how to revive a destroyed li-ion battery due to overdischarg

[bsmile]

I seemed to have standby my laptop too long that the battery was drained fully and destroyed -- it cannot be charged, and the laptop immediately loses power when the power cord is unplugged. The battery seems to be destroyed due to excessive discharge, but not at the end of its lifetime, for it can support the laptop 3 hours before this accident. I googled a bit on the internet. The information I got is that for li-ion battery, it will be dead forever if it is dead. However, it is also mentioned that there is built-in circuit in the battery to avoid it of being destroyed(different from dead?!). This arouses my least hope to get the battery back to life. Could anybody tell me whether this is doneable, or what I should do to bring the battery back to life?

[andykrej]

Have you tried to charge it for some hours? Sometimes it needs a longer time to begin charging. Is the battery-LED on?

[loyukfai]

If that doesn't help.

Heard that freezing, or even giving the cells a high voltage zap may help. Google "revive dead li-ion battery" for more information.

Regardless, lithium is a volatile chemical, so BECAREFUL!

Cheers.

[dstrauss]

DO NOT MESS WITH YOUR BATTERY - including freezing, heart attack voltage jolts, etc. There is a conditioning utility in the ThinkVantage tools, but don't know if it works with Win 7. If you can't revive it with a deep charging cycle, it's time to replace. Do not attempt any self-repairs. These Li-ion cells are unstable, dangerous, and not to be messed with by the average user.

[bsmile]

It had been charged for a full night, but doesn't work. The battery-LED light becomes orange and constantly keeps flicking, showing battery needs charging

Have you tried to charge it for some hours? Sometimes it needs a longer time to begin charging. Is the battery-LED on?

[bsmile]

DO NOT MESS WITH YOUR BATTERY - including freezing, heart attack voltage jolts, etc. There is a conditioning utility in the ThinkVantage tools, but don't know if it works with Win 7. If you can't revive it with a deep charging cycle, it's time to replace. Do not attempt any self-repairs. These Li-ion cells are unstable, dangerous, and not to be messed with by the average user.

Thanks for your warning! This is why I dare not playing with it. However, with the protection circuit, the over-discharge should not hurt it to death, there should be a way to treat it. Since now it cannot even be charged, the deep charging cycle cannot be completed.

[bsmile]

This is the information from the power management for the battery:

A battery error has occurred. The battery cannot be charged. Replace the battery.

any idea how to deal with this? Is the error a software one or hardware one? I really hope it an error on software level ...

[carbon_unit]

The battery is bad. Replace it. Batteries come with a 1 year warranty. Is it possible your is still under warranty?

[Ken Fox]

Your battery could have just "died," and the events preceding it could be simple coincidences. There are basically two ways that these batteries "die." One way is for the Lithium cells to degrade, which can happen gradually (the most common) or occasionally, all of a sudden. The other potential cause of battery "death" is that the electronics get fried for some reason. This has happened to me a couple of times and by your description and with the error message you are getting, I would bet that this is the cause in your case.

Either way, the battery is gone.

You could consider sending it out for repair to a service like batteryrefill.com. I have read glowing reports of this company but in my own experience the results were less then stellar. If you had a very old machine and few options for buying a "real" new battery, then I'd consider using a repair service, but since this is a relatively new machine with no shortage of available replacement batteries, I'd just hunt around to find a good price on a replacement battery and then buy it. Beware, however, that if you buy closeout batteries, either from the Lenovo Outlet or from ebay or another store specializing in closeout merchandise (like Overstock.com) you might get sold "new-old stock." This would be a battery that has sat around for a long time after it was made and then sold to you. A new old stock battery will almost certainly have a shorter useful lifespan than will a new battery sold shortly after it was manufactured. If the "new old stock" battery is sufficiently cheaper, I'd consider buying it, but check it right when you receive it to be sure it functions well.

ken

[bugmn]

Hi all,

the exact same thing happened to my 13 months old 9-cell lenovo battery pack (FRU 42T4644):

The battery LED would just blink at a faster rate than normal and the lenovo tool
reports a damaged battery. Measuring between the two outermost connector pins
gives 0 Volts reading. Just before the battery would last four hours when doing normal
office work.

As I'm a computer science guy and know how batteries work I opened up the battery
case and measured the cells directly: all of them are about 4.15 Volts totaling 12.52 Volts.

The explanation why the battery is "dead" lies not in in the cells but in the controller circuit
composed of a battery protection analog front end chip named bq29330 and battery fuel gas
gauge chip bq8030DBT. Either of these chips have a FAULTY bit that can be set by arbitrary
conditions. I interfaced the battery pack via i2c and could read some data at adress 0x0b:

root@raw:/home/master# i2cdump -y 0 0x0b
No size specified (using byte-data access)
0 1 2 3 4 5 6 7 8 9 a b c d e f 0123456789abcdef
00: 03 4a 0a 00 00 ff ff ff a4 e0 00 00 00 00 00 00 ?J?.....??......
10: 30 00 00 ff c8 38 d7 24 e8 30 31 63 78 XX XX XX 0...?8?$?01cxXXX
20: 08 0b 04 0e XX XX XX XX XX XX XX XX XX XX XX 0b ????XXXXXXXXXXX?
30: 0a 10 10 18 XX XX XX XX XX XX XX ab 13 00 01 89 ????XXXXXXX??.??
40: 17 XX 03 20 XX XX XX XX XX XX XX XX XX XX XX XX ?X? XXXXXXXXXXXX
50: 03 0a 08 07 0d XX XX XX XX XX XX XX XX XX XX XX ?????XXXXXXXXXXX
60: XX XX XX XX XX XX XX XX XX XX XX XX XX XX XX XX XXXXXXXXXXXXXXXX
70: XX 17 00 XX XX XX XX XX XX XX XX XX XX XX XX XX X?.XXXXXXXXXXXXX
80: XX XX XX XX XX XX XX XX XX XX XX XX XX XX XX XX XXXXXXXXXXXXXXXX
90: XX XX XX XX XX XX XX XX XX XX XX XX XX XX XX XX XXXXXXXXXXXXXXXX
a0: XX XX XX XX XX XX XX XX XX XX XX XX XX XX XX XX XXXXXXXXXXXXXXXX
b0: XX XX XX XX XX XX XX XX XX XX XX XX XX XX XX XX XXXXXXXXXXXXXXXX
c0: XX XX XX XX XX XX XX XX XX XX XX XX XX XX XX XX XXXXXXXXXXXXXXXX
d0: XX XX XX XX XX XX XX XX XX XX XX XX XX XX XX XX XXXXXXXXXXXXXXXX
e0: XX XX XX XX XX XX XX XX XX XX XX XX XX XX XX XX XXXXXXXXXXXXXXXX
f0: XX XX XX XX XX XX XX XX XX XX XX XX XX XX XX XX XXXXXXXXXXXXXXXX

The pinout of lenovo batteries is always the same: || ||||| being: + + [gap] SCL SDA [unknown] GND GND

An i2c interface is easiliy build with a spare vga cable: Pin 5 is GND Pin 12 is SDA and Pin 15 is SCL.
Under ubuntu you just need to install i2c-tools and lm-sensors then modprobe i2c-dev and run i2cdetect -l

With my R61 notebook (intel graphics) this gives me
i2c-0 i2c intel drm CRTDDC_A I2C adapter
i2c-1 i2c intel drm LVDSDDC_C I2C adapter
i2c-2 i2c intel drm SDVOCTRL_ I2C adapter
i2c-3 i2c intel drm SDVOB DDC I2C adapter
i2c-4 i2c intel drm SDVOB/VGA I2C adapter

Obviously we want to use i2c-0 as this is the interface we can then use via the vga cable.
On other machines it's often not clear what i2c device is the correct one, the trick here is to
iterate over all i2c devices via i2cdetect -y 0..n once with the battery connected to the vga
cable and once without. The diff of both runs reveals the correct i2c interface.

So what we should find out is what i2cset commands we need to run to be able to unset the
faulty bit! I hope lenovo reads this and gives us the info, I assume that some authentification
codes must first be communicated to the battery to accept the reset. Please lenovo show mercy!

Best Regards,

John

[carrausse]

Hi John,

I have a similar problem where my X61 goes in standby without notice, although the battery is charged and the AC is connected. I suspect the "chip" to be sending wrong info to the BIOS...

What tools did you use to read the chip? Software or hardware? I would be interested to deactivated the chip so that the laptop always believe the power is good even if it's not really.

My laptop is always connected to AC and I just need the battery to server as a temporary UPS when there is a power cut.

Any info will be very much appreciated.

Cheers,

Alex

[shurane]

Wow, this is good information John/bugmn. Does anyone else know about how to reset the faulty bit? Heh, querying the battery packs with i2c is ingenious; that seems like more of an electrical engineering background than computer science, but I digress.

[DeeJayPee]

Was this ever resolved? We have quite a few new batteries for 60-61s 200s etc.. that were purchased new but never used that I would like to revive. I am not clear about the custom interface. Is the battery inserted in its normal place and the battery bios codes read using the signal found on the VGA cable or is the cable clipped to the battery contacts and plugged into the vga port where they are scanned?

Thanks

DeeJayPee

[jaspen-meyer]

"once with the battery connected to the vga cable and once without."
Sounds conclusive; battery connected to custom-vga cable, other end of custom-vga cable is inserted into vga port of R61.

The commands he suggests for setting up Ubuntu are:

Code: Select all

sudo apt-get install i2c-tools lm-sensors
sudo modprobe i2c-dev
sudo i2cdetect -l # with battery<->custom vga DISCONNECTED
sudo i2cdetect -l # with battery<->custom vga CONNECTED
sudo i2cdump -y 0 0x0b

sudo i2cset ...

A snippet from 'man i2cdump'

NAME
i2cdump - examine I2C registers

SYNOPSIS
i2cdump [-f] [-r first-last] [-y] i2cbus address [mode [bank
[bankreg]]]
i2cdump -V

DESCRIPTION
i2cdump is a small helper program to examine registers visible
through the I2C bus.

[xiphmont]

As I'm a computer science guy and know how batteries work I opened up the battery
case and measured the cells directly: all of them are about 4.15 Volts totaling 12.52 Volts.

How 'about' is about? A battery well out of balance will trigger the failsafes.

The explanation why the battery is "dead" lies not in in the cells but in the controller circuit
composed of a battery protection analog front end chip named bq29330 and battery fuel gas
gauge chip bq8030DBT.

IIRC, only Sanyo-made batteries use this controller. There are five or six different controllers in the wild, and the programming and status bits are not the same. Even the Sanyo controllers do not all use the same firmware. Blindly writing new bits to a memory location is a bad idea.

Either of these chips have a FAULTY bit that can be set by arbitrary
conditions. I interfaced the battery pack via i2c and could read some data at adress 0x0b:

The bit is always set for a reason. The batteries are very very paranoid about even momentary out-of-spec condition and will throw the failsafes if anything looks wrong, even for an instant. One thing the controllers will look for is rapid unexplained voltage changes under charging/load, which can indicate damage to the battery's internal laminated separator. If the separator fails, the battery can go into thermal runaway. Likely? No. But only once will ruin your day...

So what we should find out is what i2cset commands we need to run to be able to unset the
faulty bit! I hope lenovo reads this and gives us the info, I assume that some authentification
codes must first be communicated to the battery to accept the reset. Please lenovo show mercy!

There are also two hard fuses (only one in four-cell batts) that are supposed to blow when the failure bit is set. The fuses are not generally available to the public, but you can find them on Alibaba.

[edit: Oops, I originally misunderstood 'FAULTY' as a capslock runaway. Of course, the bit's name is literally 'FAULTY'. I am a silly goose.]

[xiphmont]

Oh, and I forgot to mention... there's no real 'authentication' that I've run into persay. But all the rewrite processes are proprietary, undocumented, and different for every controller. You have to buy the devkit from the vendor and it 'just werks' without telling you in any way what it is doing. $$$, a dongle, a crappy little app that only runs in XP, an NDA and a promise not to reverse engineer.

[FryPpy]

Oh, and I forgot to mention... there's no real 'authentication' that I've run into persay.

Some more info from the net (or from the X61 forum;) but this is for X61 battry chip. Anyway protocol requires sending some "magic" bytes to get full access to battery flash and reprogram it. And for sure this "magic" bytes are different for different batteries even on the one Thinkpad series.

[xiphmont]

Oh, and I forgot to mention... there's no real 'authentication' that I've run into persay.

Some more info from the net (or from the X61 forum;) but this is for X61 battry chip. Anyway protocol requires sending some "magic" bytes to get full access to battery flash and reprogram it. And for sure this "magic" bytes are different for different batteries even on the one Thinkpad series.

Good find! I'd not seen that page. BTW, the bq8030 is only one of several chips used by batteries for the X61 and the rest of the X6x series, nor is it always running Sanyo firmware AFAICT.
There is gray market Russian software for talking to these chips for the relative bargain of juts a few hundred $$ comapred to the vendor dev kit of a several thousand.

I really need to wireshark a few of these.

[flyingfishfinger]

Some more info from the net (or from the X61 forum;) but this is for X61 battry chip. Anyway protocol requires sending some "magic" bytes to get full access to battery flash and reprogram it. And for sure this "magic" bytes are different for different batteries even on the one Thinkpad series.

I recently tried out that approach, it works for the packs I Iooked at. I reset the cycle count and capacity to see what happens; it does work. I then ordered a couple ncr18650ga to replace the old cells. Now waiting for snail mail.

R

[sheinz]

I recently tried out that approach, it works for the packs I Iooked at. I reset the cycle count and capacity to see what happens; it does work. I then ordered a couple ncr18650ga to replace the old cells. Now waiting for snail mail.

R

Hi flyingfishfinger,

Did you manage to rebuild the battery?
There might be a problem if you have Sanyo UR18650ZT 2800mAh cells. My battery had those cells and they are 4.3V rated. Controller charges them to 4.3V.
ncr18650ga on the other hand is 4.2V rated. It is not very good for them to charge to 4.3V.

[flyingfishfinger]

I did succeed, but it's probably worth starting a new thread so I can write up a proper report. Stay tuned!

R

EDIT: Thanks for the tip, I didn't check that before I recelled. I got lucky and had UR18650FMs, which charge to 4.2V *phew*.