We build and thoroughly test on the order of a couple hundred laptops during design, development, and manufacturing bring-up before we start shipping out to customers, but it’s inevitable as we ramp up production that we find new issues after systems land in your hands. There is a standard process in consumer electronics for this called “Early Field Failure Analysis” or EFFA. With this, we take back every system or module we can that has an issue or failure early-on to determine both why it had an issue and why the issue wasn’t caught in the testing that every module and system goes through. In most consumer electronics companies, all of this happens invisibly to the end user, but we feel it’s better to be transparent and let everyone know the issues and fixes as we find them. We’ll continue to update this thread as we go.
Known issues on early Framework Laptops
Touchpad rubbing when clicked - Very early in production of Batch 1, we found that the active alignment fixture that optically centers the Touchpad glass on the Touchpad PCB wasn’t tuned well, allowing the glass to be slightly off-center on some number of units. We quickly added an inspection step to ensure that Touchpads would get rejected if there was insufficient gap along the bottom edge. We’ve gotten a field report of a user seeing this issue on a unit from before we added the extra check, and have supplied a replacement Input Cover. If you are seeing rubbing on your non-Batch 1 unit, try the instructions in the Touchpad Rubbing Fix Guide.
Touchpad physical click not registering - We’re actively working on finding the root cause of this issue, but we believe a small number of Touchpads have an assembly issue on the tactile dome switch that results in it having high electrical resistance, preventing clicks from registering. Pressing the bottom center of the touchpad firmly a few times has resolved the issue in some cases.
Touchpad Cable can short if misaligned - Noticing a pattern on which module is having the most issues? We found a production error on the PCB for the Touchpad on Batch 1 units and a subset of Batch 2 that could result in the Touchpad Cable shorting if it is misaligned in a specific way when powered on. We’ve updated the guides to ask everyone to double-check that the cable is aligned before closing their systems, we’ve created a guide on how to add tape to ensure the cable can’t get misaligned, and we’re also adding this tape on new system production until we can roll in a revised PCB.
High CPU temperatures on a very small number of units - We haven’t seen any customer units with this issue yet, but a couple of press reviews noted high temperatures when running benchmarks. We’re root causing the issue, but we believe it could be related to thermal paste pump-out. We’ve switched thermal paste formulations to one with higher pump-out resistance in the meantime.
An Expansion Card bay isn’t working - Incredibly enough, we found that the non-conductive EMI shielding we use on the receptacles can become conductive in some situations. If you find that one of your bays isn’t working, you can adjust the internal shielding sticker to resolve it. There are instructions here.
System not booting, with S4/S0 blink code - We’re still trying to root cause this, but we have seen a few instances of units either DOA or failing shortly after starting use, with a symptom of not booting and only blinking out an error code on the side LEDs indicating S4 and S0 issues. We’ve found a workaround for this that we’ve added to the bottom of the article here.
System stuck at 200 MHz or 400 MHz - We’ve root caused some, but maybe not all instances of this to the EMI sticker issue. Until we can fully confirm if that is the root cause though, we want to note that some users have seen occasional instances of the CPU frequency dropping and returning to normal without clear indication of what behavior causes it.
Thank you for a roundup of all issues, these all look to be normal for batches 1 and 2 considered that they are the 1st and 2nd batches so manufacturing defects can occur.
This is awesome glad to see everything that’s being fixed. Gives me even more confidence in the laptop overall and I’m happy that I’m in batch 3 since it looks like a good amount of little bugs will be fixed in that batch. Really looking forward to getting my laptop.
Thank you for the advice on the touchpad physical click! I just did a firm press in the bottom middle of the touchpad and heard the faintest of clicks after the normal click.
Everything seems to be working fine now with the physical buttons - before I was unable to get any sort of interactivity from them regardless of the BIOS setting or other debugging information already available (running Fedora 34 so I figured it was maybe just an issue along those lines).
@nrp Re: High CPU temperatures on a very small number of units:
Has your team considered graphite pads as a CPU thermal interface material?
I’ve replaced the thermal paste in my Macbooks with these and had great success.
They’re reusable, “clean” to handle, perform within only a few Kelvin of a fresh high-end paste application and best of all they do not suffer from pump-out or degradation effects at all, which in my eyes makes them superior in long term/high temperature applications like laptops.
I suppose in theory these could even be taped to the heatsink with small strips of kapton tape from the factory, making a potential motherboard upgrade clean and easy.
I think this post shows a really great attitude that makes a difference and makes the Framework unique. A value of buying the Framework Laptop is not only to have the laptop itself, but also to experience these kind of great things. Thank you.
Thanks for sharing. What I think is even more impressive is your willingness to make the issues right in cases where it is possible. The very design of the laptop insulates against these very issues. Most excellent! I should be getting mine this week. Just need to pick it up from the friend who took delivery for me.
I’m a decision maker at a large technology company. Once I get mine and do the run through, I expect I’ll be recommending these for our fleet computers.
The upgradability makes huge financial sense in the long run and so does the repairability.
That’s a very nice idea! This will further increase reliability and the commitment for right to repair, may be a lot more expensive than using thermal paste tho
Graphite thermal pads would make temps worse though. They’re OK, but they’re really just ideal if you’re swapping coolers a lot (like with a test bench). Applying thermal paste isn’t that hard honestly, so if they put a QR coded guide on there you should be able to do it just fine. It sounds like it was a manufacturing issue on their end so it should be fine moving forward
Re: clicks not registering
I see that happen often. Sometimes 4-5 clicks before it triggers.
Re: temperature
I’ve been noticing mine running at around 36-50°C the last couple days; while having fans blowing onto the unit. Getting a lap desk that will draw air out to see if that will help.
This is something that laptop users need to come to terms with. I’m not targeting you specifically mind you, but temps aren’t as scary as people think they are.
Intel designed this chip to be safe up to 100°C. Would I ever let my laptop hit that high? Oh definitely not. Temps up to 80 though are honestly fine unless it’s throttling and you’re doing some heavy lifting while plugged into the wall. So I wouldn’t worry about the fans and all of that unless it’s specifically affecting your workflow.
Changing out the thermal paste can also help but be careful. Don’t be tempted by liquid metal (it’s kinda sketchy in a laptop imo), and high performance paste like kryonaut can actually be detrimental in a laptop. Yes, a high performance paste will let it run cooler at the DIE, but it’s still pumping out heat. I regret repasting my laptop because it runs too well now. Dang CPU runs at full tilt all the time and the chassis gets unbearably hot.
Anyways, TL;DR:
Temps under 80 are 100% fine if you’re boosting properly
Thermal paste can be a good tool, but too good and it’ll make the chassis too hot
I repasted a Huawei Matebook D with a ryzen 2500u. It’s a metal chassis laptop and it really gets warm under load, and that’s with a ~18w part (i think that’s what huawei configured it as? It goes up to 25w depending on vendor). An Intel device boosting to 60w is going to be a lapburner if you’re not careful.
Just a word of caution from someone who got too zealous on the “better paste, lower temps, more better” train.
