have two separate fulfillment centers, one typically used for retail-boxed upgrades, the other for replacement and spare parts; not that there's too much difference in the case of a battery!

The first time I called they quoted me $225 for a replacement Li-Ion battery; I was taken aback, and decided to reconsider my decision, but the second time they gave me phone numbers for both centers, and I tried the other one and was quoted a little over $125. I called the tech people again to make sure this was really the same part, and it was. Anyway, ask them for an "FRU number".

From IBM's Tips for getting the most out of your battery page comes this info:
For 560, 600, 760, 765 and 770

The battery pack for the ThinkPad 560, 600, 760, 765 and 770 series is an intelligent battery. It contains a microprocessor to monitor its capacity. Information of remaining battery capacity is passed to the system from the battery pack, and the system indicates the capacity in 1% steps from 0% to 100% with high accuracy. If the indicator appears to be incorrect, it is recommended that you cycle the battery (fully charge, then fully discharge) three times. This may occur when the battery is over charged or discharged.

Also in the same document comes this information:

NOTE : If you mainly use your ThinkPad as a desktop replacement, your battery will have a constant charge of 100%. This will negatively affect the performance of your battery. It would be better to use battery power or to remove the battery from your system while using mains power.

This information seems consistent with D. Sen's speculation that the BIOS change simply disables charging when the battery is already fully charged (or nearly so). I've run my machine hooked up to AC power most of the time I've had it, and only upgraded the BIOS a few months ago.

So, where do we go from here? First, since the problem does seem to be highly dependent on usage patterns, I'd be perfectly happy to find a workaround that decreased the "sudden drop" threshold. I'll be playing with this myself, now that I understand some of the issues better, but any tips from the field would be greatly appreciated.

Second, it seems like IBM has a real problem with the battery. At the very least, I'd like to see a real explanation from them, rather than having to infer stuff from their tech support documents. It would be even better if they were to actually fix the problem. I'll be contacting them, but if you're seeing similar problems, I recommend that you do too. I'll certainly post any responses here and on the TP600 list when I see them.

The first place I'll try is IBM's support discussion forum, as it seems that support people actually read the page.

Freezing: a solution?

mattw548 writes:

I recently looked at your article concerning the thinkpad 600 battery. I heard a helpful hint from one of the tech guys. run the battery completely dead, then remove it from the laptop and place it in a ziplock bag. Place the battery in the freezer over night (3 hours usualy does it.) Remove the battery from the freezer and place it back in the laptop. Don't be alarmed when the charge light doesn't come on. Place the laptop on a piece of cloth to soak up the condensation. The charge light usualy comes on in 15 to 20 minutes. Also don't be alarmed if it just comes on and blinks. When it starts blinking it will be about 5 minutes untill the light goes solid and actualy charges. I was told that by freezing the battery it causes the battery to take a slower more thorough charge. My battery also ran down to 50 percent then droped to 5 in less than a minute. By freezing the battery for 3 or 4 times it extended my life by about 20 MINUTES! it did run steady from 100% to about 30% then drops to 5%. that's not great but it was sure a lot better. Major note make sure you always fully charge and discharge before use. I recently over-discharged my battery and completely ruined it. Please don't make my mistake.

This sounds interesting, but I'm not really eager to try it. As I say, I'm pretty sure that it's a software problem rather than a physical problem with the battery itself.

Some details about the internals

Julius Rahmandar writes:

I took apart one of my battery pack and attempt to charge the batteries directly without the circuitry in the loop, alas the battery itself is whats gone wrong. it will not take any charge. I think I know the reason why it drops from 5o% to 0%. You see there are 6 batteries in it (each has a nominal 3.6v), HOWEVER, they are strap in parallel two at a time, to provide the 3.2A current capability, so i think for those that still works and then loose capacity after it gets to 50%, my guess is hat one of the pair is a goner... hence it goes down to no capacity when it reaches 50%. I'll send you photos of the innards if youre interested.

I've requested the photos and will post them if sent.

It's a software problem

Dave Lee writes:

One thing I have discovered is that it is not the actual battery cells that have a problem - if you discharge the battery into a load in isolation, the battery has considerable capacity left even thought the Thinkpad goes into sleep mode. Furthermore, as the terminal voltage drops to 9v (the minimum recommended for a 3-cell Li-ion battery) the output of the battery is automatically disconnected by the internal battery monitoring circuits, and is reconnected when the terminal voltage rises above 10v. This means that the basic battery voltage monitoring is working correctly - it is only the 'smart' monitoring that is at fault (where the battery calculates remaining capacity and reports this to the thinkpad).

Battery solution

Javier Valero writes:

(First sorry for my english ;) )

I have making some experiments with my thinkpad.(I have the same problem of the jumping bat), and I have reach the conclusion that the "intelligent" battery is really stupid. The problem is software, but it is not in the firmware or something in the thinkpad, but it is in the program of the microcontroler in the battery.

The solution for the problem is easy: Recycle the battery, but...not as IBM says. The microcontroler does not let you discharge all the energy from the bat, so I dont let the microcontroler of the bat comunicate to the thinkpad, and by this way I can discharge all the bat, and then get full charge.

The way to eliminate comunication is very easy: there are 4 pins in the battery, the 2 in the middle are the microcontroler comunication to the thinkpad, so before put the bat in its hole, I put 2 little papers over that pins, and I get power on. Now you dont know when the bat is empty, and the computer will power down in any moment, but you can discharge completely the bat.

You need to do this some times to recover the capacity lost.(at least some of it) A warning: when you charge it, dont use the papers, or you could overcharge the bat, and it is the worse thing you can do to a battery. If you dont understand the process (I know my englis is very poor) I could send you a couple of pics.

I'm trying this with one of the batteries, and it does seem to work. It also sounds reasonably safe. Please try this approach and let me know either success or failure.

Other resources

Last update: 22 Oct 2002

Linux on IBM ThinkPad 600

* www.levien.com