The race to make laptops thinner is pointless. I wish I could buy a thicker laptop that is more robust, repairable, expandable, and has better cooling, with the extra space filled with batteries.
I'm stunned that "mapping out" blocks of bad RAM is a thing you can do. It reminds me of the old days when hard disks had to be "low-level formatted" first (ask your parents, kids) and it would ask you for a list of bad sectors to map out. Sometimes hard disks would even come with a list of known bad ones printed on the top.
We used to buy Dells with minimum RAM and add quality third party RAM for way less than Dell wanted for it. They also use ram quantity to leverage other specs we're not bothered about. Of course they are moving to soldered.
the fact no component supplier, not even alibaba sells LPCAMM2 sockets tells you everything you need to know. The only prices I can find are in $30 range per socket
Connectors are expensive, high-speed ones particularly. $30 is fine, though it makes a giant difference at what quantity that is. Did you look at a single one or 1k?
I've been eyeing LPCAMM2 interposers for non-RAM uses - as an off the shelf connector for stacked PCB designs. The conclusion was that it wasn't produced or used at sufficient scale for it to work.
Yes, of course. It says - ball grid (array) in the article. Doing it by hand, e.g. with just hot air is proper hard. It's not hard with specialized equipment (that cost way more than the laptop, itself). Sourcing the chips is not easy either.
As for the videos and the Macs, you can see the same people (likely) soldering RAM as well.
It is - anything short of HBM can be replaced that way.
The issue, of course, is just how much time, effort and skill is required to do it.
It might become a viable commercial process outside third world countries if RAM prices spike even more - or if someone finds a way to teach an off the shelf $1000 robot arm advanced autonomous microsoldering.
Then there's the problem of forcing the device itself to recognize the new RAM. It's a device-specific trick. RAM sticks have a little tag chip that tells a PC "how much RAM this stick has", but RAM chips themselves don't - so the device has to infer its RAM capacity from elsewhere. Like strap resistors on the board. Which you have to find and adjust for the new RAM capacity. That, or fix the RAM values in firmware, which is also device-specific.
> RAM sticks have a little tag chip that tells a PC "how much RAM this stick has", but RAM chips themselves don't - so the device has to infer its RAM capacity from elsewhere
I'm surprised that in 2026 this is still true.
You could easily fit a little microcontroller on the die of each RAM chip which reports its make/model/manufacturing date, capacity and capabilities, can be triggered to run self-tests, etc. If they aren't doing that yet, I'm surprised that they aren't.
If you live in Shenzhen, it's a subway trip away. Good repair services elsewhere can do it too. Either way any warranty you have is probably getting called into question after that, even if that's supposed to be illegal in some places (e.g. the US - they're supposed to prove that your modification/repair is what caused the warranty fault, but in practiceā¦).
Bash me all you want, but after CPU, GPU, memory controller & a host of other things integrated in modern System-on-Chips (SoC), I've long held the belief that next logical step would be to integrate RAM.
Optimal: piecemeal. As in: X sockets on a board, each socket takes a chip that adds CPU core(s), GPU cores, and some amount of main RAM. More chips, more powerful computer.
Keyword: socket. Read: replaceable. Bad chip out, good chip in. Slow chip out, faster chip in. Soldered onto board = reliable, but a defect could junk the whole machine.
Manufacturing processes for RAM and processor cores are not one and the same and use different tradeoffs. It's the main problem in-memory processing is running into, which is akin to what you're describing here. The die space required for the same level of performance makes the idea of on-die RAM intractible. The best they're able to do with RAM optimised manufacturing processes are tiny RISC-V cores that lack most of the feature sets. What apple does with moving the RAM chips on to the package really is the only viable option.
> Bash me all you want, but after CPU, GPU, memory controller & a host of other things integrated in modern System-on-Chips (SoC), I've long held the belief that next logical step would be to integrate RAM
Haven't Apple (A-series, M-series) and others been copackaging memory for a while? (SiP or MCM?)
Am I the only one who wonders if there is a way to get the kind of performance we want out of non-soldered components/SOC designs? Like from what I've heard, it's a "physics says no" problem, but I'm not sure if that's actually true (not being an expert in this technology) or if it's just the excuse many companies use to justify soldering everything to make serviceability impossible. And it would be really neat if this is actually being looked into by labs right now (possibly with AI's help), since areas like that are where I think (non) LLM AIs might be very useful. Same thing for medicine, materials, etc.
Effectively there are no socketed CPUs for laptops, though, for way more than a decade already. Saying that as I'd fully support modular components, as in desktops.
While it's true that you can get better performance out of soldered memory than a socket, that's also not why vendors use it. Desktop computers, with their very much not soldered memory, get great performance after all. The laptop vendors solder the memory because it lets them make more money off selling units with more memory soldered in. That's all.
"Smartphone makers glue them because it's cheaper and they can make more money ..." No. It's what the consumers want. Thinner phones, water resistant phones, and thinner laptops.
Since the OP mentions Lenovo, look at their other product lines for ones that still have socketed RAM. I'm currently running an E16 with 40GB RAM, 8GB soldered, 32GB socketed. And aside from that it leaves my previous X1 for dead in terms of performance. Lenovo's replacing the carbon fiber case, the whole reason the X1 line is called "Carbon", with crappy black plastic was the last straw.
The race to make laptops thinner is pointless. I wish I could buy a thicker laptop that is more robust, repairable, expandable, and has better cooling, with the extra space filled with batteries.
I'm stunned that "mapping out" blocks of bad RAM is a thing you can do. It reminds me of the old days when hard disks had to be "low-level formatted" first (ask your parents, kids) and it would ask you for a list of bad sectors to map out. Sometimes hard disks would even come with a list of known bad ones printed on the top.
We used to buy Dells with minimum RAM and add quality third party RAM for way less than Dell wanted for it. They also use ram quantity to leverage other specs we're not bothered about. Of course they are moving to soldered.
Prebuilt branded PCs have always had inferior quality, compatibility, and expandability problems, all with overpricing.
Like the exact case with phone batteries that are not easily changable by the owner, soldered rams are evil...
LPCAMM2 sockets were supposed to provide all the benefits of soldering while still being a socket. Anyone know the state of adoption?
You can't buy a single LPCAMM2 RAM module on Amazon. eBay has some.
The new Framework 13 Pro laptop ships with LPCAMM2.
https://frame.work/blog/framework-laptop-13-pro-deep-dive-lp...
But no ECC support.
yeah but there is very little interest in that model compared to ai max line
Lenovo uses it on their latest thinkpad t14.
the fact no component supplier, not even alibaba sells LPCAMM2 sockets tells you everything you need to know. The only prices I can find are in $30 range per socket
Connectors are expensive, high-speed ones particularly. $30 is fine, though it makes a giant difference at what quantity that is. Did you look at a single one or 1k?
Honestly, reasonable.
I've been eyeing LPCAMM2 interposers for non-RAM uses - as an off the shelf connector for stacked PCB designs. The conclusion was that it wasn't produced or used at sufficient scale for it to work.
All of the new NVIDIA data center GPUs use SOCAMM2, quite similar to LPCAMM2.
https://www.koreatimes.co.kr/business/tech-science/20260420/...
> I then had to manually transcribe this, which [w]as āfunā.
Can't you select and copy the text in the photo (on most smartphones and PCs at least)?
It it possible to desolder and replace the offending RAM if one were sufficiently motivated? I've seen folks doing that to soldered SSDs in MacBooks
Yes, of course. It says - ball grid (array) in the article. Doing it by hand, e.g. with just hot air is proper hard. It's not hard with specialized equipment (that cost way more than the laptop, itself). Sourcing the chips is not easy either.
As for the videos and the Macs, you can see the same people (likely) soldering RAM as well.
It is - anything short of HBM can be replaced that way.
The issue, of course, is just how much time, effort and skill is required to do it.
It might become a viable commercial process outside third world countries if RAM prices spike even more - or if someone finds a way to teach an off the shelf $1000 robot arm advanced autonomous microsoldering.
Then there's the problem of forcing the device itself to recognize the new RAM. It's a device-specific trick. RAM sticks have a little tag chip that tells a PC "how much RAM this stick has", but RAM chips themselves don't - so the device has to infer its RAM capacity from elsewhere. Like strap resistors on the board. Which you have to find and adjust for the new RAM capacity. That, or fix the RAM values in firmware, which is also device-specific.
> RAM sticks have a little tag chip that tells a PC "how much RAM this stick has", but RAM chips themselves don't - so the device has to infer its RAM capacity from elsewhere
I'm surprised that in 2026 this is still true.
You could easily fit a little microcontroller on the die of each RAM chip which reports its make/model/manufacturing date, capacity and capabilities, can be triggered to run self-tests, etc. If they aren't doing that yet, I'm surprised that they aren't.
Why bother with the expense if you are permanently soldering it in anyway?
If you live in Shenzhen, it's a subway trip away. Good repair services elsewhere can do it too. Either way any warranty you have is probably getting called into question after that, even if that's supposed to be illegal in some places (e.g. the US - they're supposed to prove that your modification/repair is what caused the warranty fault, but in practiceā¦).
On YT thereās dosdude soldering ram on M series SoCs
Good BGA experience and a steady hand seems required
Of course it is possible. Unless you're dealing with manufacturer who actively works against that, like Raspberry Pi.
Bash me all you want, but after CPU, GPU, memory controller & a host of other things integrated in modern System-on-Chips (SoC), I've long held the belief that next logical step would be to integrate RAM.
Optimal: piecemeal. As in: X sockets on a board, each socket takes a chip that adds CPU core(s), GPU cores, and some amount of main RAM. More chips, more powerful computer.
Keyword: socket. Read: replaceable. Bad chip out, good chip in. Slow chip out, faster chip in. Soldered onto board = reliable, but a defect could junk the whole machine.
RAM uses a different process than the logic nodes you describe. RAM uses tons of capacitors, along with the the transistors.
Then there is the HBM which is an expensive process to mount on the chip.
Manufacturing processes for RAM and processor cores are not one and the same and use different tradeoffs. It's the main problem in-memory processing is running into, which is akin to what you're describing here. The die space required for the same level of performance makes the idea of on-die RAM intractible. The best they're able to do with RAM optimised manufacturing processes are tiny RISC-V cores that lack most of the feature sets. What apple does with moving the RAM chips on to the package really is the only viable option.
I'm using laptop with Ultra 7 268V. It already has RAM on chip. I think they call it "processor package". I don't think it's socketed, though.
> Bash me all you want, but after CPU, GPU, memory controller & a host of other things integrated in modern System-on-Chips (SoC), I've long held the belief that next logical step would be to integrate RAM
Haven't Apple (A-series, M-series) and others been copackaging memory for a while? (SiP or MCM?)
Those aren't socketed though, are they?
Am I the only one who wonders if there is a way to get the kind of performance we want out of non-soldered components/SOC designs? Like from what I've heard, it's a "physics says no" problem, but I'm not sure if that's actually true (not being an expert in this technology) or if it's just the excuse many companies use to justify soldering everything to make serviceability impossible. And it would be really neat if this is actually being looked into by labs right now (possibly with AI's help), since areas like that are where I think (non) LLM AIs might be very useful. Same thing for medicine, materials, etc.
If the CPU can be socketed, there is no reason RAM can't be.
Effectively there are no socketed CPUs for laptops, though, for way more than a decade already. Saying that as I'd fully support modular components, as in desktops.
While it's true that you can get better performance out of soldered memory than a socket, that's also not why vendors use it. Desktop computers, with their very much not soldered memory, get great performance after all. The laptop vendors solder the memory because it lets them make more money off selling units with more memory soldered in. That's all.
It's partly a thickness problem. Sockets add height to boards.
Besides, since RAM isn't getting cheaper for a while, this isn't a big issue now.
Are any customers asking for a laptop that is 1 or 2mm thinner?
That's not all.
"Smartphone makers glue them because it's cheaper and they can make more money ..." No. It's what the consumers want. Thinner phones, water resistant phones, and thinner laptops.
Well that was good to learn!
Since the OP mentions Lenovo, look at their other product lines for ones that still have socketed RAM. I'm currently running an E16 with 40GB RAM, 8GB soldered, 32GB socketed. And aside from that it leaves my previous X1 for dead in terms of performance. Lenovo's replacing the carbon fiber case, the whole reason the X1 line is called "Carbon", with crappy black plastic was the last straw.
i had to move a very very small resistor in the correct place to make a laptop disregard the soldered ram and only use the sodimm