Last night I encountered a strange issue where I thought my dGPU was being limited to 30W, but it turns out that the laptop simply wasn’t charging. I did some testing and found that the non-working ports (1,2,4) have 0A draw while only port 5 has an actual draw on it. I have no idea how or why this would happen. I tested both with and without expansion cards and saw the same result. Does anyone have ideas, or is this possibly an RMA issue?
System:
Framework 16 Ryzen 7 7840HS w/ RX 7700s
Fedora 42
BIOS 3.07 (stable)
The ports still work for usb, display out, and etc.
Update: I tried a mainboard reset and that didn’t help.
Update #2: I was able to get the ports working again! Per a suggestion in another thread I disconnected the battery in bios and rebooted, which seems to have worked!
Try reflashing the bios? Framework would be the best ones to ask, so put in a support ticket. But I think the firmware/bios has separate parts for each charging port. If those failed to update I could see them not working to charge.
On a FW16, the EC can be reset with this proceedure:
Power off the laptop, not suspend, power off.
Remove the PSU (disconnect the power brick)
Waited 120 seconds (its actually about 45 seconds from when the power light goes off and the brick is disconnected, but I say 120 seconds in case the laptop takes a while to switch off.
This will power down the EC. I.e. reset it.
Switch the laptop back on again will start the EC up again.
One way to tell if the EC is reset is to use “ectool console”. It has timestamps on it. If the timestamps start again with small values, your EC has been reset.
Ports 1,2,4 are connects to one power rail.
Ports 3,4,5 are connected to a second power rail, so maybe you have a hardware fault with one of the power rails.
That could certainly explain the pattern. I vaguely recalled hearing a detail like this before, so I was assuming there was something similar going on under the hood.
When you talk about resetting the EC, is the battery still plugged in?
Fascinating. If this happens again I’ll try that (hopefully it doesn’t). Turn off, unplug, wait, turn on.
Just to provide another detail for future readers, Fedora/KDE did detect the AC connection and automatically activated my AC performance profiles when I was plugged into the bugged ports. This is part of why I had originally suspected that something had gone wrong with my dGPU under bios 3.07, as my device was still running with my AC mode scripts. Only a closer investigation showed that I was actually on battery.
