AMD FW13: erratic behavior with official 32GB refurb SODIMMs

Not really sure what happened here. Sometimes the laptop doesn’t boot at all, sometimes it boots fine, sometimes the 780M graphics don’t work, sometimes booting takes forever (not just upon initial memory training and not just on the Framework logo). There appears to be some strange fitment issues–the WiFi antenna wires can get in the way or hit it, which is surprising for an officially validated kit. I initially thought some kind of seating issue caused this but upon reseating ‘sometimes it doesn’t boot at all’ became ‘always’.

I previously had some trouble with this machine with my display, which had some very strange green corruption that manifested whenever the machine was under load (finally got this replaced, it was a saga to get this approved) but the RAM (the G.Skill 5600MHz 2x16 CAS40 kit which is on the validated list for AMD) was never an issue.

My instinct would be to test these in a different machine, but I don’t have any other machine that uses DDR5, much less DDR5 SODIMMs. Since they’re refurbs though I assumed that they will have been checked, so honestly, my thinking is that somehow, both times, one or both of the sticks didn’t get in there properly. But the serial failure to boot (fan ramping up and down, the power button LED, the screen might? have been on but there was no text or images, no other sign of life) scared me off trying anything else anymore without a concrete plan, and I went back to my (apparently!) trusty G.Skills for now.

Anybody experienced similar issues with the official 32GB sticks? Other ideas on how to approach this? My strong preference is to make sure that these work before concluding that they’re in some way defective, because my experience trying to get things RMAed is…well, I’ll just say that if you can avoid it, you should.

FWIW the OS is Win11 24H2, but I don’t think that’s terribly relevant.

Would it be possible for you to send any pictures of the issues you are experiencing?

There have been some reports of issues when running / updating to 24H2, so try seeing if these issues occur on another OS or older version of Windows if you can get one.

Hmm, I’m not sure pictures would add much, and would require me to try installing the new sticks again blindly in order to get. Let me go through what happened again, maybe we’ll get some clarity from there:

  • First boot was, as far as I noticed, fine (this only happened once). I’m not sure what happened here but notably the keyboard cover was off, and I didn’t intend to turn it on, but somehow hit the power button.
    ** After connecting an external keyboard (I was uncomfortable using the internal one while it was half connected like this) I managed to shutdown and close’r up.
  • After this, I got an incredibly slow boot, with the graphics card not being properly recognized and thus the resolution and scaling being out of whack, I’m assuming because it was limited to the fallback driver or couldn’t load the proper one fully. I did not attempt to do anything drastic in Device Manager (like updating or rolling back the driver) and unfortunately did not make a note of the error code there (there was a down-arrow, not an exclamation point, iirc). I did try starting the device but naturally, that didn’t work.
    ** At this point I decided I needed to reseat the memory. I figured that one of the sticks must have shifted upon installation of the keyboard cover, and the fact that it worked at all on the first attempt meant I got lucky.
  • After that, the previous symptoms continued a couple more times. I tried loading optimized defaults in the BIOS; this did not help but did not make things worse.
  • Eventually, even this stopped happening, and the machine failed to POST at all. Power would turn on (or at least the power LED around the power button will–I did not note whether the LED on the chassis turns on or not), the fans may ramp up and down, but nothing displays on the screen.
    ** After this happened three times, I gave up and went back to the old sticks.
    While there was no sluggishness in booting and general computing this time, the graphics were still not recognized properly until I removed the device via Device Manager and let it be detected again, at which point everything was fine.

So, I’m really not sure what happened here, but it might be multicausal, since some symptoms happened with both the old and new sticks (the graphics not being recognized properly) and some only with the new ones (slow booting and computing and eventually no booting at all)–perhaps some combination of bad seating compounded by some detection/configuration issue in Windows with respect to this particular SoC/iGPU that arises when RAM is replaced (wrong memory or device location or something)?

I have no ideas here other than reseating again, maybe trying one stick at a time, making sure not to accidentally power on in the process. At that point, if stable, I can run my usual battery of memory tests, if not…well if not, we’ll get to it when we get to it. Thoughts?

(Honestly, I need to get into the habit of taking photos and notes when doing this kind of thing, not just with laptops, because otherwise it’s easy to lose track of what happened when in response to what. Not very scientific…)

Just a quick update here. I’m not quite sure what’s going on but for some-odd reason, when I have 2x32GB installed, and the iGPU dedicated memory allocation is set to Auto in the BIOS, the above-mentioned symptoms occur–specifically, an initial white screen followed by the GPU driver not really working, scaling being incorrect, and fan running at all times. The allocation is also not set to the amount suggested by the BIOS help text.
My guess is that (a) clearing the BIOS and/or (b) reverting to the official drivers rather than AMD’s latest (no other issues to date) will help matters here, but since it works in Gaming mode (which I was going to use anyway) I just ignored this for the time being. That said I would like to make sure that Auto mode actually works properly, for completeness’ sake, so plan (a) to try booting into other OSes and seeing if the issue persists, (b) to DDU the current drivers and revert to the latest official release (which unfortunately is still woefully out of date), and if that doesn’t do it, (c) to reset the BIOS. With respect to (c), I’m assuming that fully resetting the BIOS involves disconnecting the battery for a time, since there is no separate CMOS battery on AMD and Intel 13th gen and on?
Any other ideas or things I should be aware of? FWIW, BIOS is already updated to current release.

I would start by reverting the drivers to Framework’s custom version. While it is older, you are guaranteed stability by Framework when using those drivers.

I would try updating all your software to the newest possible or newest provided by Framework (for drivers and BIOS). I have read reports of devices crashing and unable to update, but once they are updated they are fine. Maybe once you do that, you could try the auto mode.

Of course, there is also the possibility that the auto feature was just added in without much consideration causing the crashing issues. With your 64GB RAM, auto should just be defaulting to Gaming mode, so there isn’t really a benefit there though. I understand the urge to make sure that every part works, but if it is a BIOS issue, I’m afraid you’re out of luck.