When you set the charge limit, do you disable the battery extender at the same time?

When you set the charge limit, do you disable the battery extender at the same time?

Here's what I did this morning - hopefully this answers your question:

1. My charge limit had been correctly adhered to yesterday, plugged in with charge limit enabled to 80% and the battery extender disabled. I then closed the lid to sleep it and then unplugged it.
2. My laptop had been sleeping, unplugged for 18h+
3. I first plugged it in, then pressed the power button. Upon waking, the framework logo showed. So it was in a deep sleep / hibernation.
4. after 10 min or so plugged in, it passed my charge limit. This was now the 2nd time since upgrading to 3.07 where I had had the charge limit working, but then it stopped working.
5. I rebooted and swapped both charge settings: I limited the charge to 100%, and enabled battery extender.
6. I rebooted again, and only changed the charge limit to 75%. After booting though, it was still charging past 75%.
7. I rebooted again, and limited the charge to 100% and disabled the battery extender.
8. I rebooted again, and only changed the charge limit to 75%. After booting, it stopped charging and began to discharge to 75%.

We found a known issue that "If the battery extender mode is disabled, and charge limit is set below 100%, the charge limit will not be enforced causing the battery to charge to 100%."

You could try to enable battery extender first, then see whether the charge limit works normally or not.

We found a known issue that "If the battery extender mode is disabled, and charge limit is set below 100%, the charge limit will not be enforced causing the battery to charge to 100%."

The issue is reproducible both with and without the battery extender mode enabled.

@quinchou77 @kiram9 @JohnAZoidberg There are many confirmed reports of the broken charge limit at the Ryzen BIOS 3.07 BETA announcement thread:
https://community.frame.work/t/framework-laptop-13-ryzen-7040-bios-3-07-release-beta/65582

And Framework have shipped 3.07 as stable, even with this known regression....

Hi,
I don’t have a FW13 AMD, but I do have a FW16.
I have not observed the battery problem. I limit mine to 70% and it stays at 70%. I disable the battery safe feature because I don’t need it. If you power off and unplug the laptop every night, it never will trigger the battery save feature.
But, my understanding is that battery charge/control is done by the EC (embedded controller) which is 100% open source. So, if people are seeing battery charge problems, they could look at the source code and suggest fixes. I have made changes to the EC code that fix some bugs in the EC, but none of those are charge related.

Algorithmically, the battery needs some input parameters:

Fixed set by user in the BIOS:

global charge limit
battery save enable/disabled
battery save trigger setting.

Measured values:
4) battery current charge level
5) EC uptime.

State that must survive a power off of the EC / laptop
6) state of hysteresis cycle.
1-charging to the limit, and switch to state 2 when charge limit reached.
2-stay idle while AC plugged in, unless the charge has dropped below "limit -5", then switch to state 1
Note: There appears to be a bug in the current EC code, that causes state 2 to actually discharge, instead of stay idle.

The rules:
Say you have set the global limit to 80%.
That is fine until the uptime triggers the battery save. It then is looking to use the 98%, then less over time.
That set of rules is odd and suspect in my opinion, but is consistent with what problems people are reporting. I would have chosen a different set of rules.

My suggested work around is therefore to never trigger the battery save function or set the global limit to 100%. Then reset the EC so the uptime is reset to zero and not triggering the battery save function.

To reset the EC:

1. power off the laptop, not suspend.
2. unplug the PSU for 2 minutes.
3. plug power back in.
4. power on laptop.
5. use ectool console to check the uptime has been reset.

Summary: unplugging the power for a few seconds does NOT reset the EC.

Hi Everyone,
We have reproduced this issue and have a fix that we will be releasing to stable before the end of march.

@kiram9
If the fix is a change to the EC. Please point me to the git commit so I can have a look.

by the end of march

Is this still happening? March is almost up.

@mercertom 3.08 beta has been posted to forum https://community.frame.work/t/framework-laptop-13-ryzen-7040-bios-3-08-release-beta/66858

@quinchou77 doesn't exist on lvfs-testing, as the notes claim.

I just get:

Devices with no available firmware updates:

@mercertom LVFS is restored, you can try it now.

@quinchou77

I still get

Devices with no available firmware updates:

The LVFS seems remain unstable.
'fwupdmgr refresh --force' is intermittently failing.

Devices with no available firmware updates:

There's also a fwupd bug that we've been hitting recently, that I think causes that error message.

fwupd 2.0.6 is affected, 2.0.4 and 2.0.7 is not.

See fwupd/fwupd#8510

Thank you, Framework, for acknowledging the situation and following through with firmware v3.08 to fix this; I Appreciate it.

@timwenger Thanks for the feedback. Assuming you've disabled the battery extender in the BIOS, the battery limit should work correctly on your system with version 3.08. Just sharing below information. I'll close this issue tracker and follow up on the other open one regarding the related issue.

We have identified an issue in version 3.08: when the battery charge limit is set below 100% and the battery extender is enabled, the device incorrectly charges to 100% after AC power is disconnected for more than 30 minutes. We are prioritizing a resolution for this issue.
To use the battery limit function, disable the battery extender and then set your desired charge limit.

Just wanted to say that I'm having the same bug on BIOS 3.07! Frustrating to say the least, especially since I was instructed by FW support to upgrade BIOS version after receiving a faulty replacement part. When will 3.08 reach stable?

@timwenger Thanks for the feedback. Assuming you've disabled the battery extender in the BIOS, the battery limit should work correctly on your system with version 3.08. Just sharing below information. I'll close this issue tracker and follow up on the other open one regarding the related issue.

We have identified an issue in version 3.08: when the battery charge limit is set below 100% and the battery extender is enabled, the device incorrectly charges to 100% after AC power is disconnected for more than 30 minutes. We are prioritizing a resolution for this issue. To use the battery limit function, disable the battery extender and then set your desired charge limit.

@quinchou77 Yes, battery extender is disabled and the battery limit is working correctly on my system in 3.08. Thanks for also informing users of the outstanding 3.08 issue when the battery extender is also enabled.

Since the posts here are all about Windows, I want to note that I'm seeing this with Ubuntu as well. Since the 3.07 upgrade, the laptop charges to 100% despite the limit set to 80%.

Will try if 3.08 fixes that once it's available.

Thanks for being transparent!

Hello,

I have updated from 3.07 to 3.08.
With battery extender disabled, the battery limit is working correctly.

However, here it is indicated :

This release modifies the battery charge limit functionality to add a 5% float range. [..] the battery will not start to charge until the battery has dropped 5% below the charge limit.

Yet, once the battery charge has reached the limit I have set in the bios, the charge alternates between charge and not charge every 30s - 1min.
I was able to notice this since my Fedora system plays a small sound notification each time the system charge or stop to charge. For now, I have disabled system sound as it is very annoying.

I post my issue here as the problem is directly related to this first issue, but if you prefer, I can open a new issue.

Don't hesitate to ping me if you need more information.

Thanks in advance,

I have seen the mini-cycle behavior on my FW16.
There are a number of edge cases, here are a few:

1. It can do the mini-cycling at almost any point. At the charge/discharge point and at the Idle/charge and idle/discharge points.
2. If the battery re-calibrates itself at a particular moment, it can go on to charge to 100%, ignoring limits.

These are all fixable by fixing bugs in the EC source code.