Exactly. with the 180w you STILL get battery drain when maxing performance on this laptop. Of course a lower wattage PSU will not allow max performance. It’s common sense.
He is doing the equivalent of buying a Ford Mustang without an engine, putting a Honda Civic engine in it, then returning the Mustang because he can’t get the same horsepower as the other Mustang owners.
He has posted this a few locations in the forum, without citing any source. I’m going to assume it’s just frustration and made up.
He isn’t following support suggestion to use the proper power supply and so they are at least letting him return it to avoid any future support requests from him. Smart move I think.
Basically, if you are not using the framework 180 watt power supply, then the laptop auto defaults to a 60 watt power profile, no matter which settings you have in Windows. The actual firmware of the laptop is programmed to do this. My guess is that it is the 13 inch framework firmware, which it is falling back to.
This means that in any mixed use workload; the GPU uses a Max of 35w typically instead of upto 100W. This is true even if using a 100W or 140W adapter.
To make matters worse; the whole laptop is jittery when using a 140W adapter (and will have the same problem with any adapter supporting over 20V) because there is no fallback firmware for over 20V power draw. So the laptop seems to ramp up and down in power randomly and the GPU and the entire system in general hitches randomly and frequently.
This isn’t a problem if using a USB-C adapter which doesn’t support over 20V (which most 100W and under chargers do not)
Interesting. I’m either using the 180w charger or a multiple one that outputs 100w on the plug that I use for the laptop. Is the code for the firmware available (I suspect not, since it comes from Insyde). I am not doubting you, I’m just trying to understand it better. If it is an issue with the firmware around how PD is being handled, I expect that it would eventually get resolved. Framework has adjusted it already, but I can’t recall if that was for the 13" or the 16".
No, I didn’t do that, let me explain my feelings. My frustration is real because I cannot use the laptop with any adapter other than the included 180 watt framework charger which I did not buy because I have no use for.
While I could use that charger at my desk, the laptop isn’t portable IMHO with such a big brick. I returned the Asus flow x13 in 2022 for the same reason. They forced me to use the included brick with its clunky non removable cable and large power connector.
I didn’t “buy a mustang without an engine”, let me give you a better analogy. It is more like I bought a mustang without a fuel tank because it said it supports any size fuel tank. Then when adding a 15 gallon tank VS the original 18 gallon tank; the mustang now has 100HP instead of 450HP.
I can tell you that personally from looking at the power draw of the laptop itself via a USB-C analyzer that there is absolutely nothing preventing the laptop from running at, let’s say, 85 watts on the GPU from my 140-watt power supply because the laptop only exceeds 90-watt draw when it is charging its battery. Once the battery in the framework laptop is fully charged; the laptop won’t draw more than 75W at full load from the USB-C port.
My charger is 140W so that is 75W more… Clearly in low cpu usage scenario; the 140W charger could fully push the GPU to 100W. In a higher power scenario; I wouldn’t be able to get the full performance.
I fully understood this when I ordered the laptop and was fine with it.
I waited 1 year for the laptop to arrive and it is locked to 60W total for the full system. This means that under a full CPU load; the gpu gets around 25W max and under a mixed load it gets 35W max… Under an almost fully GPU load it can get 48W max.
So in the best case scenario regardless of which non 180W power supply I use; I can get 1/2 of the gpu performance. Most of the time it is 1/3 or less. Effectively speaking; sometimes the dGPU isn’t that much faster than the iGPU performance if it was set to say 65W purely on the APU.
When my 2021 Alienware m15r5 w/ 3070 offers 4x the performance in the same title… It really seems to be a total joke.
I was willing to work with framework to solve the issue (even work with them on the firmware since I am a software engineer) and willing to wait.
They pushed for a refund instead because clearly they have no intention of supporting any other power adapters outside of their own.
Please edit this post to remove the ad hominem attacks in the first and third paragraphs. It is not tolerated in the Framework Community as it is in violation of the community guidelines. Your post will be hidden until you do so. Further personal insults or other violations of the community guidelines may result in this thread being closed.
1 Like
The issue is directly in the firmware. They sent me the firmware to look at specifically since I am a developer. I will not share it here on the forum because that would be wrong.
Since at first they told me that I should not have this problem and that I just needed to use Windows 11 in order for the laptop to work properly.
It wasn’t until I debugged the USB-C power input using my little tool and also did logging using hardware info and GPU-Z and reported the information back to them that they then started recommending only the 180 watt power supply.
I suppose that if I were to take the time to patch the firmware and send it back to them that they would likely publish the firmware for everybody and the problem would be fixed. But the question is really if I want to take the time to learn all of the intricacies of the firmware and update it to work with all power supplies.
The thing is it would be extremely tough for me to fix the firmware considering that I have no way to compile and deploy the firmware myself in order to even test it. So regardless of me having the code access, I cannot actually change the firmware or test it.
Thanks for the reply. I agree, please do not share it. It sounds like a bug, and I can’t imagine that they won’t patch it at some point. It sounds like it is not great now, and it will be ridiculous if a 240W power adapter does not allow the machine to run at its max. I am not surprised that they would not officially commit to anything, though.
Hopefully they will fix the firmware at some point if this is true. Framework has publicly proclaimed they will be working on improving the firmware and update frequency so I have hope. And when the time comes that greater than 180w PSUs come out, you can bet there will be many complaining in here if they are not working properly.
Telling you to use Windows is very discouraging and contradicts their pledge to support Linux. Many, myself included, only use Linux. I hope that was a mistake on their part.
Also, from personal experience. I use a 100w supply by my bed and it quite often will start pulling from battery when the cpus ramp up for a bit. Not even using the gpu at all and it will still go down.
I would hope they would fix the firmware. I mean I don’t want to return the laptop. I like the design overall and the ability to upgrade it. I obviously wanted it so much that I waited 1 year for it without canceling the order.
They did mention that in support that they did support linux with their drivers. For me I was using Windows 10 as I don’t have any systems with Windows 11; the whole OS goes against my Ethos.
I wish I could switch entirely to Linux (part of me ordering this laptop w/ full linux support was going to have me dual booting windows and linux) but I do develop software for windows. So I wanted a system where I could use my Windows 10 and then boot into Linux when I am not needing windows (which would actually be a good portion of time).
The 100W USB-C isn’t utilized properly which is why it will draw from the battery. The only portion which uses the 100W charger would be the max charging speed of the battery. It can actually pull upto 125W from my 140W charger because it can use the rest of the power to charge the battery. But as I mentioned before; it wouldn’t pull more than 80W from the 100W USB-C no matter how I used it.
Also if you have a 100W charger in your room; make sure the cable you use is certified for 100W; otherwise it will only use 65W and the laptop definitely pulls power from the battery quite often when using a 65W charger.
You can run windows 10 in a vm from linux for your development, depending upon exactly what it is. I have windows 11 on the 2230 in my 16" and pass that through to the vm unless I have some need to reboot to windows for something like a firmware update.
Virtual machines don’t work well for debugging nor do they work well for games because I can’t just pass through my graphics cards.
I do gaming software development so I really need power and memory. This is also part of why i bought the laptop because it can utilize 96GB ram.
Thanks for the reply. That makes sense. If the Framework 16 doesn’t wind up working out for you, hopefully you find something that will. I will be curious to hear where you land. Have a great day!
1 Like
I don’t currently have any other system on my radar… I haven’t even started looking. I was just hoping to fix the framework with the team haha… I’ll post back when I figure things out.
I might just keep my M15r5 for now w/ 5800h, 64GB ram & 3070m.
I think you are bluffing, or perhaps using a smokescreen to convince people that you have some sort of insider knowledge.
The embedded controller mediates charging and power limits, and it is running a fully open source firmware. I have extensively documented how to both make changes to it and install a new version (admittedly for an earlier model, but the principles are the same), and have written the very software that the community uses to do so. The package power limits you’re talking about being set by charger power availability and OS power plan are exactly here.
I would be happy to document how you can change them and flash your embedded controller to test this stuff out, for the good of the community. 
2 Likes
I definitely am not bluffing because that would be pointless. I wouldn’t want to return my laptop if it just worked. After all I waited literally 1 year to receive it and wasted about 2 weeks trying to make it work properly.
I was sent a link to the firmware directly in customer service after explaining my background. They didn’t say if it was public or private but I assumed it was private. Considering my background; I am used to companies sending me information that shouldn’t be shared.
I see that the link that you sent is indeed the link that they sent in customer service.
I would be open to discussing with you about how to flash the firmware.
If you know of any links (or have personal knowledge) of the background of this firmware (documentation) then I would appreciate that too.
I have no problem actually fixing the problem not only for me but for everyone. After all I really wanted something like this to exist.
I am upset with framework for pushing me down the refund path instead of the path of fixing the problem. At least they did push for a refund instead of just ghosting like I am sure some other companies would have… however it still is unfortunate that the problem isn’t something they seemed to want to work on.
If you notice in that file starting on line 512: you can see part of the problem. The wattage for 140W = the same as 100W and under. The max package power @ 65W is 65W and it is 75W above that.
Also in the top of charger.c it literally checks to see if charger=180W and otherwise makes a totally different calculation for power.
I spent about 30 minutes tracing through the different paths and it is clear that things are always handled differently if the 180W charger isn’t connected. Since the same math isn’t used (like on line 500 where everything that isn’t 180W has a hard limit set) then the power isn’t calculated properly or is limited.
Anyway I look forward to your response.
If I am reading the code correctly I believe that the section starting at line is for ac power only - no battery. Above that is for battery only, and above that is for ac and battery.
In the charger.c file it looks to me like some alert is being scaled by 1.2 when the power supply is 180w. I haven’t dug deeper than that to try to understand why only for that case. I’m not sure what if any impact it would have.
The firmware is based on the ChromeOS Embedded Controller firmware. DHowett has written about it at length on his website, https://www.howett.net/
I find it hard to believe that any company is sending private IP without requiring you to sign an NDA, let alone indicating that it’s private. Also, if they just sent a GitHub link without asking for your username first, you should know that is public automatically.
1 Like
Thanks for the information.
As for github, there are things which are not listed but could be linked to directly, similar to pastebin. Also you would be surprised what is on github which isn’t for the public really to see… either way I didn’t want to out them if I wasn’t told that I could share it.
Also I worked for Amazon in the Vine program for years; I didn’t sign NDA’s overall but was sent stuff that was unreleased and was sent beta firmware, code and such for various products over the years without ever signing anything. So to me this type of communication is not new.
Well that is what it implies; but possibly there is some issue where the firmware isn’t setting the battery flag. The rest of the code for “UMA” has the same calculations for every adapter basically, so that clearly isn’t where the problem is.
The AC + DC section has the power switch from the 95/85W mode and this is what SHOWS UP in Radeon Control center.
However it isn’t what actually happens with the laptop. The power used doesn’t actually change at all.