Toshiba Satellite A205 Battery Charging Issue

[canadaboy25]

Hey guys my Toshiba Satellite A205 is having some troubles charging.  Now before I get flamed about how old and crappy this laptop is, I realize that it is 10 years old and out of date in a big way but it has a lot of sentimental value and it does everything that I need it to do.

A long time ago I shorted the battery ID pin to positive battery voltage on the battery connector which killed the laptop.  Members on this forum told me that my Winbond system controller chip was probably fried.  So I replaced the chip and sure enough, she was up and running again.  Has worked great for over a year since then.  Now, my problem is the battery is always at 0% in Windows XP.  The funny thing is that the laptop will still run for a long time on that 0% even though the orange low battery light is blinking.  If I shut the laptop off, it usually won't turn back on and the battery light will just blink.  Even stranger is that sometimes the battery will start to charge and cruise right up to 100%.  While charging though, Windows XP will sometimes say "on AC power, not charging" and then click right back to charging again.  The laptop has a led which shows that you have an AC adapter connected and another led that is orange when the battery is charging and blue when the battery is full.  This orange charging led is on even when the percentage is stuck at 0%

This was all starting to really confuse me so I started by ruling simple things out.  Instead of plugging the AC adapter in, I connected the laptop to my power supply so I could see how much current it was drawing.  With the laptop on and the battery supposedly charging it was roughly 1.5 amps.  This seemed really low for a running laptop that was charging a battery.  I pulled the battery out while the laptop was running and the current didn't even change a bit.  Aha so it clearly isn't charging the battery.  Since I had already mucked with the system controller which is connected to the battery's three data pins, I thought that would be a good place to start looking.  I tested continuity between the battery connector and the system controller on the BATT_DATA, BATT_CLK, and BAT_ID pins.  The first 2 tested out just fine but the ID pin has no continuity.  I have attached a picture of the system controller (pardon the flux).  I circled the via which is connected to pin 21 of the controller and should have continuity with pin 3 of the connector but it doesn't.  I have also attached the pages from the schematic with the battery connector and the controller on them because the full schematic is too large to attach.

Now another thing that makes me suspect a problem with the battery's connection to the controller is that I measured the voltage of the battery while the battery led was blinking and the computer wouldn't turn on and it was 12.2 volts.  The battery is a 10.8v battery so at 12.2 volts each cell is about 4.07 volts.  I don't know about you but I think that should be plenty of voltage for the laptop to start.

So basically I was thinking of running a wire from pin 3 of the battery to pin 21 of the controller because there is no connection there.  Does this seem like a good direction to go?  Or am I missing something obvious?  I'm just a little nervous to start running wires, especially straight to the battery connector.

The battery charger chip itself is obviously working because it has the mosfet that sends power to the battery shut off when the laptop is plugged in and the battery is full.  I think its just a problem with the system controller.  I may be completely wrong though because I don't know what kind of data the system controller is sending back and forth with the battery.

Any help will be greatly appreciated and if I have made something unclear or left something out, let me know.
Thanks

[Rasz]

BAT_ID is a straight connection, have you tried following this trace and looking for a break?

[canadaboy25]

BAT_ID is a straight connection, have you tried following this trace and looking for a break?

Both sides of BAT_ID just go straight to a via and inside the board so I can't tell if it is broken or not.  Yesterday I got no resistance between the two points and today I got 200 and some ohms between them.  I ran a wire straight from the connector to that pin of the chip and it made absolutly no difference.  The orange low battery light just blinks when I try to turn it on when it isn't plugged in even though the battery is at 12 volts.  When the laptop is off I can plug it in and it will draw about 1.5 amps and the battery will go up to 12.6 volts but it stops all charging the second you turn the laptop on.

[canadaboy25]

I just found something else out.  Pin 75 of the system controller is connected to the CHGEN pin of the battery charger chip.  The pin needs to be pulled low to enable the charge.  When the laptop is off and unplugged, the pin is 3.3v.  When I plug the laptop in, the pin drops to 0 volts and the current on the power supply jumps up.  When I turn the laptop on, the enable pin stays low for 8 seconds and then jumps back up to high, disabling the charger.  So for some reason the system controller stops letting the battery charge when the laptop is on.  Also, the system controller will only allow the battery to charge after shutting the laptop off if you unplug the charger and plug it back in.

[Rasz]

1 whats the link to your previous thread?

what might be bad:
charger chip
power sensing resistor divider

does it use 2 or 3 pin charger?
does windows see battery at all? batterymon or something?

[xavier60]

Has another battery been tried? Although the cells might be good, the battery's monitoring circuitry might be faulty.

[amyk]

I just found something else out.  Pin 75 of the system controller is connected to the CHGEN pin of the battery charger chip.  The pin needs to be pulled low to enable the charge.  When the laptop is off and unplugged, the pin is 3.3v.  When I plug the laptop in, the pin drops to 0 volts and the current on the power supply jumps up.  When I turn the laptop on, the enable pin stays low for 8 seconds and then jumps back up to high, disabling the charger.  So for some reason the system controller stops letting the battery charge when the laptop is on.  Also, the system controller will only allow the battery to charge after shutting the laptop off if you unplug the charger and plug it back in.

That behaviour suggests there is a detection/ID for the AC adapter and the EC prevents charging while running if it deems the adapter to be providing insufficient power. Is there an ID pin on the adapter? I found the full schematic but it doesn't seem to contain one.

I know that other laptops which do have adapter ID will run, but not charge the battery, if the ID is missing.

[Rasz]

That behaviour suggests there is a detection/ID for the AC adapter and the EC prevents charging while running if it deems the adapter to be providing insufficient power. Is there an ID pin on the adapter? I found the full schematic but it doesn't seem to contain one.

I know that other laptops which do have adapter ID will run, but not charge the battery, if the ID is missing.

We had the same idea, thats why I asked if charger has 3 pins in the connector
I have an older Asus (core duo, not even core2) with ordinary barrel charger, but detects voltage drop and wont charge if supply is significantly below 19V.

[canadaboy25]

1 whats the link to your previous thread?

what might be bad:
charger chip
power sensing resistor divider

does it use 2 or 3 pin charger?
does windows see battery at all? batterymon or something?

I have a couple of old threads because I was chasing the issue from all sorts of different directions.
https://www.eevblog.com/forum/repair/battery-connector-shorted-on-laptop/msg820112/#msg820112
https://www.eevblog.com/forum/microcontrollers/winbond-wpc8763ldg/msg873013/#msg873013
https://www.eevblog.com/forum/repair/help-finding-battery-connection-pins-on-schematic/msg835672/#msg835672

I don't think it could be the charger chip because it works just fine.  When it is enabled, it charges the battery up to 12.6 volts and cuts out.  The only time it stops working is when it gets told to stop by the EC.  So I think it's ok.  I let the laptop charge until the battery was full and the orange led switched to blue.  Windows sees the battery and says it is at 100% capacity.  If I unplug the laptop the percentage will tick down a few numbers then drop to 0% and the orange laptop battery led will start flashing.  This is ridiculous because the battery is still at 11.5 volts under load.  The charger is just a two pin charger.

Has another battery been tried? Although the cells might be good, the battery's monitoring circuitry might be faulty.

Unfortunately this is the only battery I have.  I still have the circuit board from the dead original Toshiba battery.  I  soldered some cells to it but it is totally dead.  I'm wondering if it might be the battery at fault.  The only way the system controller knows what voltage the battery is at is by communicating directly to the battery on the BATT_DATA and BATT_CLK lines.  I probed these two lines with my scope and the clock line has clock pulses on it and the data line has data going through it so they must be communicating.  I'm just wondering if the amazon battery is wildly out of calibration or if it is no longer communicating correctly with the controller.

That behaviour suggests there is a detection/ID for the AC adapter and the EC prevents charging while running if it deems the adapter to be providing insufficient power. Is there an ID pin on the adapter? I found the full schematic but it doesn't seem to contain one.

I know that other laptops which do have adapter ID will run, but not charge the battery, if the ID is missing.

We had the same idea, thats why I asked if charger has 3 pins in the connector
I have an older Asus (core duo, not even core2) with ordinary barrel charger, but detects voltage drop and wont charge if supply is significantly below 19V.

That's interesting.  It is a 2 pin charger but I'm wondering if maybe its not getting quite enough voltage?  I am using my power supply leads hooked up to a barrel connector and cable that I cut off a wall wart.  But then again the laptop was doing the same thing on the original charger...


I have the datasheet for the controller chip inside the old, dead Toshiba battery.  So using that data we should be able to decode the data on the battery data lines if it comes to that.
I just really don't feel like shelling out $150 bucks for a new Toshiba 6cell battery when I don't know if it is the problem.

[canadaboy25]

Well I tore open my Amazon battery to see what chip was inside.  It is a sn8765 which from what I can tell is a ripoff of the Texas Instruments bq20z75-v180 or similar.  The datasheet can be found here http://www.ti.com/lit/ds/symlink/bq20z75-v180.pdf

It turns out that the chip inside the battery does all of the calculations for how much capacity the battery has left and the run time until empty.  Basically every battery statistic is calculated inside the battery and sent to the EC via smbus 1.1.  So it looks like it is the battery after all?  Now that I know the protocol I should be able to see what data the battery is sending the EC and see if it is indeed giving it false information.  I also have a spare Winbond EC chip if it is the fault.

Any ideas?

[Rasz]

Windows sees the battery and says it is at 100% capacity. 

kinda not what I asked
Windows doesnt see the battery - windows talks with motherboard (EC) using ACPI
there are dedicated programs that do (batterymon), afaik it communicates with battery directly
batterymon will display all the info battery provides (capacity current max), serial number, often charge cycles, and whats most important it will show you current charge/discharge current

[canadaboy25]

Windows sees the battery and says it is at 100% capacity. 

kinda not what I asked
Windows doesnt see the battery - windows talks with motherboard (EC) using ACPI
there are dedicated programs that do (batterymon), afaik it communicates with battery directly
batterymon will display all the info battery provides (capacity current max), serial number, often charge cycles, and whats most important it will show you current charge/discharge current

I'm not arguing with you I'm just asking, but how can Windows ever talk directly to the battery?  The only data lines coming from the battery go straight into the EC, nowhere else.  So there would be no way to talk to the battery without talking through the EC?  Unless I'm missing something.  (Which is quite possible)

I don't think it matters anyways because I have found what looks like a problem with the cells inside the battery.  I did a little bit of research on the smbus protocol and found a nifty guide from TI explaining how it works.  I hooked my scope up and was soon able to see what data the EC was asking for and what it was receving from the battery.  With the laptop off and unplugged, the EC reads 12 values over and over again.  Everything looked more or less normal.  The cell voltages were 4.05v, 4.067v, and 4.064v.  The temperature was 24.65 degrees C which was fine.  A few values looked a bit strange but the one that looked outright wrong was the RelativeStateOfCharge Field.  It showed 1%.  That didn't seem right when the cells were all at 4 volts.

I then charged the battery up to full and the state of charge read 100%.  I turned the laptop on with the scope connected and waited until the orange low battery light started blinking.  As soon as it started blinking I froze the scope so I could look at the charge percent and cell voltages.  Sure enough the charge was at 2%.  I checked the cell voltages which were 3.928v and 3.815v on the first two but 3.181v on the third cell...

So it would seem as if I've been chasing a controller problem when I clearly just have really crappy cells that faceplant under load but spring back up in voltage before you can get a meter.  One bad cell would explain why the battery voltage was at 12 but it still said that it was dead.

As for why it won't charge while the laptop is running, I have no idea.

[Rasz]

I'm not arguing with you I'm just asking, but how can Windows ever talk directly to the battery?

it doesnt, it uses ACPI
maybe EC passes battery communication thru, and this is how batterymon works?

[canadaboy25]

I'm not arguing with you I'm just asking, but how can Windows ever talk directly to the battery?

it doesnt, it uses ACPI
maybe EC passes battery communication thru, and this is how batterymon works?

Yeah, that would make sense.  I put better cells into the battery and then reassembled the laptop.  I installed batterymon but it doesn't seem to show a lot of information.  Lots of fields are blank and some are not updating at all.  However, the battery has been running for over 2 hours with no problems.  The orange light hasn't started blinking and the percentage meter has been decreasing in a uniform fashion.

The only problem that I am facing now is that it refuses to charge while running.  I have improved my power cable so the laptop is not being starved of power.  It charges just fine when off but if you turn the laptop on it will run off of AC but won't charge the battery.

[Rasz]

>doesn't seem to show a lot of information

normal for chinese clone batteries I guess, but now you know computer has zero problem communicating with battery, so its either something bad in charger circuit, or software issue (newer laptops even have drm in batteries)

[canadaboy25]

>doesn't seem to show a lot of information

normal for chinese clone batteries I guess, but now you know computer has zero problem communicating with battery, so its either something bad in charger circuit, or software issue (newer laptops even have drm in batteries)

The battery doesn't even have a thermister inside it.  It's obviously a problem with the EC or an input to it because the EC is not pulling the chargers pin low.  I just don't know what.  Maybe reading the data from the EEPROM that the EC reads its code from could offer some insight

[Rasz]

you said its charging when turned off, so EC has no problem activating charging circuit (pulling CHG_EN low)
BQ24721 is also fine since it can charge and switch to battery

the problem must be with logic? EC sees something it doesnt like that tells it to disable charging
btw supplying enough power is one thing, making sure computer  sees it is another, have you checked R511
 and around?

>The battery doesn't even have a thermistor inside it.

wait what? that 'seems' like a problem, doesnt the original battery have one?


you could always sniff charger smbus and see what its told to do and what it reports back both when it is charging and with computer powered on
http://www.ti.com/lit/ds/symlink/bq24721c.pdf


btw you said it stops charging  8 seconds after power on, what if you stop in the bios?

[canadaboy25]

you said its charging when turned off, so EC has no problem activating charging circuit (pulling CHG_EN low)
BQ24721 is also fine since it can charge and switch to battery

the problem must be with logic? EC sees something it doesnt like that tells it to disable charging
btw supplying enough power is one thing, making sure computer  sees it is another, have you checked R511
 and around?

>The battery doesn't even have a thermistor inside it.

wait what? that 'seems' like a problem, doesnt the original battery have one?


you could always sniff charger smbus and see what its told to do and what it reports back both when it is charging and with computer powered on
http://www.ti.com/lit/ds/symlink/bq24721c.pdf


btw you said it stops charging  8 seconds after power on, what if you stop in the bios?

Well I'm guessing that they just put a resistor over the thermistor pins of the chip because the battery is reporting a constant temperature of 24.65 degrees to the laptop.

It seems to be working now though.  I cycled the battery from full to empty and back a couple of times and that seems to have reset something.  It now charges the battery when running and doesn't cut out anymore.

It seems that it was only the battery the whole time and I had the laptop torn apart for no reason.  If I just bought a new battery like a regular person would have it would have worked.  Oh well it's working and it was free.

Thanks for the help.

[Armadillo]

Was that from the previous post?

[james_s]

I ran into an issue like that some years ago, rebuilt a laptop battery with cells salvaged from a new surplus battery for some other laptop, it still reported the battery was low because the intelligence inside the pack kept track of that. I don't recall what we ended up doing. Some inkjet cartridges do the same thing, they will report empty long before they actually are and if you refill them you have to reset the chip or it still says empty.

[canadaboy25]

Haha yeah, but as stated, the laptop is ancient so buying a new battery was not happening.

I've had the same issue with the cells not being recognized as good cells by an IBM laptop but this Toshiba seemed to accept them just fine.