Please keep this thread on-topic. Any off-topic replies will be deleted or split off into another thread.
1 Like
At least for me on BIOS 3.05, what is happening is if the battery is less than the max charge, then the dGPU will vary in its wattage. When fully charged it is fine, but once it drops a tiny bit, youāll see frame spikes as the wattage starts to fluctuate for half a second.
2 Likes
Mine shipped yesterday from Digikey (Ordered 28/11/2024, shipped 16/12/2024, about 2.5 weeks while the site indicated 32 weeks from factory).
Now Digikey shows 24 in stock for immediate shipping; get them while theyāre hot!
3 Likes
Thanks, reordered from digikey again as Mouser werenāt even trying to look into why they couldnāt supply me the PSU.
Since it doesnāt look like anyone else has posted this example yet, hereās a screenshot of my Framework 16 on Windows drawing power from battery while playing Helldivers 2 on the 180W charger (started playing at 8pm). Didnāt have more sophisticated logging software at the time unfortunately.
I knew a higher wattage charger would be nice, but I was surprised to see the RX 7700S drawing so much power (I guess all those cinematic explosions are pretty demanding). Thanks all for the info on 240W chargers, Iāll be picking one up as well.
2 Likes
Oh noā¦ We had a similar problem with other chargersā¦ looks like this Delta 240
w has the same āstuttering wattsā problemā¦
I recieved my unit today. Can confirm:
-stock 180w Framework charger: 100w dGPU pulling during 3DMark
-Delta 240w: power oscillates between 50 to 100w, as described in another thread.
My bios version is 3.03, now Iāll upgrade and see if it gets better on the newer version.
EDIT: I upgraded to bios 3.05. The charger now works flawlessly. While the stock one, on max performance, while running Furmark + CPU burner (10 threads) pulls ~20W from the battery, the Delta 240W loses 3-5W from the battery while keeping everything at max TDP: 100W GPU and 30-35W CPU. Now I dont have time to test further, Iāll keep you posted.
4 Likes
Got my 240w PSU, now to test it works!
1 Like
Can confirm that gGPU performance sucks if the 240w charger is plugged in and you try and game at anything less that ~90%
I am going to try and create a ticket for this and get it escalated beyond level 1 to report as a bugā¦
4 Likes
It looks like the same issue which was raised by Obasav in his earlier comment. He created github issue . There is comment from kiram9 (who works for framework) that he will investigate it try to fix in next release.
5 Likes
Thats much better than the traction I am getting in the support case.
1. May we know if you've followed this guide in installing the Windows 11? [Windows 11 installation on the Framework Laptop 16 DIY Edition](https://guides.frame.work/Guide/Windows+11+Installation+on+the+Framework+Laptop+16+DIY+Edition/305)
2. āCheck if the charger works stably with other USB-PD devices, if available.
3. If possible, test the laptop with a different 60W or greater USB-PD charger to confirm whether the issue is related to the Delta charger or the laptop.
4. Were you using a dock station, hub, third-party charger, or an eGPU before the issue started? If yes, kindly provide the specifications of the components used.
5. Disconnect any external devices (e.g., USB peripherals, external displays) while using the charger to rule out power conflicts.
6. Fully charge the battery to 100%, then let it drain to 0% under regular use. Recharge to 100% without interruption and see if this stabilizes performance.
7. Try to reproduce the issue with other ports.
These arent all terribleā¦ But I stated when I created the issue that this looks to be a firmware issueā¦ and needed to be sent to engineering. Then I get thisā¦
#6 is the one that makes me madā¦ as this isnāt healthy for the battery.
A temporary workaround for it (at least from what Iāve seen in my experimenting) is to set the laptop to balanced mode. The dGPU performance will be more stable but may not be the same level you would normally get in performance mode.
1 Like
This is required for battery calibration. Youāre right that with li-ions it produces more wear / stress, but doing it occasionally should not meaningfully impact battery health when normal usage wear is added in.
5 Likes
False for lithium batteries. Thats generally a myth from when ni-cad needed this:
https://saft.com/energizing-iot/lithium-ion-batteries-use-5-more-tips-longer-lifespan
Contrarily to some received ideas, Li-ion batteries donāt have a memory. They donāt need regular full discharge and charge cycles to prolong life. Itās actually the contrary: the smaller the discharge (low DoD), the longer the battery will last, the more cycles it will be able to do.
1 Like
Itās not about a charge memory like ni-cads had / have.
With Li-ions itās done to calibrate / recalibrate charge level estimation. Li-ions have a rather flat voltage/charge percent relationship so it harder to gauge current charge level. And with the normal wear that goes on with use, the charge level estimation that the device believes can drift from what is reality. Doing a full zero to 100% cycle allows it to recalibrate itās charge level estimation.
Whether or not charge level estimation could be an actual an problem here, with this 240w adapter issue, thatās another matter. But just customer service asking one to do a full zero to 100% cycle to recalibrate charge level estimation is not a problem or something to be mad about, in my opinion. Itās likely just a stanadard step they ask people to do when there are presented with some battery or charge issues.
5 Likes
I tried this last night. With the Fedora 41 change in how it handles power profiles, it actually did not help.
Actually, with the 180w adapter, I still lose battery in balanced mode now (where it used to maintain). This is seemingly related to how tuned handles the power profiles vs how PPD did.
Having the 240w adapted made the above a non-issue.
2 Likes
EDIT2: Bad news everyone. Turns out my quick testing (Furmark) is an outlier. Using both the 240w or the 180w adapter on Furmark it consistenly pulls 100w.
While normal gaming or 3DMark, the issue persists even on 3.05 for me as well.
Itās 100% a firmware issue, other USB-C chargers less powerful than 240w have the same problem.
Framework pls fix, we waited so long for the 240wā¦ only for it to be sluggish in performance. Plenty of us here for you to ask logs to, weāre here to help! 
3 Likes
And my charge level estimation is not currently inaccurate. If it was - and I could confirm that the Framework didnāt just measure voltage to get the % - then I would consider it.
However, a full 100-0% discharge is bad for the battery - and not something I am willing to do for the above request. It shouldnāt be done as a request unless its something like āwhen I unplug my charger, it immediately drops to 75%ā or something similar.
Limiting the amount of drain is the best for lithium batteries - with the ideal rate being from 40% - 80% if you really want to prolong life. Ideally, I would set my charge limit to 80% and just be happy - but the current issue with the 240w prevents that.
Im not āmadā at Framework for this issue - but rather just want to ensure they are looking at this from the firmware side - which seems to be what is being done in the bug filed by @Obasav - so it seems the right people are on this from the Framework side.
Had I thought to create a bug - I would have - I figured the best way to do that was through a support case. I was wrong 
If you could add your update to this bug?
I have to collect some data tonight and I will try and provide it as well - but it sounds like you have some solid data.
Adding this to the bug - but I drained my battery to test this out. Again, this does not happen at 100%ā¦ or even 90%
I start to notice it in the 80% range - and at 50% its really noticeable.
Here is a video of the issue happening.
Sorry, I had added to my post at about the same time you replied.
I agree that charge level estimation is not likely to be an actual issue in this case.
But it is a normal thing to ask, to me. And the impact of a single full zero to 100% cycle is low. A drop in the bucket. Also, considering how easy and quick Framework batteries are to replace if and when one get to that point, itās not worth worrying about, in my opinion.
Certainly, it can be annoying when it feels like customer service is asking about things that arenāt likely to be an issue at play. I imagine Iād be annoyed as well. The fact that there is a github issue that someone said they will be looking into is good, though. Anyone seeing this problem can add themselves as tracking it, plus add to the reactions, to let them know there are a number of people seeing this issue. Looks like you commented on the github issue as well.
I wouldnāt think the Battery management system chip uses just voltage to estimate charge level. They also attempt to track current flow in and out. And the algorithm might try to factor in predicted wear from the number of cycles, depth and temperature during the cycles, it all plays into wear. But in the end, charge level estimation still drifts, sometimes a lot. So a full zero to 100% cycle to recalibrate estimation is something one should do anyway, if you normally never let your battery go through a full cycle like that.
1 Like