Major issues with FW16 & dGPU: Ports failed after using 240W third-party charger

Hi everyone,

I’m writing this because I’m incredibly frustrated and, frankly, disappointed with my experience with the Framework 16 and the RX 7700S module so far. I’m hoping for some advice or to see if others have faced similar hardware failures.

The Issue:

I realized quickly that the dGPU cannot be used properly for gaming without a 240W power supply. This wasn’t immediately clear to me during the purchase process. Since the official Framework 240W charger has a long lead time (Q2), I looked for alternatives based on community reports.

I purchased a Ugreen 500W charger (which features a dedicated 240W port). It arrived today, but after only 2 minutes of use, the power LED on the laptop disappeared. Now, the USB-C ports 1+2 on my Framework 16 are no longer working. I’ve tested them with other devices, and they seem completely dead.

My Frustrations:

Stability: For a laptop at this price point, I expected much better stability and clearer communication regarding power requirements for the dGPU.

Compatibility: It’s frustrating that the official charger is unavailable for months, forcing users to look for third-party solutions that seemingly risk frying the mainboard.

Efficiency: Seeing how unstable the power draw is, it feels like the current x86 implementation for this modular setup isn’t where it needs to be for gamers.

The Dilemma:

I likely need to contact support for a RMA, but I don’t have a spare device, so being without this laptop would be a disaster for my daily work.

Has anyone else experienced port failure when using high-wattage PD 3.1 chargers? Is there any “soft reset” for the ports/controller I should try before sending it back?

Thanks for any help.

I have a similar port power issue (using the official FW 240W charger), and it appears to be related to the overcurrent protection (a tricky thing to solve, and is part of the reason it’s taken so long for usb-c powered devices that pull this much juice to come out).

You should be able to get your ports working by doing one of the two methods of resetting the EC:

  1. Reboot into BIOS, have the AC unplugged. Click the “disconnect battery” option, then save and exit. Wait a second or two for it to fully shut down, then plug it back in and power up. This has worked for me every time I’ve tried it. The battery disconnect forces an EC reset.
  2. Shut down the device and unplug. Wait 120 seconds. EC should be reset. Plug the device back in and turn on, port should be working.

As a batch 1 member, I also currently share your sentiment of frustration. At this time, I don’t think I could recommend the FW 16 to non-enthusiasts. Between various thermal, power delivery, and mechanical issues I don’t think the average user could handle it.

2 Likes

Thanks a lot for the suggestion! I disconnected the battery via BIOS and it actually worked! My ports are back to life.

I’m still a bit wary about the power delivery stability with the dGPU and third-party chargers, but I’m relieved that the hardware isn’t fried. I’ll be extra careful until I can get my hands on the official 240W brick. Really appreciate the quick help!

I don’t have experience with the Ugreen 500w dock, but:

  • What BIOS version are you on? For me (Delta 240W brick and Ryzen 7840HS) both BIOS 4.02 and 4.03 solved power stability issues with both the 7700S and 5070. Every other version had issues with third-party chargers, only the OG 180W Framework brick I got in 2024 was working well. BIOS 4.02 and 4.03 solved stuttering with lower power bricks, like my Baseus 140W. The dGPU should be smart enough to pull what it can reliably, this is my case.
  • What power settings are you on? Make sure to try both Balanced and Max Performance, Balanced should never discharge the battery and Max Performance should draw a bit from the battery if the charger is not enough. Although, I used my FW16 for gaming with 180W for more than a year. Using ~75 out of 100W on the dGPU hurts performance just a tiny bit, <10% on my experience, so 240W charger are better but not mandatory.

I’m on BIOS 4.03 and updated the keyboard and macropad firmware as per the manual before I started.

I’ve made some progress: I was experiencing heavy battery drain even on ‘Balanced’ because Windows/Adrenalin kept prioritizing the dGPU over the APU. I managed to fix the stuttering in that scenario by forcing the APU for certain tasks and enabling Radeon Chill (capped at 60 FPS).

However, I’m still struggling with performance. Even on the ‘Max Performance’ profile using the original 180W Framework charger, I’m still getting stutters during gaming. It seems like the power delivery/throttling balance isn’t quite there yet for the FW16 when pushed.

Has anyone found a sweet spot in the Adrenalin settings or Windows power plans to eliminate stutters while using the 180W brick?

You may be interested in my other thread: Blocking GPU fans to increase CPU airflow - Framework Laptop 16 - Framework Community

I get stuttering in some heavy games like BF6 due to CPU thermal throttling. Those plugs eliminated the issue for me.

1 Like

Thanks a lot, I really appreciate the help and the insights! I’m going to take a closer look at those settings and suggestions later when I’m back at my laptop. Fingers crossed that I can get it running smoothly now.

1 Like

I think there must be something wrong.

As I said, I had no issues playing with 180W only for over a year. No stuttering, only battery drainage when in full performance.

It really seems to depend on the game’s demands. For example, when playing Factorio, everything runs perfectly fine. However, in a more demanding title like Diablo 4, the experience becomes quite stuttery—especially when there are a lot of mobs on screen—even with a configuration that avoids battery drain entirely.

I’ve noticed that if I tweak the AMD Adrenalin settings to reduce the stuttering in Diablo 4, the battery starts to drain slowly. If I switch to ‘High Performance’ to get the best frame rates, the battery drains quite rapidly. It’s a tricky balancing act between performance and power draw, and I’m still trying to find that ‘sweet spot’ where the game is playable without sacrificing the battery or hitting thermal limits.

Since Diablo 4 is notoriously CPU-heavy, I suspect that the heat soak/thermal congestion mentioned earlier by @TechPriestNhyk is exactly what I’m experiencing. The system just can’t dissipate the heat fast enough when both the CPU and GPU are pushed, leading to those performance dips.

This is why i decided to go with a 240W PD 3.1 Charger.

On a positive note, I don’t experience any issues at all during my daily work. When it comes to coding, the Framework really shines, and the performance is excellent—especially for my dual-boot setup. I’m honestly very glad to have moved away from NVIDIA; it’s such a relief to be rid of the headaches that came with using NVIDIA GPUs under Wayland and Linux.

1 Like

Regarding the CPU thermal soaking diagnosis, try to do a Cinebench run. If you score sub-average, could be a liquid metal/PTM issue. If your score is good in a repeatable way, thermals aren’t an issue.

You can check here for many, many results on both 7840HS and 7940HS in Cinebench.

Also, check out Framework Control Center, an user-made service+GUI lightweight tool to manage and monitor TDP, temps and noise. With that you can see how much watts your laptop agrees to pull with/from the charger and can set fans to run 100% (very noisy but should help diagnose thermal problems).

I also recommend you DDUing the drivers away completely and reinstalling the driver pack.

You can also reinstall the BIOS running the .exe again… It helped some users in old releases (can’t find the posts right now, it was around the 3.06 release IIRC).

I used to run Helldivers 2, AlE IV and VR games like Pavlov VR or VTOL VR with the 7700S with both 180W and 240W chargers without stuttering or issues. Don’t surrender, when you diagnose the issue and games are smooth you’ll thank yourself :slight_smile:

1 Like

Thanks for the suggestion! You were spot on.

​I actually did a clean Windows reinstall when the GPU arrived to be sure, and I think I found my mistake. In my frustration, I had been experimenting with different driver packs. This time, I stayed with the original AMD drivers instead of using the framework packs.

​I’ll also check out the tool you mentioned – thanks for the tip!

1 Like

Ah yes. The non-OEM drivers have always been a source of headache for me in the past, especially on laptops where devs have finer control over monitor passthrough and TDP tables.

Just FYI (to 5070 users) I’m still on 581.80. If it works don’t touch it.

I’m very happy that my hours wasted on troubleshooting were helpful to someone else… Enjoy!

:slight_smile:

(PS: give Radeon Super Resolution and AMD Smooth Motion a try… they really help getting to 120-165fps to really take advantage of the monitor!)

1 Like

Thanks for the driver tip! I’m currently sticking to the ‘don’t touch it if it works’ rule as well. GPU performance is solid now, and I finally managed to fix my stuttering issues.

However, the temperature of the CPU is still very uneven. To put it into perspective: my Cinebench R23 multi-core score is only 13,572. This is significantly lower than the expected performance, confirming that the uneven thermals are causing the CPU to throttle hard. I’ve started reading through the ‘uneven thermals’ thread, but honestly, it’s quite overwhelming at first glance.

I’m currently torn on how to proceed:

  • RMA or Mod? Should I even bother requesting a new mainboard, or should I just take matters into my own hands and replace the liquid metal with PTM7950 directly?

  • Airflow Hack: I’ll probably try the airflow mod first to see if I can get the temps under control that way.

To be honest, I’m not happy about this. In my opinion, a company like Framework should deliver a reliable cooling solution as a baseline. We should be able to optimize on top of a working system, not fix a broken one. I haven’t had thermal issues like this since 2003…

It’s also frustrating that these issues aren’t transparently discussed before buying; you really have to dig through dozens of forum threads to find out. I only started looking here once my expansion ports acted up. Now I’m looking at another 3-4 days of troubleshooting and modding just to fix ‘the basics’, which is a massive waste of time.

I’ll give RSR and Smooth Motion a shot though, once I’m sure the chip won’t melt! :slight_smile:

Update: I just tried the airflow hack, but it didn’t help at all. My Cinebench score was 13,668 with the mod, which is still well below where it should be. It’s clear now that the LM application is the primary bottleneck. Improved airflow provides no real advantage if the heat transfer from the die to the heatsink is fundamentally compromised. I definitely need to fix the LM situation (or swap to PTM7950) before any external cooling mods can actually have an impact.

You should get significantly better CPU thermal regulation by replacing LM with the Honeywell PTM.

How long have you had your machine? There was a nice web form you could fill out to get sent a PTM pad to swap in for the LM if you had a machine that had the LM in it, but the form was only open for … I think a few months. It’s possible you can still submit a support request and get FW to send you a PTM pad free of charge.

I’ve been using my machine for about 9 months now, and until two weeks ago, I used it strictly for work on Linux. It was perfect—honestly much smoother than my old setup. I was so happy with it that I decided to replace my desktop PC (and its RTX 2080) entirely with the Framework. I wanted to move away from Nvidia anyway, so I ordered the dGPU, and that’s when the trouble started. I definitely should have spent more time in the forums before making that decision, but I’m learning the hard way now!

Textbook LM issue behavior.

Get some PTM and replace it if you feel comfortable doing so. It’s easy with a hairdryer, isopropyl alcohol and the framework tool. I’d say it’s 7/10 difficulty, if you’ve worked on computers before shouldn’t be a problem. Plenty of suggestions on the Uneven CPU Thermals thread as well. LM is solid at room temp so it won’t spill, but be careful not to toss any piece around the MB.

In my personal opinion, when I get a problem I know how to fix safely RMA is wasteful and overkill, but that’s me, it’s well within your rights to expect a working product. :slight_smile:

3 Likes

I’m reaching out again because, unfortunately, the issues haven’t stopped. After just a few days, the charging problems returned. To be honest, this is becoming quite overwhelming.
I’m dealing with the exact same issues described here, which has made the experience quite frustrating over the last few days.

As an owner of a 1st Gen Framework 16, I’m at a point where I don’t know what to do next. I have no desire to become a professional repair technician just to keep my daily driver running, and I am not in the mood for a long RMA process. On top of that, I’ve already had to spend a significant amount of extra money on a new charger, cables, USB testers, and other parts just to troubleshoot issues that shouldn’t exist out of the box.

I am currently waiting for a delivery of PTM7950 arriving tomorrow to see if it improves the situation, and I will likely use it to replace the Liquid Metal (LM) entirely. It’s incredibly frustrating that as a customer, I’m the one investing both time and additional funds just to make a premium laptop functional.

I feel that much of the creator coverage focused heavily on the benefits of modularity while underplaying the potential downsides. I am now seriously considering whether I should return the device and switch to a non-modular laptop. I just want to code and relax while gaming, but this experience has been anything but smooth. I hope this post helps others understand what to expect before they buy.

I’m now seriously considering whether I should return the device (if still possible) and switch back to a non-modular laptop. I love the concept of Framework, but I need stability for my work and downtime. I’m sharing this update so that others considering a purchase are aware of the potential challenges involved.

I think this response has always been my concern with the FM16 and how people are pitched it in a lot of media.

I love fiddling with my machines, even without modular designs, I’ve always been taking apart my laptops, pasting, repasting, tweaking heatsink pressure, modding case openings, etc. to see if I could get any minute performance/heat improvements. As a result, I often bought many a secondhand laptop specifically because I was aware the risks inherent in my personal use. I’ve had used secondhand machines that’ve been rock solid even with mods, but I’ve also encountered brand new machines that’ve been glitchy and power limited/thermal throttled out of the box.

As a first gen machine, I didn’t expect the 16 to be the most stable or reliable. Hell, I’ve had Alienwares and Clevos that’ve been in production for multiple generations, still suffering thermal BSOD or BIOS/EC power limitations, and these were name-brand, prebuilt machines, with a few skus they should’ve troubleshooted from the factory. I had an Asus x13 that hard power throttled (CPU dropped to unusable freq) whenever I plugged in a third party USB C charger, despite being advertised as having USB PD charging (this included plugging into powered docks). You had to buy a proprietary Asus charger to have full use of a marketed feature. After 2 bios updates and no fix in sight I sold that machine.

The FM with it’s modular DIY capacity is just that many more points of failure. They not only have to deal with factory hardware, they need to deal with customers bringing their own parts to the table, self assembly on top of manufacturing bugs, design bugs, they need to troubleshoot users bring their own OS’s be it windows or linux, etc. etc. all of this long often after the product ships, because they’ve (so far) maintained board compatibility inside chassis designs.

All Dell needs to do is hopefully send out devices with the most recent drivers(something they still fail a lot) all design and assembly happened long ago and won’t change for the generation.

I think you can see how this leads to compounding sources of problems in a FM.

My one positive I’ll throw in for the FM is that as a modular design, it also has the possibility of fixes in the future. My now dead clevo had no such chance, same with the pressure cooker AW15.

If you buy a clunker Dell or Lenovo, it’s either the OEM fixes all issues by a year into the product cycle, or you’re probably stuck with them forever. With a FM, it’s possible years later, you’ll still have active hardware and ‘hopefully’ software support.

1 Like

Hi everyone,

I wanted to follow up on my previous post. After a few days of tinkering and a lot of trial and error, I have some positive news to share. Today, my PTM 7950 arrived, and after a marathon 6-hour session of assembly and disassembly, I finally have my Framework 16 running exactly how I wanted it.

Here is what I did to turn things around:

  • Thermal Overhaul: I decided to replace the stock Liquid Metal (LM) with PTM 7950, following this guide. Before flipping the laptop over to let it cool, I used the Framework Control Center (thx @TheLPeink) to set the fans manually to 5% and ran Cinebench until all cores were at about 80°C. This made the process much easier, and the heatsink came off quite smoothly using just the back of the Framework screwdriver. It took some patience, but it was worth it.

  • Clean Driver Setup: I manually installed all drivers from the Framework bundle (extracted the .exe) to ensure a perfectly clean baseline.

  • Catching up with AMD: My last Radeon card was actually an ATI Radeon 8000—back then, it gave me nothing but trouble with blurry textures in Counter-Strike! I’ve been on Team Green (Nvidia) ever since. To bridge that massive gap in knowledge, I had to disappear into the darkest corners of the internet to catch up. That’s where I discovered the Radeon Software Slimmer, which allowed me to strip away the bloat and install only what I actually need.

    • If anyone is interested, I can drop a list of packages and services which are safe to debloat in the AMD Radeon drivers.
  • The Result: The CPU is finally delivering significant performance with stable temperatures and proper wattage. Interestingly, fixing the thermals also automatically resolved the issues I had with the charging/discharging state switching.

Gaming Performance (Diablo 4): Before these tweaks, I was getting a very unstable 120 FPS. Now, I’m seeing a solid 160–180 FPS on medium settings.

Benchmarks
1st Cycle with PTM 7950:

Stock LM:

Future Potential: These are the results from the very first cycle. According to sources like Igor’s Lab, PTM 7950 needs about 10 thermal cycles to fully “burn in” (meaning the phase-change material settles and reaches maximum conductivity). I’m confident that my Cinebench scores will improve even further once that’s done. I’m also keeping an eye on the “Airflow Hack” mentioned by @TechPriestNhyk for later—but for now, I’m just happy it works!

Final Thoughts: I’ll be honest—I was incredibly frustrated at first. I didn’t plan on spending so much time on “the basics.” But in hindsight, diving this deep into the hardware actually made me appreciate the device much more. I’m honestly a bit proud that my “rusty” technical knowledge was still solid enough to get everything sorted.

Anyway, I’m just glad I can finally stop staring at hardware monitors and actually start using the laptop. It was a steep learning curve, but seeing the results now, it was worth the gray hair it probably gave me! :smiley:

Lastly, I want to thank everyone here who commented on my initial post. Your feedback and encouragement really kept me going and pushed me to find a solution instead of just giving up on the machine. Cheers!

2 Likes

Your persistence paid off. Good work!!

1 Like