Hardware/Software:
- Framework Laptop 13 (AMD Ryzen 7 7840U)
- BIOS: 03.18
- OS: Arch Linux, kernel 7.0.3 linux-zen
- Battery: 219 charge cycles, ~56% of design capacity remaining
Symptom:
The system does not detect AC power when the charger is plugged in while the laptop is awake and running. cat /sys/class/power_supply/ACAD/online reads 0, and the battery continues discharging (status: Discharging, current_now shows active drain) even with the charger connected.
However, if I plug the charger in while the laptop is fully shut down, it charges immediately and continues charging normally once I boot into the OS. If I then unplug and replug the charger while the system is running, AC is not detected again — the only way to restore charging is to fully power off and plug the charger in during shutdown.
When this first occurred yesterday, the laptop needed to sit fully disconnected and powered off for roughly 30 minutes before charging would work again on reconnect. Today, the same shutdown+replug sequence works immediately, with no wait required.
Over the last month, I noticed that when my laptop is connected, the OS sometimes shows it’s charging, but then for a few seconds it shows that laptop is not charging. This was happening before the issue became critical. The laptop is almost always wired up — I mostly use it as a desktop and only move it from room to room. I have the original charger, but with two different cables: one is the original, and one is an Anker.
What I’ve ruled out:
- Charger, cable, and port — tested the stock charger with multiple cables across multiple USB-C ports; the same charger and cables charge my phone and another laptop normally
- TLP charge thresholds — TLP (1.10.0) reports
Battery Care Plugin: generic,Supported features: none available, andcharge_control_start_threshold/charge_control_end_thresholdare not exposed in/sys/class/power_supply/BAT1/at all, so TLP has no active control path on this hardware currently - Suspend/resume — a plain suspend/resume cycle does not restore AC detection; only a full shutdown does
- Recent updates — no package updates in the days before this started; last kernel/TLP changes were from early May (6.19.11 → 7.0.3, TLP 1.9.1 → 1.10.0), so this doesn’t appear to be a fresh regression from a recent update
Given that a full shutdown (not suspend) is required to restore charging, and the first occurrence needed a longer disconnected wait than today’s, this feels like it could be an EC-level fault state rather than something OS/driver-side. Has anyone seen this pattern before, or know if 3.20 addresses AC-detection/charging-state handling specifically?