Hi all, I had a problem with a Linux OS potentially physically destroying a new laptop battery. It didn't attract any answers in the hardware section, so I thought perhaps this Linux forum might find some experienced eyes who can help. See my post describing the issue. Thanks!

Milos

If you boot up the Live CD again, does the battery start charging? Have you tried resetting CMOS? (Disconnecting and reconnecting the CMOS battery after 5 minutes or so)

1) If I boot from LiveCD again, it doesn't start charging.

2) If I boot up Windows, it doesn't charge, and going to the IBM Battery Information (on the thinkpad battery manager icon in the taskbar), shows 10.72 V or something like that and 0.00A, with 0% charge. Windows charges and discharges fine my second battery (which I'm now using), although even the second battery gets smoothly down to "60%", and then all of a sudden drops instantaneously from 60% to 4% and gives me warnings to shut down or plug in. So that battery I guess is miscalibrated now or something (but that battery is on cycle 158 - quite old). On the other hand the above "fried" battery was new - maybe on cycle 30 or 40..45..

3) I haven't tried CMOS.. 1) how would that help, 2) is it safe to do -- will it create any problems for the rest of my PC (except losing the clock time), like losing hard disk info, etc..?

Thanks!

Milos

It shouldn't hurt anything, and may help. It's worth a shot.

Ok, I reset the CMOS (disconnected, reconnected CMOS battery - and on next reboot it told me the CMOS checksum failed, so indeed it got cleared, and that it took on default values). I just set the boot order correctly and tried again.

1) Still, the "fried" battery neither charges in Windows nor in Ubuntu. Furthermore, in Windows the Thinkpad Battery manager or whatever shows "Charging" with 0 Wh energy out of 47.6Wh full capacity, 11.23V battery voltage and 0.00A current. I.e. it's trying to charge but zero current is flowing. It's still amazing to me that the (Ubuntu) system could, by discharging the battery, fry it. BTW, the battery manager in Win tells me that it's actually on cycle 76. Anyway, it's pretty new.. and worked perfectly, charging up to pretty much the full 47Wh consistently, before the fateful time I tried the Ubuntu LiveCD without wall power.

2) Going off this battery topic to another issue.. about my CMOS resetting. Weird, after I reset the CMOS, I would actually get stuck at the F1/F12 screen (go to BIOS or choose boot options). I could go to F1 (BIOS), but F12 would freeze the machine, and not pressing anything would also not start the hard disk boot (even though it is first on the boot list, it worked fine before, etc). I was really getting worried that I would never be able to start the machine again from my HD, since I could only get into BIOS, but not boot (from HD or CD). But then I forced the system to do a full memory test and full diagnostics on boot rather than quick boot, and it then booted to HD (Windows) or CD (Ubuntu LiveCD). Then, setting the settings back to normal, it keeps working as it did originally, so everything's fine again. Somehow it got its settings happy again.

I have in general had the experience that my A31 is very flaky, though I love it, whereas others here have reported them to be very robust. From the kind of problems (sometimes not waking up from sleep though rarely, taking forever to realize you're trying to wake it - even needing to press keyboard keys to wake it from suspend on occasion, etc)... I suspect the BIOS. I have the latest BIOS/embedded controller firmware, but I wonder if I should stick with the old versions, i.e. downgrade back to the version that was originally on it. I'm concerned that the new BIOS may be for other things, and that the backward compatibility to my A31 may not be the priority right now since they're so old..


Anyway, about the battery issue.. I still have no clue what's going on. Is there a battery charge cct on board? Is it possible to reset it? Is there a pinout online for what the 4-6 contacts of the battery do, so I can see what's wrong with it with a multimeter, etc.?

Thanks!

Milos

Not aware of anything like that, but why don't you take volt and ohm readings and compare them to your other battery?

Milos:

This is purely a guess on my part, but did you ever consider that you had a defective battery that just happened to fail on you while running Ubuntu?

I'm only speculating that Ubuntu may have had nothing to do with killing your battery; that it was going to fail anyway and was only a coincidence that it did so while running Ubuntu. It could have just as easily failed while you were running Windows.

Possible?

No, I think that's unlikely, I think there is a cause-effect relationship. Of course anything's possible in theory, but one needs to consider reasonable probabilities.

The evidence is:
1) Before trying Ubuntu, the battery that failed was in very good condition, and recharged pretty much to full capacity still (before a battery fails it tends to degrade in capacity gradually for well over a year -- as I have seen in my other two batteries, one of which I still use). Windows ran it for 60-70 cycles (not too many cycles), with no perceived degradation and no signs of wear. The Thinkpad battery manager for Windows is very good btw, compared to standard all-purpose battery managers and gets the capacity right, while the all-purpose aren't accurate in general (for various laptops), as I see from reading threads.

2) The battery did not fail at a random point in time. It failed exactly at the point when the battery discharged to a fully discharged state - because Ubuntu didn't realize it was drained, I guess. Windows never allows this to happen - it shuts down at about 3-4% left. This first try of the Ubuntu LiveCD came up on my A31 with barely half the hardware autodetected (sound half-worked, dimming display half-worked, thinklight worked, display off didn't work, suspend didn't work, hibernate didn't work, video ATI Mobility Radeon 7500 worked only at a lower 1024x768, but scrambled at the native 1400x1050 res). So it would not be surprising to me that the power management driver didn't get something right.

3) I can see the battery has an open-cct voltage (11.23V), but the laptop can't charge it (0.00A charging current) after this full discharge. This says to me that the charging cct inside the battery is confused by its communication with the laptop, rather than the battery cells being busted. This could be due to some battery controller IC (maybe the embedded controller?) messing up.. I'm not sure how the OS can affect it, but perhaps by messing with the EC registers.. no idea. Or, batteries are maybe not supposed to be discharged all the way.. Anyway, I get the feeling this could be revived if one knew what was wrong.

But I guess no one's seen this before, since I'm not getting answers or confirmations from anyone else who has seen this.

Anyway, don't understand me as hacking at Ubuntu, I'm not, I just want to know what happened and why.. so I can fix it or be aware in the future.

Thanks,

Milos

Your battery is not dead. You need to use another, less intelligent charger to resurrect it. Just loading it with, say 50 mA, for a couple of hours will give it an initial load and your laptop will recognize it again as a healthy battery. If you do not have necessary skills/equipment ask some friend who has. I'm not familiar with Ubuntu, the reason of your trouble is of course misconfigured (or not configured) power management.

I ran across this on the OpenSuSE wiki ... says there is a problem on Thinkpads with the ATI driver draining batteries . So maybe this is a combination of a battery that was drained and a Thinkpad that doesn't like what it sees electronically in order to recharge it. An anomaly but possible. I agree dummkopf, there should be a way to partially recharge it outside of the machine.

http://en.opensuse.org/SDB:Reducing_Thi ... ACPI_sleep

I don't kmow about the A31 batteries but the 600 series batteries had a failsafe circuit that opened up when the battery discharged to a certain level. At that point you could not charge/discharge the battery anymore but if you applied a charge directly to the cells and got them back up above the cutoff voltage it started working again. It was dangerous and I am not recommending that you do this.
Running healthy battery cells clear down can damage them. Only cells with nothing to lose should be run down to try and rejuvenate them.

Everyone, thanks for your replies.

I tried manual charge. When I try to charge the battery (I use a current source but limit the voltage to no more than 12-13V), the voltage immediately maxes out and no current flows, i.e. there's a massive internal resistance (or the charge circuit inside is blocking charging). When I disconnect the charging power supply from the battery after this, the voltage stays at about 11.1V. If I try connecting the supply at a higher voltage a few times, nothing happens (no current flows), and the battery returns to 11.1V after disconnect.

On the other hand, when I try to discharge, the voltage of the battery drops to 0.2V pretty much right away, and again no current flows.

This still looks to me like a messed up charge circuit than a bad cell, because a bad cell I would think can only be shorted or opened. There are 9 of them, and usually only one dies first (they are rarely so well balanced that they all conk at the same time), so I don't see how they could drop from 11.2V to 0.2V with no energy added or extracted.. seems more like a capacitance at the front storing it and the battery always being disconnected.. thoughts?

Milos

Nope, a completely expired cell behaves like a capacitor. It charges/discharges fast.
I'd never try charging these batteries without limiting the current.

So actually you charged that battery a little bit - can't do it without current! How did you measure the current?
I'd still say use low 30-50 mA current at least for an hr before deciding the battery is gone bad.

If it's a new battery isn't there a warranty? If not, I think I would find the cheapest battery I could find (or borrow) to make sure the machine works. If it does, buy another good battery. I don't see any other options for you.

Yes, but to go from 11.2V for a 9-cell-series battery to 0.2V, it means that *all* 9 cells need to charge/discharge fast (near instantaneously in my observations) at the same time. I find it hard to believe that they would all be so well matched after 60-70 cycles that they would all empty at the same time. On the other hand if it was just one cell that is "behaving like a capacitor", you would expect the open cct voltage to vary between 11.2V and 10.0V (i.e. around 1.2V) between charging and discharging.

Actually, I tried watching it.. when "charging", with about 1mA current the voltage goes from 0.2V gradually to 13V over about 10sec. Raising the current will try to put the voltage way above 13V (high internal resistance).




I am in fact controlling the charge current by hand and limiting max voltage - very safe. The current never goes (eyeballing the scale) much above 1mA for sure. I.e. if I try to crank up the charge current, the voltage instantaneously increases to the limiting voltage of 12V or so that I set, so the supply doesn't allow the current to increase above 1mA or so level because the voltage would go through the roof.



Well, I might have put in a few electrons , but I wouldn't call it charging the battery a little. It's small enough that it's not clear whether it made it into a cell or just onto some capacitive on the electrode.

30-50mA definitely won't go as I mentioned - the voltage goes up above 12V (tried up to 17V actually, briefly) with no increase in current - still 1mA or less (probably near zero).

Actually, I just tried to see the actual currents.. when discharging you see a flash of about 1mA, but the voltage pretty much drops instantaneously from 11V to 0V. Disconnecting it sits at 0.2V, and as I mentioned charging it can go back up to 11-12V in 10sec or so, with 1mA. It rests at 11.2V. Keeping the voltage at 12, 13, 14, 15V limits the charging current to 1-2mA, but it is actually non-zero. Even going to 20V keeps it at 2-3mA, but I was worried about leaving it there in case the cells aren't balanced and most of the voltage is dropped across one bad cell.

Further thoughts?

Milos

Harryc, thanks. I will, and I already have another (older not so great) battery, but I'm just intrigued at what could be wrong with this battery because I was pretty sure it was in good health. I guess draining it all the way might have damaged it.. or Li-Ions have to be charged up reeeeally slowly (low current) if they're bottomed out all the way..

Milos

I thought these cells are connected in serial - then the weakest cell determines the parameters of whole battery.

I admit I do not know what's going on inside of your battery.

Well I ripped apart the battery, and thought I'd let you know what happened.

It seems my battery's smart circuitry may have a low-voltage trip point and may disconnect the cells from the terminals if you ever go below that voltage (permanently).

Says so here: http://www.batteryuniversity.com/partone-12.htm:


... and it's consistent with what I see.

With no expectation of it working, I ripped apart the battery to see what's going on. Indeed, from the outside I cannot charge it (max charge current ~1mA to avoid putting the voltage through the roof). However, directly to the cells there's no problem charging them at 0.4A or so. No heating, gassing, etc; all cells balanced in voltage, etc. I expected this since the cells were ok.

So, there's some chip on the little control board which permanently disconnected my cells from the terminal. FYI, there's a PIC16F73 microcontroller which is the smarts, and a Texas Instruments bq2060SS battery charge monitor chip, and 3-4 small chips one of which is surely the culprit.

Milos

You used the word "permanently". Does it mean your battery is paperweight even after you managed to get some charge in it? Sorry to hear it if it's so.

What I mean is that the laptop won't charge it, even after I put some charge in the cells with no problems. So, it is not taking a cue from the voltage - but rather seems locked into a "disabled" state. The website I quoted above seems to suggest that this is expected -- that there's a mechanism that disables the battery if cells go below a certain voltage, and that the battery after that may be "unservicable".

It seems to me possible for this to happen because there are FET switch chips on the little control board inside the battery that can control whether the battery is connected to the outside, and there's also an EEPROM memory chip and a microcontroller (and also a charge monitor chip). I have no idea what they do, but I guess it is possible to store a permanent change into the EEPROM so the battery circuit always keeps it off from then on.. (even if the circuit loses power and comes back up).

It would seem you could reset this if you knew where to look, but of course I don't. I'd bet there's even a way to reset the battery from the outside, because the 5 pins, I found, are V+, Clock, Data Pin, Thermistor, Ground. So there's a serial communication pin between the battery and the laptop..

Anyway, if someone knows they can post, until then I give up further attempts. Basically, the battery seems fine, but a bit in the EEPROM is telling it to remain shut off, and I don't know how to reset it.

Milos

How far do you have it charged up? On the 600's once you charged them up to a certain point the circuit reconnected.
Be careful to not explode those things.

Interesting. The voltage on this battery is actually at the voltage that a normal cell is at, 3.73V - nominal for a Li-Ion cell is 3.8V. That is, it's not really low, it rebounded after being drained.. even before any charging attempt.

So, if the circuit reconnects once the voltage is high enough, I would think it would already have done that. On the other hand, if it reconnects after the charge capacity is high enough, then I should try to charge more..

I'll check the Li-Ion charging specs online (voltage limits, etc) and then try to charge a good part of the way. I only charged for about 10-20min at 0.4A, so I didn't make it far at all.. the battery has 4400mAh capacity, and I only put in about 50-100mAh.

Milos

Do you have another battery laying around that you could 'borrow' the charge control circuitry from and transplant it into the 'dead' one?

You are thinking exactly what I was thinking. I do have a battery I am currently using, which is both older and in worse shape than the one that died. So, I was thinking, fine when this one dies or becomes near useless in capacity, I'll see about transplanting its circuitry into the 'dead' one.

My only concern: the circuitry keeps track of the capacity, etc., history, so I donno if changing the cells on it would mess things up or not. There's a chance though that, so long as it is not "tripped off" as my 'dead' battery seems to be, then doing a full discharge/charge might force the battery charge tracking chip to adjust its levels and give you the correct charge level (on your desktop)..

Thanks for the idea.. if I do it (much later), I'll post whether it works.

Milos