G73 Battery Pinout

I'm trying to determine the pinout of the battery so that I can talk to it. However, I'm trying to do so non-destructively, meaning no hammers and chisels (yet). If I can figure out the pinout, I've got some I2C tools which could talk to the battery, see what's there, and eventually build a larger power battery (LiPoly packs with a BQ2060, perhaps, or maybe just a simple microcontroller like the MSP430 with analog support circuitry).

I know it's not practical. Buying one or more ASUS-made batteries is cheaper and easier. It's probably safer too. But, this is more fun than watching brain-dead TV or spitting out pointless twitters. Might learn something too.

1. Does anyone have the pinout? This would be best, of course.
2. Does anyone have a dead G73 battery that I can beg/borrow/steal for dissection?
3. Are other ASUS batteries known to have a similar or the same pinout? This would mean I could get another battery (dead or not) for experimentation.
4. Who else does LiIon laptop battery hacking? Most of what I've seen is older and/or fairly crude one-time things.

At some point, it'd be nice to have a 16-cell or better battery to connect to my G73 that uses the built-in charging circuitry. The closest I've seen to a large laptop battery is the stuff at BatterySpace, and it looks like TSA's worst nightmare. Plus, I don't know how trustworthy they are.

Portable Li-Ion Battery Bank ( 14.8V 12.6Ah, output 15-24V, 180Wh) for Laptop and Mobile Electronics

 

Sounds like it'd be easier to cannibalize an existing g73 battery pack and just add more cells. IIRC, most batteries have a standard pinout, and everything I've read has been quick and dirty hacks that look like you're attaching a bomb to the computer.

 

the best thing you should as DCx said just add more cells in parrel i think this could work and you do like 12 or 14 cells battery

 

The problem with adding more cells to an existing battery pack is that the battery management chip in the pack may not work well with that. Some implementations appear to be quite simple and easy to work with. You could easily read/write settings to reflect the higher capacity so that the charge management system on the notebook would then accurately charge and display status.

Other implementations are like encrypted Rube Goldberg designs. They have a full microcontroller (undocumented and unsupported) that does not allow any easy post-factory control. You have to be a major buyer from the chip manufacturer to get any information.

In addition, some implementations appear to act like the HP ink cartridges. They register decreases in capacity, but they will not detect or accept any increases. So, if the battery reaches 30% of rated factory performance, no battery change back to 100% capacity will be seen. This means that the discharge/charge management system on the notebook will charge the battery less and allow less discharge. You have to decrypt the EEPROM, change values and then re-encrypt. I believe this is what happened to Kevin's attempt to rebuild a battery pack in 2006 (see http://forum.notebookreview.com/notebook-news-reviews/213793-diy-laptop-battery-rebuild.html here in the forums).

You say that the pinout is standard. Great! What is it?

I can guess the meaning of 5 or 6 pins:
- V+ (Vcc)
- V- (ground)
- T (thermistor possibly the same as SFT or safety enable?)
- I2C SCL (clock)
- I2C SDA (data)
- SFT (safety and possible the same pin as the thermistor)

There are 9 pins. What are the other 3 or 4? Intermediate voltages in the pack for balancing? Is the battery balancing done on the laptop motherboard? I had thought that the BMS equivalent was on the battery, but I'm now thinking that it's not. Perhaps only the overcharge/undercharge and fuel gauge circuitry is actually in the battery pack. This would make sense because it would keep replicated circuitry to a minimum (once for the laptop instead of for every battery).

I won't scruple at cracking a battery pack, but I wanted as much information as possible before I irretrievably throw away $140. If I could find out the pinout, I don't even have to do that. I can manufacture a fake battery header (just a connector attached to a cheap Spartan 3 proto-board), snoop the I2C protocol, and find out what I want.

I think the key is figuring out which pin is the SFT pin to enable the battery. Perhaps it is sensing voltage on the therm pin?

Asus has to be doing the same thing on a bunch of batteries, because it just doesn't make sense to do it radically differently for all their 14.8V battery packs and charge systems. They've got to have a common block of circuitry that they just re-use with minor modifications in the UL80Vt, G71, G72, G73, G51, G53, and so on.

I assume that Pin 1 is ground and Pin 9 is V+ from the battery package, but this is what I find when probing the laptop pins with the laptop off, battery removed and AC adapter removed.
Pins 1 & 2 are shorted or low-impedance (< 5 Ohms) path
Pins 1 & 6 are shorted or low-impedance (< 5 Ohms) path
Pins 1 & 7 are shorted or low-impedance (< 5 Ohms) path
Pins 8 & 9 are shorted or low-impedance (< 5 Ohms) path

With AC power to the laptop:
Pin 1: 0V
Pin 2: 0V
Pin 3: 3.3V
Pin 4: 3.3V
Pin 5: 3.3V
Pin 6: 0V
Pin 7: 0V
Pin 8: 0.2V
Pin 9: 0.2V

So, I assume that I2C bus (SDA and SCL) is on Pin 3, 4 or 5, with weak pullups. Pin 3 or 5 is probably probably the thermistor to keep SCL contiguous with SDA.

 

Since batteries use SMB bus, to ensure complete compatibility with I2C you need to run it below 100KHz clock rate.

The BQ2060 would be good enough for your use.

The MSP430 family really is overkill for what you would be doing. You could look at some of the PIC micros that have enough analog inputs to monitor the battery voltages and a few digital outputs for charge control.

 

ohh i am sorry man seems that you are clearly better then me in this stuff good luck then in your project :]

 

The MSP430 family is functionally overkill, but have you seen how much less current they use compared to Atmel or PIC chips? I guess the PIC line now has a "nano-Watt" line, but I'm not as familar with it. I've already done commercial low-power designs with the MSP430.

Also, I'm familiar with the clock rate of I2C as I've done extensive work with it. Thankfully, the MSP430 has a hardware block that does I2C and lets the chip sleep most of the time, only interrupting when a full symbol has been transferred or an event like start/stop. It's a lot more of a pain to do I2C on an FPGA chip. The PIC and Atmel chips do a fine job of I2C but just use more power.

Really, I'm hoping that I can just use an existing real fuel-gauge chip like the BQ3060 (the BQ2060A only supports up to 10Ah) and only use a microcontroller or FPGA for reverse-engineering. The FPGA would only be for data-logging if the microcontroller isn't fast enough. I don't have a real logic analyzer or protocol analyzer but have written special purpose code to do triggered logging+analysis before.

 

We use two variants in the MSP530 line, have been for quite a few years, If I had known then what I know now, would have chosen something else, albeit I am not sure what. The main draw for the MSP430 is the display controller built in and the low power consumption,we use it in battery powered devices.

Like the MSP430 we use a couple variants of PIC’s as well, we are using parts that are getting older, and I see a time in the near future where we will have to recode to use a more current part. Microchip has some nice ultra low power devices these days.

I have a lot of experience in charging of NiMH, which is really a whole different beast, so I am not really of much help otherwise.

 

Thanks, othonda. I was hoping that someone would know the pinout or have a dead 14.8V ASUS battery for dissection. Even a high-res photo of the G73 (or similar) mobo might do the trick, since they might have helpfully put SDA, SCL or such silk-screened onto it.

 

One of the reasons I chose a Sager over the G73 was the fact that you can get access to full schematics threw the service manual.

 

Hey, thought I'd jump in on this thread.

I have a few spare 18650 cells that I'm going to merge into a 6-cell battery on my Asus G50vt. I just got my soldering iron today and I feel like a total n00b since I'm gambling no IC changes will work fine. I'm basing this off the fact that my battery has reported everything from 36% to 12% wear in the past 14 days and the max capacity has fluctuated accordingly (one time something reported it was 106% charged). Hardest part for me will be disassembling the battery safely.

OOC, for the G73, how will that work? I thought the battery was completely underneath the laptop so adding cells was highly inconvenient.

I'm going to put my cells in sets of 3 and just make everything parallel, duct tape it all together, and pray I never have to face a TSA person.

 

I wasn't really concerned about making it look pretty. Mostly, this is a project to better understand the battery system. There's so much manufacturer-inspired ignorance that it just bothers me. It feels like the manufacturers are attempting fraud by deceit. Maybe Asus is better (more open) than their competition?

If your wear percentage has changed, you might have an intelligent battery instead of a brain-dead one that makes you buy more batteries.

Disassembling the battery safely might be safer with a plastic tool rather than metal. Try a plastic orange peeler tool. That's what I might have to use if I can't figure this out non-destructively. Or, you might try a hot-knife. That might be tremendously unsafe, though: heat+Lithium could be more exciting than you wish. Be quick and precise with short strokes?

 

So I just duct taped together 12 more cells to my 6-cell battery and it WORKS. Currently charging, and HWMonitor and Everest are both reporting a 51% wear level to start with. I'll post pictures sometime after it's done, I'm so excited. It looks completely ridiculous and it's funny every time I see it.

 

What was the charger chip in your battery? How many pins are there? Can yo decipher the pinout? Have any high-resolution pictures of the battery protection/fuel-gauge circuit board in that battery?

 

Wouldn't it be easier to first try out talking to the battery using some linux box or smt ? I did some work on SMB through the PCI ports and as I recall, reading it in linux was very easy using i2c-tools ...part of lm-sensors. I did the same on windows, but since I don't have direct access to the SMB there, I used winio.dll to access the SMBus controller on the ICH7 and ICH8 boards...and some nforce mobo. However, sticking 2 wires on the battery from a MCU is more fun

Edit: Just found my test app, it even works on my G1S's SMBus

 

I have no idea/the same as any other smart battery/no/no/no.

Here's what I did:
I never dealt with the chip as I only disassembled the top (the sticker + a thin piece of plastic). Inside my cells are arranged kind of like this:
|-=+-=+-=+|
|-=+-=+-=+|

Each | is a vertical connection. Each = is a cell. + and - indicate polarity. I added cells, and did not replace any. This is generally frowned upon and is not a good idea because of 1) differing wear levels, which infers 2) differing discharge rates when connected to other cells, and 3) I must match the voltages of all the batteries as close as possible before connecting them.

I should preface the rest of this by saying this was an extremely crude job. I had at my disposal: a leatherman tool, of which I used the knife+pliers; a soldering iron, solder, duct tape, some kind of thicker speaker cable (idk what gauge, I want to say ~18 or 20?), and a 24V DC fan (in place of a multimeter, I'll explain later).

At the cathode side there was a very convenient terminal for soldering. At the anode side I wasn't so lucky, I had to break part of the plastic to get to it, but I was able to squeeze a bare speaker wire and some solder in there to connect it. I broke off enough plastic to run my wire outside. Prior to this I tried soldering my new cells together, but I had protected cells without tabs, and the surfaces were too smooth and narrow when put together to do a respectable soldering job. I managed one battery-to-battery solder connection but ended up duct taping the rest of the cells together under tension from the duct tape. They're holding up well.

After DTing the cells together I tested the relative battery levels by listening to the speed of a 24V fan I had. I had run down my original cells to about 40% prior to testing and to my surprise I was almost spot on. I much prefer a multimeter, however. I am lucky because my cells were all relatively the same voltage (+/- 0.02V) when I measured them about a year ago (when I first got them and had a multimeter). If the cells differ by too much they'll try to charge one another upon their first connection as fast as possible, given the low resistance of the batteries. So I was very wary of any heat generated or sparks shown when I connected all the cathodes and anodes together (cathodes first!). I was lucky and there were none.

I soldered the speaker wire to the terminals and duct taped everything back up, put it in my laptop, and it immediately detected and began charging it far past its original capacity. I can only tell because Everest lets me read the charging rates, and when the cell was reporting 99% capacity, it was still charging at 26W. About two hours later it was charging at 5W and dropping, which I expect means it was nearing the end of charging. I have yet to fully cycle the battery so it still reports 50% wear, but hopefully that will change over time.

I took several pictures but none are of the PCB/SMB, since I never touched that.

 

@Predator_MF
Using the SMBus controller would be easier? Maybe. I've already got my own I2C code and know how to use it. Learning lm_sensors, building a parallel-port adapter, and re-learning linux (I grew up on SunOS and BSD, but use Windows now) would probably be more work. Not that work isn't bad. I mean, if I didn't like doing this stuff, I'd just buy more batteries.

Honestly, it's just kind of neat to have a use for this old MSP430 proto board I've had sitting around.

@dtd00d
I think I'm going to just buy and tear apart a spare battery. Maybe I can use vacuum forming to expand an existing battery and make it look semi-normal. I wish notebook companies had a std form-factor for their batteries.

 

@dtd00d
Please post your images. They should be fascinating. Or, as the online community is so fond of phrasing, "Pics or it didn't happen!"

 

Hey thought I'd give a small update (no pics yet because of some problems, sry!):

It looks like I need to modify the EEPROM or something somehow. I see what you guys were talking about with the controllers and all that. The battery doesn't detect that it's any less than 20% worn, though if I force Windows to keep going, I can sustain 0% for another few hours. I tried calibrating everything once already but the battery won't pick up the extra capacity.

It seems to be charging itself just fine and has no problem reading voltages, but capacity is measured in mWh, so after it supplies a certain amount of energy, it decides it's at 0% when it's really at about 60-75%. Right now I've been running for a little over 2 hours and the voltage is at 10.8v (11.1v nominal, 12.x volts fully charged). The cells I added are protected cells in case something goes terribly wrong.

One thing I made sure to do when adding to my battery was ensure the IC never lost power (ie: I never took out the original batteries).
Although I'm an undergrad EE, I have no experience with the SMBus, interfacing with it, what the pins are for the battery, nor why Asus needed 9 of them. The list you guys had above makes sense, a GND, V+, DATA+, DATA-, and some other stuff interspersed.

If there were some way to disable the mWh calculations on the IC, I think my problem would be solved (assuming the notebook would read the %remaining from the voltage instead of capacity).

So... Anyone have a schematic for the parallel port adapter?

 

You can build one, or you might try this:
SparkFun Electronics - LinkM - USB to I2C

It's a USB-to-I2C adapter with libraries to control it. If I didn't already have proto boards with I2C on-board, that's what I'd get. If you need the LPT based board that the unlocking program uses... here it is:
http://www.i2ctools.com/Downloads/LPTtoI2C/LPT-to-I2C_Hardware_Users_Manual.pdf

 

I just thought of a way to defeat the coulomb-counting problem though. I think the current sense resistor is external, meaning that you could change its value or use a second resistor in series to make a voltage divider and thus distort the measurement. All it is doing is getting a voltage across the shunt resistor and (with calibration) deriving the amperage. It is probably calibrated to that resistor.

See (for the BQ2060A) this:
http://focus.ti.com/lit/ds/slus500c/slus500c.pdf

Can you tell me what the battery management chip is? Are there any bits of silkscreen text on the battery's PCB?

 

If its the BQ206x series, I think the data on the EEPROM isn't actually encrypted. It's protected by a checksum. So, rewriting the parameters and the checksum might be enough. Make a backup before doing it, I guess, and then power-cycle the BQ206x to force it to re-read the config from EEPROM. The documentation isn't clear (on purpose, I'd guess).

 

I really appreciate your input and knowledge, Brians, but I need to take things a little slower as I'm new to this kind of stuff. In the meantime have some rep! But I need some more help...

Hardwarewise, can you clarify how to get from the USB adapter to the EEPROM? The USB to I2C adapter has 4 pins, and I'm not sure where to put them in terms of battery terminals. So so far it would be:

Computer >> USB to I2C >> ??? >> Battery terminals

How far off am I?

Is there no way to do this without peeking at the board itself? I am currently at school and only equipped with a leatherman and soldering iron. The plastic on the battery is very rigid and I'm afraid it'll snap in places I don't want it to if I apply enough force to break it. As the leatherman has only an unserrated knife, I'm left without a saw to break into the battery itself. I'm also afraid if I disconnect the power from the chip, the whole battery will be compromised.

This leads to my concerns about adding a resistor in series--unless I'm really crafty there's a very good chance of disconnecting that part of the board... though that resistor is probably separate from the main battery supply. Can you explain how the Rsense works? From the PDF you sent it looks like the Rsense is in the range of 10-100mOhm (not MOhm), which is quite small!

As for software I'm not too worried about it yet, but I'd like to know exactly what I'd need to do for all this stuff before I go out and spend even more money. Currently what I have kind of works, I just need to power my computer back up after the battery thinks it's at 0% and I'm good for another 2-3 hours.

 

First, my thought is that you are better off just keeping what you've got if you want to just be practical. If you are motivated by curiosity and a certain level of OCD... you're in good company.

Sorry for going so fast, but you said you were an undergrad EE. I figured you had EE experience just dripping out your ears, unlike me. I'm a CS grad, which means that they lock up the screw drivers in the cabinet if I get closer than fifty feet. Can't trust a CS student near HW, ya know.

OK, so the connection to the battery is going to be four wires: Vcc (positive voltage), Ground, I2C Data and I2C Clock. You can see the silkscreened info for which pin is which on the USB adapter on page 6 here:
https://www.sparkfun.com/datasheets/Components/LED/LinkM_datasheet.pdf

You got the connection chain perfect.

Which pins are which on the battery? I don't know, and I haven't cracked open my battery (yet). That's the easy way. You could also crack open your laptop and trace the circuitry there if it's easier. I haven't done that either. I'm chicken; risking a $130 battery is easier to contemplate than risking my $1400 laptop. You can guess, if you like, put a voltmeter on the battery pins (while it's connected) to see which pins are Vcc and ground. You could put a logic analyzer on them to see which ones are showing I2C activity. I still say cracking the battery is easiest, because I don't *have* a logic analyzer, and getting wires to the battery while it's connected seems like it'd be a pain.

The specific circuitry for measurement of the shunt isn't detailed in the BQ2060A document. It could be a straight voltage measurement, or it could be a more sophisticated arrangement like an isolated amplifier. See here:
Shunt (electrical) - Wikipedia, the free encyclopedia

*If* you went whole hog and modified the shunt component, you would have no worries about changing it, because it's not part of the power path for the gas gauge IC. You could just temporarily jumper that pin to ground while you worked on it, too. Until the battery is delivering power, Vshunt is identical to ground, right? The IC measures current by sensing the difference between ground and Vshunt, and using Ohm's Law (along with the known value of Rshunt) to calculate the current.

I = V / R

If you are being rational about it, either treat it as an EE learning exercise (cheap tuition on a semi-practical project!), or leave it alone. Frankly, I think the best approach is to buy another battery, disassemble it more completely and experiment upon it rather than on something you depend upon. I depend upon my laptop battery, so a small mistake is very costly. If I experiment on a backup battery.... all of a sudden, the cost to any mistake goes way down and I don't get ulcers and hand tremors. Maybe you might feel the same way.

 

I forgot a couple of things. You'll need to connect only three of the four I2C adapter pins: Clock, Data, and Ground. Do NOT connect V+ to the battery V+. The battery V+ is near 12V, and the adapter is +5V. You'll probably smoke the USB port on the host computer. Bad. I would not guess that the USB-I2C adapter is sophisticated enough to avoid carbonization if attached to a +12V high-current battery.

 

Well stopping is an option anytime but it'd be fun to at least look into what it'd take. Yeah I'm an undergrad EE, but we're taught more things like theory and how to make the best amplifier from a BJT or MOSFET. Not how to communicate with a schematicless battery over a two-wire interface!

Aha, great!

Aha, ok. I actually had a spare dead Dell battery which has the same 9 pin interface as my Asus. I ripped it apart with my leatherman and it actually has some stuff labelled on the pin junctions as follows (numbers going left to right with the horizontal plastic bar of the battery port on top):

1 +
2 +
3 C
4 D
5 BATT_PRE?
6 SYS_PRE?
7 BAT_ALE?
8 -
9 -

This battery may be a little more complex than mine as it has a second board to it to light up some LEDs and indicate what charge is left. It's extremely difficult to make out what chips are used in it since they either do not have any ink printed on them (you can read what's kind of engraved) or they have some plastic coating on them to obscure that. The voltage regulators and some resistors are unmarked as well, and there are two large flat capacitors.

Anyway in terms of logic analyzers I can do an oscilloscope which is a bit more low-level. I don't think I'd be able to read any data but I'd at least be able to tell what pins are operating at DC/data/clock. Maybe I'll bug my friends in IEEE if I get to that.

I don't know why but I assumed that you'd want another resistor in parallel to lower the resistance, but I see I want a lower current rating. In essence it'd have the same labelled capacity, but report a lower discharge rate. My only concern is the magnitude of the resistor--was I correct when I found the resistance is in the mOhm range? That's incredibly tiny and is basically just a thin wire. I think my limited resistor collection begins at 1Ohm.

Haha, I'm in the same boat as you. This is my original 6-cell battery I am modifying, but I bought a 9-cell awhile back. Having two is the only reason I'd think of playing with the bomb we know as Li-ion cells.

lol ok, as an undergrad EE I can at least attest to the effects of overvolting that which is not made to be overvolted and magic smoke and all that. I know the effects of this and it doesn't take long in lab or IRL to become very wary of what voltages are coming from where and when they shouldn't touch.


SO, to outline what I would need to do if I really wanted to go through with this:

• Find out what my I2C chip is
• Get the USB >> I2C adapter
• Hook up pins to their respective parts on battery. +Vdd remains disconnected and battery does not need to be plugged into laptop.
• Use software to change certain offsets and the checksum, assuming data is not otherwise encrypted.

OR

• Open up battery to get to the controller chip
• Find the shunt resistor
• Add approximately double the resistance in series (since my battery capacity is now about 3x what it was before).

 

I still wonder why you make your life hard while there's an integrated SMBus controller in almost every laptop out there and you can get it up and running with less than 500 lines of code or even less...assuming there are dozens of open source libs out there. No soldering, no risk of getting something messed up, less time. And yes, I'm a hardware guy and I love going low level, but there are times that I do prefer the easiest way. Get any linux, install lm-sensors...it will find any SMBus devices ...open up any C editor (or java or even delphi), 50 lines of code and you are there.

 

Oh. Any chance you can point me to a program that will work with the on-board SMBus controller? And do you know if that controller supports writing to an SMBus device? I tried googling a few things and I found eeprog, Linux SMBus (I2C) 24C32 EEPROM reader/writer, however at first glance at least it looks like it still needs a parallel to I2C port or something.

Edit: Seems like you're right. I assume if I follow this guide I can try writing to my battery's EEPROM (or rather, reflash the whole thing since it's EEPROM, right?).I just have to find out what controller I have and what documentation goes with it.

 

One, does that SMBus controller connect to the laptop battery SMBus or is it just connected to internal items like the fan controllers, thermistors and so on? This is possible, but I didn't think it likely. All this does is let you talk to a read-only chip because it is likely to be locked.

Two, do you really think it's connected to the I2C bus between the gas gauge chip and it's I2C EEPROM? This I think is highly unlikely. It would be needed if you want to rewrite and adjust the settings without having to guess the password on the gas gauge chip.

 

You can only rewrite the EEPROM if your gas gauge chip is 1) connected to the SMBus from the south-bridge chip-set, and 2) has and allows I2C pass-through commands. Right now, you don't even know which chip it is, do you?

 

What would be different if you connect directly to it ? You sill don't know what chip it is, you still don't know its address, you still don't have a clue what to do with it. My point is, you can just scan the SMBus for all answering addresses and see what's behind each one...with lm-sensors, the very setup does it and tells you what's behind every address (if known). And yes, I do have a working software for talking on the SMBus (with my own written drivers) and it even does work on the G1. I have ICH7 and ICH8 drivers for windows, both are working fine. I'm just pointing out what would be easier for me, I guess only you know what your goal is

 

Predator, do you think you it's possible to write to the battery on an ICH9M chipset then? I can give give you remote access to my computer if you'd like to diagnose the situation.

I found the SMbus controller in my computer, here's the copy/paste from everest:
Bus 0, Device 31, Function 3 Intel 82801IB ICH9 - SMBus Controller [A-3]

I also think we can all agree I need to find what chip my battery uses. I don't want to tear my battery apart that far though. Can this be done via software?

 

Predator, you've used lm-sensors on the G1 to talk to the battery gas gauge chip? If so, neat! I thought that the topology was different.

I thought that the south-bridge was connected to a charge controller chip on the motherboard via SMBus, and that the charge controller was connected to the battery's gas-gauge chip via another and separate I2C/SMBus bus.

Normally, ACPI is used to communicate with laptop batteries, and ACPI is an amorphous mass of functionality to me. I mean, it's got its own compiler, for Pete's sake. I think it's like all the best elements of Forth and ASN.1 tangled together for infinite flexibility and no thought for simplicity.

Anyways, I thought that there were (in Windows) no means of synchronizing between the ACPI driver and user-land code doing SMBus/I2C transfers. Unless there is synchronization, you run the risk of trampling on transfers already in progress. Even in Linux, don't you have to turn ACPI off in order to do custom SMBus activity?

Thanks, Predator, I'm learning...

 

Uhm...I never said the battery is connected to the SMBus, I just said scan it, see if it is there and if so, connect to it. I'm not familiar if it is connected or not, I can check that out on my G1 for you.... Also, an idea for replacing the USB<>I2C adapter is to use a PC-s SMBus to connect to external device (such as a battery as in your case). I've never worried about messing things up, I've already did a PCIx1 some time ago and used the SMBus directly with no trouble at all...that's when I wrote the ICH7/8 drivers (and some weird nforce mobo's one) ...and I have those working under windows and linux with not much effort, just couple of PDFs online. The only thing can go wrong is if you hang in i2c-start condition for too long, the PC just hangs up....I got that only couple of times, besides that is totally risk free....I don't have to turn ACPI off to use the SMBus, I just address the hardware directly under windows (using winio for example), under linux there was even less work - just use the lm-sensors lib to talk to any SMBus device - also risk free

 

Brian,
I was hoping you could help me out, I am interested in your original focus of this post; the pinout for the G73 battery.
After a power outage, my battery stopped providing power to my laptop. The laptop still sees the battery (charge level), but no charging or use of it remains. I did verify that it is the battery by trying a good know battery, I did not however verify that it charges, just that it will run the laptop.
I have the battery disassembled and was hoping to reverse design the board, so if you have any info on that, I would grateful.
I am thinking about modifying the battery in a few ways;
1) Fix the "NO Output" problem,
2) Install some additional circuitry and a connector to allow external charging of the battery from a large bank of deep cycle batterys (I'm wanting to minimize power loss, my current situation is deep cycles -> Inverter -> Laptop Battery Charger)
a) Any designs or ideas for the step up converter would be appreciated also.
Thanks in advance.

 

Interesting! I haven't had the spare time to follow up on this, but am fascinated. Can you post pictures of the chips and tell how you disassembled the battery? They are glued together, and I couldn't think of a non-destructive way to open them up.

 

Hi Brian,
are you still working on this project?
if so, i have some interesting info to share with you. It looks like most manufacturers are switching to the same pinout. In fact, some batteries can be switched with other brands, so long as they fit the port.
I have disassembled a dell battery, which has points in this order:
P-,P-,C,D,TS,SYS,ID,P+,P+
This directly corresponds with what you have posted.
Thus, i have arrived at the pinout being: Neg,Neg,Clock,Data,Therm,SYS,ID,Pos,Pos
(- term 1,2)(SYS/ID is GND on off? 1,6,7)(+ term 8,9)
I fyou like, i can upload a pic of the board.

 

I'm not working on it currently (time is short!) however, I'd love a picture. Please share! When I get some time, perhaps I can get back into this project.

 

Hello guys.
My G73jh battery was dead after 2 years. I've opened it, all cells having 0 voltage. They were some unprotected LG 18650, so i bought Panasonic NCR18650A to replace them (these are higher capacity and protected). I did the diy and connected the battery to the laptop. It shows that is charging but 0% charge. Voltage is reporting normally through HWINFO64 (14.48v).

I tried to charge with laptop closed, but nothing. It will not charge (though it's showing it's charging) and it will certainly not boot up from the battery alone. I checked the circuit board to find out if something is preventing it to work, and there is a Texas instruments IC (bq20z95): Battery Management - Battery Fuel Gauge - BQ20Z95 - TI.com

Reading through technical documents, i found the pins that you connect to reset it, but with no success. Maybe i just need to connect exactly these pins together without touching any other pins, but it's too small. Also i found out that the chip can use SHA-1 authorization, which will prevent me running any function if it is activated. Also there is a reset function(software) but i don't know if it has to do anything with the hardware function:/ I'm thinking of trying to charge cells manually individually but i have no charger, plus now the tabs are welded together.

I was stuck for sometime, but when i went to sleep i made this logical conclusion:
1) Laptop will not boot alone from battery since the chip won't let it as it sees 0% charge
2) Since the previous cells were dead, the chip remembers that they were on 0% charge so it's confused a lot now. Also before while the battery was dead, when you connected it to the laptop, it would try to charge and then show dead. So maybe it uses a different charging method to charge the cells now, or maybe the protection circuit on the cells will not let them charge with this method (at that current)
3) Which gets me to this point: I have to reset the IC. When the manufactures build this circuit board the batteries are not attached. So flashing of the IC is done before is soldered to the board or after it's soldered but before connected to the cells. Still something complicated maybe happening. But if it is this way, there should be some jumper to reset the chip after they connect the cells first time. There some pins on the board i suspect are for this thing. Though i have no professional experience or knowledge working with laptop battery circuit boards. Neither i have a flasher or similar equipment.

Any help is appreciated. I can post pictures too.

Kind Regards,
Ken


EDIT!! I found this site saying they have a flasher that supports this chip!:
BE2Works > Home
The software is able to reset the EEPROM!

 

Same exact thing happen to me... it feel good i'm not alone who try to upgrade g73's cell

 

Neat that someone is working on this. Please post any success or info you find! I haven't had much of any time to work on this, but it really should be possible if you download the bq20z95 datasheet, peek around with google and have some patience.

I remember thinking it should be possible to work on the EEPROM image for the bq20z95 with some judicious value rewrites and then just recalculating the checksum. It's probably a lot more complicated than that, though.

 

Hello guys. I've actually went deep in this. After much research i found out that there only a few technicians in the world able to do this. What needs to be done is to reset this "fuel gas" IC on the circuit board of the battery. When cells fail, like in my case, this texas instruments chip (which controls everything about the cells-since the original cells are not protected) writes a permanent fail flag in it's internal EEPROM.

In order to reset the chip you have to have one of the evaluation units made by texas instruments. These cost $50 both(you only need one) plus $30 shipping if you leave outside US. With this "flasher" unit and proper knowledge you can reset the chip. Since i'm no expert at this, i found someone that made software that makes this a piece of cake(at least much easier). It costs $10 for a 24-licence and it will allow to write new capacity value, so it will report correct capacity and time to the laptop.

There is a catch though. This chip MIGHT be password protected. From re-search i found out that another ASUS laptop battery with similar chipset, has password on it. So probably this will have too. In order to fix this, you have to read the chip, unsolder it, solder a new/empty one in, and flash it with the data.

I don't have the experience, though i could make this. Problem i don't have the equipment or the money to spend on this. Actually i'm going back to UK for studies at the end of September, my housemate is electronic engineer, maybe we could end-up with something. Still, there better solutions out there:
-Find someone that already does this thing and pay him to do it
-Find a used g73jh battery(that it's not dead) and try to transplant new cells there. Theory says that if you don't stop supplying voltage to the chip, it will not understand that cells are changed)
-Get a cheap chinese battery replacement (i already did) and transplant new cells there.
-Hardwire battery so it bypasses the cheap. I know it will supply power correctly, i don't know if it will charge correctly though.(Still, it will not report data to laptop like normal)

Anyone that wants to help, i will be living in Bath, UK if someone has used batteries so i can experiment onto. Alternatively you can find more details in the thread here: Battery EEPROM Works &bull; View topic - Newbie questions. Asus G73jh battery with BQ20z45

Maybe we can find someone that will do it for payment. If you want me to post pictures here, reply.

Kind Regards,
Ken

 

Thought to share these photos:




New Panasonic cells:



S**tty LG cells:



Spot welder?? 12V car battery with 2.5mm solid cooper wire. These are the worse welds. You need quick hands, gloves(first time i burnt my finger), and something to secure tabs while your trying to weld them). Still, it works with no damage to the cells.



You can see to the top left a big chip. Is texas instruments cell controller/gas gauge: BQ20z45



Rest of the circuitry:

 

This is a simple reply to state that your post is awesome. You have really dug into this and it is appreciated!

 

Great experiment there TnF. Thanks to my basic electronics subject that I can understand some stuff that you are trying to perform. My G73JH battery is still kicking as new with its 5-8% wear level after 2 years and 3 months, though I very rarely use it (I always leave 40-45% charge on it) as I'm always on AC and UPS. I hope you'd yield good results from your experiment :thumbsup:

 

The problem is not achieving a result. I'm pretty confident i can do it, though it will cost me a lot that way. Except if i open battery re-cell service in the end; Which i can't since MEng studies take lots of time.
Still i would buy a used g73jh battery that is not completely dead to experiment on it.
I'll try to come up with more solutions, at the end of September though.

 

@Brian Sorry for the late reply, but i had other things to do
While I don't actually have a G73, i have realised that ALL notebooks have pretty much the same connector (I arrived at this page looking for pinouts for laptop batteries) and i once rammed a fujitsu battery (don't ask why) into my ASUS (K43S) and it did start up. So yeah. I believe the pinout for those applies the same, really, since 1) All your previous conditions meet the labelling on board and 2)Fujitsu batteries that half fit into Asus computers also seem to work
@TnF Are you sure the batteries have any charge at all? Buy a voltmeter from any random shop and check the voltage of EACH CELL. If it reads 0V or <3V then the chip is doing its job, really. The best way to get them to likely work again is to take any adjustable AC adapter that electronics shops sell at 4.5-5V and charge them until they have at least 3.4V or so, then reconnect them. The voltage in fact sounds like the chip is definitely protecting the cells, cos Li-Ion cells for 3 in series should NEVER exceed 13V. Ever. Especially if your charger reads 14.4V on the side (which also means the batteries are disconnected!).
By the same principles, it MAY be possible to revive the old cells (done it twice! ) using a AC adapter at the same settings.

 

Thank you, Yusiang. The connectors certainly weren't all the same, but maybe they are now? There is significant engineering incentive to have all the same parts (battery charger PCB in the laptop and battery protection/fuel-gauge PCB in the battery). However, there is also a great incentive to have the batteries not fit in other laptops. That makes sure that we buy new batteries and don't get them from some generic and cheap source. From a materials perspective, the laptop battery should not be $120. It should be closer to $50.

 

The connectors are actually very similar across laptops, and follow a general pinout pattern, but sometimes that pinout is just slightly different from one manufacturer to the next. That's great you were able to start up the computer on a different laptop battery, and actually could give us a great clue in terms of what battery chips are compatible with the G73/74.

Yuisang mentioned recharging Li-ion batteries with 4.5-5V... I can confirm this works, but should not be used on a full or nearly full battery, only one that can accept a lot of charge (ie: somewhat dead or used). You can just use an old phone charger and cut the cables, as phone chargers range from about 4.8 to 5.3V and only put out 1A at most so you're less likely to kill something. You should only connect the battery to this voltage for a few minutes at most, as it's a rather harsh way of putting charge back into it.

 

Yusiang1998 Voltage in EACH CELL is 0. I can try to revive the cells (Constant current - constant voltage) but there is not really a point of doing this. The point is to try to reset the microcontroller on the battery circuit board. As i said before, i need to go back to UK first to start experimenting on this.