Basically, the way DIMMs are built with edge connectors causes trouble with signal integrity. That limits the maximum speed and is one reason why LPDDR is soldered.
CAMMs replace the edge connector with a grid of contacts like on modern CPUs. That’s easier on the signal, allowing for faster speeds and even LPDDR on modules. Downside: They need to lie flat on the mainboard and thus use more space. Then again, laptop RAM is typically mounted in a flat configuration already so it’s mostly a new challenge on the desktop.
In case you’re wondering why they announced CAMM2 and where CAMM1 went: The original CAMM was a proprietary module from Dell; their spec was refined into the JEDEC-standardized CAMM2.
Ah, I did note the grounding bad around the edge but hadn’t thought about where the actual contacts were. Ironic, considering I deal with high speed signaling professionally lol.
Thanks for the info, let’s see where the industry goes…
Basically, the way DIMMs are built with edge connectors causes trouble with signal integrity. That limits the maximum speed and is one reason why LPDDR is soldered.
CAMMs replace the edge connector with a grid of contacts like on modern CPUs. That’s easier on the signal, allowing for faster speeds and even LPDDR on modules. Downside: They need to lie flat on the mainboard and thus use more space. Then again, laptop RAM is typically mounted in a flat configuration already so it’s mostly a new challenge on the desktop.
In case you’re wondering why they announced CAMM2 and where CAMM1 went: The original CAMM was a proprietary module from Dell; their spec was refined into the JEDEC-standardized CAMM2.
Ah, I did note the grounding bad around the edge but hadn’t thought about where the actual contacts were. Ironic, considering I deal with high speed signaling professionally lol.
Thanks for the info, let’s see where the industry goes…