Frankenstein laptop after repacking Asus A32-N61 Battery 18650 cells DRM fail

[josh-nix]

Hi,

I'm hoping you know how to jumpstart this battery (picture attached)
Asus A32-N61 (original battery)
Rating 10.8v 5200mAh

Inside, of the 6 cells, 2 were internally shorted, 1 was open circuit and the other 3 were in good condition.

I've repacked it with 6 new Panasonic NCR18650B 3400mAh cells that I bought from you guys recently (tested to 3200mAh by Opus C3100 charger @ 1A)

At first I put the cells in at a low voltage.
First pair of cells was 3.5v and the other two pairs were 3.7v

`acpi -b` reported "27% charging at zero rate, battery will never fully recharge" That % didn't change after 20 mins so tried manually charging each pair to 4.17v
After charging, at rest, the pairs were all 4.10~4.12v

The laptop wouldn't power up from the battery.

Now `acpi -b` reports 95% charging at zero rate, battery will never fully charge."
When I pull the AC power, the laptop dies instantly.

Does it matter what order you connect the BMS terminals? I started from ground and moved linearly towards the +11v
(I mean would it be better to start with ground and +11v and then connect the 2 middle terminals afterwards, or does it not matter?
The BMS seems to be operating because it's talking to the laptop, reading the cell voltages and estimating state of charge.

I haven't pulled the sticker off the BMS chip, but it seems to be
Texas Instruments BQ20Z45R1 Datasheet: https://www.ti.com/lit/ug/sluu387a/sluu387a.pdf

==================

I found
https://www.laptopu.ro/community/laptop-battery-chip-reset-and-repair/asus-a32-n61-issue-after-recelling/

Seems very technical and specialized (and equipment cost is for refurb pros only). I'm hoping you know some simple trick to make the BMS reset?

==================

Thanks for reading and any help/tips/suggestions that you can offer.

[SMdude]

Ha!

Been there done that!
If you tie the reset pin on the controller to be in a reset state it won't record the fatal failure of the cells being disconnected.
Once it sees this happening it records a failure and the battery will never work again.

Next lesson, those cheapie replacements that claim more capacity are garbage. They won't even give you as much run time as original and will be useless in about a year.
It is one of those sad cases where you just need to pony up and spend a bit extra and get a genuine replacement. Been there done that and now happy! I spent enough in the meant time to have bought 3 originals!
However, if you wanted to be really tricky, get one of those cheaper replacements(don't put it in your laptop and charge it) and rebuild that (tie reset low)with the good cells before you first put it in the laptop and power/charge it up. Also, make sure on the first charge cycle you let it charge up fully, otherwise it will be dead in no time at all. That is when it flies out the door at a great rate of knots!

The only other way is to find some software/device that interfaces with the battery and you reset the eeprom parameters.

Good luck!

[josh-nix]

SMdude, thanks for your reply.
In my case it wouldn't have helped to tie the reset pin while removing the old cells, because the original battery had 2 shorted cells. So the battery had already failed.

Yes, the cheap generic (fully assembled replacement) batteries are garbage. I'm not sure if the BMS have planned obsolescence built in, or if it's only because they put garbage cells in them.
But at best I'm aware that cheap replacement batteries will have garbage cells in them, possibly 10 year old cells.
I've tried 3 cheap replacements for my X220 Lenovo laptop. The first replacement which was supposedly a 9 cell, didn't work at all.
The 2nd replacement worked for 2 weeks.
The 3rd replacement worked for exactly 1 year.
Then I reverted to the original X220 battery that now lasts about 45 mins.

Yeah, it wasn't fun repacking the cells. I had to dremel the plastic on the sides because the original had tabs of the perfect length spot welded, which is thinner than what I've done, where I got tabbed cells and soldered on the tabs. I had to redo all of my solder joints. Initially using a decent amount of solder so the solder's internal flux could wet the tabs. but once the tabs were wet I had to remove 80% of the solder to make thin solder joints. Was a hassle.
Was just an unreasonable fiddle to get all the cells in place because there's pretty much no space inside those stupid battery cases. Even just lining up the cells to solder them and push them down, and put paper between the anode/cathode and the tab so I didn't heat the 18650 etc. Was such a hassle.
And to end with the stupid BMS refusing to self reset after seeing the cells are fully charged. Really not great.

[josh-nix]

SMDude, is it worth buying a generic battery, tieing the reset and then replacing the cells in the generic with new cells?

This is my plan:

Plan A: If there is an easy way to talk to the BMS and reset it, I'll do that.
Plan B: I dunno.

I've already hacked the shit out of this laptop in other ways. I could just hack it's power delivery.

(read all steps before you judge, I'll make it safe in later steps)
(If a step fails, go to hack option B)
Hack option A:
1. Bypass whatever output safety switch is on the BMS so that the power flows from the battery to the laptop, so it can power on, on battery.
(assuming that works, it means I can discharge the battery and run the laptop, and the laptop will be able to read the battery % in software from the disabled BMS)
2. Get a 3S BMS from Amazon, and wire it up to the 3 cells (in addition to the original BMS)
3. Connect the 19V input to the Amazon BMS.
4. Unbypass the output on the original BMS, and connect the output of the Amazon BMS to the output of the original BMS.

At this point, the laptop can charge/discharge software can read the battery % and there is short circuit or over-current protection.
And I can change the cells at will.
If there's enough space, I could even hack the bottom of the laptop so that there are 18650 holders there. And I can just swap out cells on the fly (assuming cells are at the same voltage)
I could even hot swap cells while running on battery power only. Since it's a 3S2P battery.

Hack option B. (I think option A is viable and so this is very unattractive to me)
1. Remove the original BMS, lose the software battery % reading.
2. Install a "big" buck SMPS inside the laptop.
3. Rewire the cells into 6S so the batt voltage will be 21.6~25.2v
4. Feed the 6S into the buck, feed that into the 19v input of the laptop, so it thinks it's AC powered.
5. Install a 6S BMS for charging.
6. Oh crap, how's it going to charge, since the ac input is 19v. Nah, it's a terrible idea. Too complicated and bulky and inefficient.

Option C:
Throw the 2011 Laptop in the e-waste or make it into a low-power server with shitty IO, and get another laptop.

[abdulbadii]

shot few close up pic.  and acquire  schematic on DCV || to batt. supply lines to Mobo, upload here

[amyk]

There is some useful information here on resetting the BMS: https://www.eevblog.com/forum/chat/experiences-with-ti-bq-management-ics-when-revelling-laptop-batteries/

This is the software for your particular BMS:
https://www.ti.com/tool/BQ20Z45-R1_BQEASY-SW

There are "unseal" and "full access" keys, but the defaults (36720414, FFFFFFFF) might work. Otherwise you can try some power/timing analysis while attempting to bruteforce it (pick it like a lock.)

[SMdude]

SMDude, is it worth buying a generic battery, tieing the reset and then replacing the cells in the generic with new cells?

Well that is what I had planned to do before trying the actual supposed genuine battery, which seems good so far a year later.
Of course there is risk to it in that if you stuff up and it records a fault then you are back at square 1. I would try with a failed but still working pack to put it into and take it back out from reset just to be sure before you try a new one! Also if it has an external eprom I'd copy the contents, that way you may be able to restore it as a "virgin" battery later on.

Some of the tools posted in the other thread that was referred to look good, but yet another rabbit hole to jump down..

[josh-nix]

SUCCESS! (kind of, technically)

@abdulbadii I took a few pics of the face of the BMS board, is that what you wanted to see?

@amyk Thank you very much for the links.
The one suggestion was interesting, that some of these chips will reset if you connect the battery output voltage to the + terminals where the power is supposed to go, (bypassing the mosfets momentarily)
So I tried it but it didn't reset the BMS.

The BMS hacking is not worth my time to revive a 2011 laptop. I'd rather spend my time working and just buy a new laptop. Sad that the DRM bullshit wins like that. But anyway.

So as a proof of concept test for my idea, of installing my own BMS inside the laptop, then saying fuck you to the BMS...
I said fuck you to the BMS and soldered a wire from the +V batt to the output of the 2nd mosfet. For whatever reason the mosfets are connected in series. I guess as a safety incase one mosfet accidentally turns on. Both mosfets are supposed to turn on to let the power through, it seems.

But then when I pressed the power button with just the battery installed, nothing happened.
Then I plugged in the power cable, the charge light comes on.
Unplugged the power cable, tried to power on, nope.
Plugged the power cable in, powered on, immediately pulled the power cable out and the laptop kept running.
Then I proceeded to use the laptop for 15 minutes on battery power.
I ran `acpi -V`
The laptop can tell me when it's on battery vs AC power. (says AC: offline or online respectively)
But now instead of showing me the battery percentage (as it did in my previous tests)
Now instead it says "Battery status unknown. 0% fuck you too."

I re-tested. And it won't power up without the power cable in. But once it's on, I can pull the power cable.

That doesn't scare me. I can add a button to fake power on the barrel jack momentarily.

The lack of battery % from the BMS is slightly disappointing. I was hoping that I could use the BMS as a software fuel gauge... even if it's wildly inaccurate and just gives me some info based on the voltage at least. So I can write a service that pops up a message when the batt is low and shuts down my laptop if the battery is very low.

But since the BMS's official response to my "fuck you" is "fuck you too" I'll need to completely remove the BMS and implement some alternative for knowing when the batt is low. A really basic hardware method is one of these voltmeters that have an alarm (image attached)
I've used one on a AA battery powered RC car that I modified to run on 2S LiPo. Works well.
I really don't mind cutting a hole in the top of my laptop for the voltmeter to be visible. It's a frankenstein laptop already anyway.

But it would be nice to have a software solution for capacity. I could probably program an arduino to read the voltage across a shunt and work out the current and therefore runtime, then feed it to the laptop via USB.

But my only experience with microcontrollers is programming PIC's in assembler. I'm totally noob with Arduino and have never even read from an ADC before so I'd have to figure out the potential issues with reading the shunt voltage which would be close to a rail voltage etc.

But I'd only do something like that later, after spending a few weeks/months with a simple solution that works (which would be simply using a 1-8S LiPo voltmeter/alarm.

I found this for measuring current
https://learnarduinonow.com/2015/05/11/reading-current-shunt-with-arduino.html
Instead of using ADS1115 could I just feed A1 through an opamp with some gain, to overcome the 4.8mv/step issue without adding an ADC?

[amyk]

For whatever reason the mosfets are connected in series. I guess as a safety incase one mosfet accidentally turns on. Both mosfets are supposed to turn on to let the power through, it seems.

As they have a body diode, they don't need to be on to conduct in one direction. Since current can flow in both directions to/from a battery, one is for stopping charge, and the other is for stopping discharge.

But it would be nice to have a software solution for capacity. I could probably program an arduino to read the voltage across a shunt and work out the current and therefore runtime, then feed it to the laptop via USB.

But my only experience with microcontrollers is programming PIC's in assembler. I'm totally noob with Arduino and have never even read from an ADC before so I'd have to figure out the potential issues with reading the shunt voltage which would be close to a rail voltage etc.

If you really want to, you can program your own "fuel gauge" ICs, maybe even a TI BQ series like what's in the original pack, that will do exactly what you want. If you're determined to make your own BMS using ADCs etc (which will unlikely to be as accurate), then the relevant specification for the protocol between the laptop and the pack is at http://sbs-forum.org/specs/sbdat110.pdf .

[josh-nix]

Sorry if some of the sections are a bit rambley. I put headings.
TLDR, skip to the end and see image attached.

@amyk thank you very much
For the PDF link, I've downloaded it for safekeeping. Your knowledge on this stuff is incredible and I appreciate the time you've taken to contribute to my thread, thank you very much!

2 Mosfets for +Vbatt
Oh, yes of course what you're saying makes sense Regarding the 2 mosfets with internal diodes for charge/discharge. Yeah it was pretty late and I wasn't really thinking about charge, I was laser focused on being able to discharge haha.
What made me think they put 2 mosfets in series for charge purposes is that they put some white sealant on one side of both of the mosfets. And those were the only components on the board where they added it. So I thought maybe they're super paranoid about liquid getting onto those contacts or whatever and series'd them for extra safety.

A ceramic knife is useful for electronics
I didn't feel like de-soldering the cells so I cut/scraped the sealant off with a ceramic knife while the BMS was powered up.

====

I did a bunch of reading and thinking
So I skimmed through the SMBus smart battery protocol PDF and also looked to see if anyone has implemented smart battery on arduino before.
And I found this and read through the code
https://forum.arduino.cc/t/smbus-a-smart-battery-emulator/645352

And it looks like basically exactly what I had in mind for doing a fuel gauge with an arduino, in terms of measuring voltage and current.
Regarding your PDF for SMBus protocol for smart battery, I actually wasn't planning on emulating a battery. Too much hassle (I'll explain in a bit) I was just thinking to talk to the laptop via USB and just write a service to notify me or shut down, etc.

But really, after reading about your suggestion to implement another TI chip, with it's fine pitch SMT and whatever else, probably requiring a custom PCB, and all the hours that would need to be invested into doing something like that.
Or the SMBus battery spec.
Or the arduino sketch I mentioned...

=============================

Clarity on the goal
I took a step back and asked myself "What do I want?"
"I just want to use my laptop on battery."

=============================

Devils advocate, do I really need a fuel gauge?
Interestingly I recently bought a 18V vacuum cleaner. A ton of 18V power tools don't even have an onboard "fuel gauge" (neither voltage or capacity measuring in neither the battery nor the tool)
They simply fade when they're almost dead, and then the tool stops working when the battery voltage is too low. (2.5~3.3v/cell etc)
When I think about the lithium cells that I've used for my past projects. Or my various 12v 100Ah Lead Acid batteries that I rely on during utility power failures (rolling blackouts aka load shedding)
If I want to guess SoC. I consider/guess what my load has been, and guess how long, then I just reduce the load to the minimum I can easily do (put PC to sleep etc), look at the voltage and take a guess how much power is left.
Yes it's crude and low tech, etc. But it works for me.

When my 18V vacuum cleaner is low, I just put the battery on the charger and put in another battery. No big deal.
I don't time how long I use the vacuum for. But I know that the manufacturer says it'll run for about 10 mins on a 2.5Ah 18v battery. So after a few mins I press the button on it, that shows 1,2 or 3 lights.
Which is most likely just a voltmeter, but could be SoC. I've got a cordless drill that has no meter on the battery or tool. I just pull the battery out and check it with a voltmeter.

So even if I simply have a 1-8S Voltmeter and alarm module connected. That'll make a noise if I need to shut down urgently, if I forget about the battery. And if I'm not available to shut down... the BMS will cut the power. And I can easily check the voltage to guess about SoC in combination with guessing how much I've used it, etc.

============

Strategies to make micro-managing remaining capacity un-necessary
1. Make the battery fucking big (high capacity)
2. Make the battery swappable

I will do some combination of these.

============

What's really important
As you can see. When I needed to crunch numbers on this laptop, back in 2013 and it was thermal throttling all the time. I didn't worry/care about how the bottom of the laptop looks...
I focused on my objective, which was to get what I needed done (solve overheating issue)
So I made my own heatsink and thermal epoxied it to the heatpipe. Then stuck those peel-and-stick-on little copper heatsinks (which mostly fell off after a year etc (then I pulled off the rest to avoid losing them) but even the heatsink I made, plus the holes in the bottom, work incredibly well.

So what's important to me now?
1. Saying "fuck you" to Asus and their battery DRM.

How do I successfully say fuck you to the DRM?
If I was a wizard with power glitching etc and felt up to brute forcing the password (or trying random passwords) then that would be a great way. Cos then I'd restore original functionality with the high capacity cells that I've installed.

But I'm not such a wizard, nor do I want to invest in becoming one.

Even talking to the battery at all with the SMBus is too much hassle for me. I've never done anything with SMBus. The SMBus protocol and smart battery protocol is totally foreign to me. To set something up and learn about the protocol isn't a good use of my time.

So conceptually, how do *I* (personally) successfully say "fuck you" to the Asus DRM, and restore the ability to run on battery?

1. Spend the minimum amount of time to get it running.
It doesn't have to be high-tech. Doesn't need fuel gauge. It just has to work, be usable, not destroy the battery and safe.

Why?

Asus put in this DRM to make me suffer by coercing me into doing 1 of the following
1. Pay their inflated prices for their genuine battery. (it's not in production anymore)
2. Buy chinese fake batteries that last 2 weeks to 1 year max.
3. Waste countless hours of my life hacking their DRM.
4. Accept defeat and buy a new (they would hope Asus) laptop
5. Accept defeat and only use the laptop plugged in.

All of the above involve accepting suffering inflicted upon me by Asus. And I don't accept that.

So for me to win, I need to use the laptop on battery AND not suffer.
That's how I successfully say "Fuck you Asus."

So this is how I will do it
1. Connect a cheap off-the-shelf dumb BMS (for safety)
2. Install a voltmeter/alarm module (guess SoC and get feedback)
3. Hack the case (probably just the bottom) as necessary to make stuff fit
4. Install something so I can press a button to fake power at the barrel jack DC input for long enough that it will turn on.
Be happy.

I'm going to use 18650 holders to make it super replaceable. Fuck them and their shitty battery chassis that's a royal pain to get cells in and out of.

Later if I'm feeling very frisky with an Arduino, I might do some sort of fuel gauge. Most likely I'll get a new laptop and lose interest in this one before that can happen.

==============

TLDR
Final proof of concept test for powering the laptop from a dumb battery (see image attached)
I wanted to be able to guarantee for myself that the laptop is not reliant on esoteric shit (current state, that could change) with the current BMS.
So I tested powering it from a battery without a BMS. Success! (same as last's night test, I provide AC power, press power button then pull the plug and can boot and reboot, etc on battery power)

Obviously that was just a proof of concept test.
I didn't leave the DC barrel power plug in for more than a second. And would not run the batteries dead like that without any BMS at all.