So, from reading all the messages in this thread. Everyone is happy now.
Many E0 memory can use either E0 or E1 interposer.
Some (some Dell and Lenovo) E0 memory must use the E0 interposer.
The E0 interposer is relatively cheap. $20.
The FW13 Pro comes with an E1 interposer.
I don’t think there are any remaining incompatibility reports, so long as the correct interposer is used for your specific memory chips.
Maybe FW could update their FW13 Pro memory compatibility table, with an extra column for which interposer (E0, LGA644 or E1 LGA666) is needed to make it compatible.
Me just waiting for a FW16 AMD with dual LPCAMM2 memory slots and ECC.
Aren’t CAMM2 already designed to be capable of having two channels per one module to essentially replace two SO-DIMM modules/sticks you used to have to use before?
I haven’t seen any mentions of dual LPCAMM2 being an option anywhere. Such a setup would require a CPU supporting four channels, which are currently only server-grade ones.
It isn’t but for platforms like strix halo that has more memory channels you would either need bigger lpcamm modules or two of them (which then come with the whole bag of issues with mixing memory sticks we just got away from with dimms XD)
Digging through the Framework subreddit and this post, it isn’t clear to me if the RAM I picked up secondhand would work natively or not. It is listed in that Google Doc as being untested/unconfirmed. It has center dots that I haven’t seen on other E0 modules, but unlike other E1s I’ve seen, the dots aren’t colored or recessed. I’d assume the dots would be serving the same purpose on my module. Anyone have any guidance or thoughts? Going to pick up an E0 interposer just to have one handy. I wish I had more confidence to cancel the 16 GB module I have configured in my pre-order now.
It is specifically the M561K2LC4DE0-CCTJS module. Front and back pics attached.
Looks like the official compatibility page got updated with unofficial data gathered by community, though the original Reddit post with the g.sheet wasn’t mentioned.
It seems like the incompatible list consists of every model that was incompatible on any platform in the spreadsheet; even Samsung M561K2LC4DE0-CCT is listed there, despite being reported as working on Ultra 5 325. I wonder if it’s just precaution.
Scouring the Framework subreddit, the only report I could find of Samsung memory working with the Ultra 5 325 was specifically M561K1LC4DM0-CCTGM. I can’t find any pictures of any of the CCT* modules to see which ones have the circular pads in the pour for those center rows and which ones are just solid copper. And I haven’t seen any postings of someone saying a specific module works in the Ultra 5 325 but doesn’t in the Ultra X7 or X9. According to the reasoning in this post and others as to why this workaround functions, it is all due to the E1 transposer pins getting shorted to ground because early E0 RAM modules don’t have those circular pads.
These are the pins located along the middle of the connector (see the aforementioned forum thread for some great photos showing this, or JESD318B). RFU pins are supposed to be safe to use for such new additions because modules and motherboards are supposed to leave them floating, JEDEC says of these pins:
Reserved for Future Use. Circular pads are required within the CAMM2 footprint to assure future RFU pins are not shorted to ground.
However, on some E0 modules, these pins are not left unconnected, instead, they don’t exist as independent pads, they are just part of the large ground pour that surrounds the existing pads.
My question is with the module in my post, are those circular pads actually independent pads? They don’t look like the circular pads in other pictures of modules that have them.
Yeah that looks like an E0 module (rectangular PMIC+separate buck converter) and individually isolated pads for the center rows. That should in theory work with an E1 interposer.
I suspect even with Crucial E0 modules with isolated center row pads, it’s better to use a E0 interposer.
I have the X7 mainboard with Crucial E0 module in an old 13 chassis, which works under normal conditions. However the old chassis do flex and if I pick the laptop up from lower left corner & press up from bottom cover at the LPCAMM location, the laptop would freeeze.
Tighting the 3 LPCAMM shield screws really hard fixes this problem.
Now that I see this post, I wonder if the problem is with the E1 interposer + E0 module combination. I suspect if you slightly warp the MB & the LPCAMM screws are not extremely tight, the E1 center row pins might touch the E0 module’s ground plane. And that would cause a short.
I’m the one with the CCTGM on the Ultra 5 325, and it does have the circular pads in the center. I can’t confirm whether it works on the X7 since my preorder for that is in batch 16, though.
EDIT: that is a photo of the CCTGS, (32GiB model), but the back of the CCTGM looks identical and even has the same identifier text. I have both, just got confused between the two letters.
