A non-GPU Expansion Bay Shell and a Graphics Module interposer are the items needed.
The wrong interposer, which will not work with OcuLink, is this (fan-only) interposer. If you wish to compare & see the difference.
1 Like
Is the expansion bay shell not what I have with my Framework 16?
If thats the case then why is my interposer not working with my expansuion bat shell
There are 3 possible situations one can have. And for each, I’ll list the items needed to use the OcuLink module Josh is making.
1) Framework 16 with the Graphics Module (AMD Radeon RX 7700S) - Click to show
You brought your Framework 16 with this: Framework Laptop 16 Graphics Module (AMD Radeon RX 7700S)
You need a non-GPU Expansion Bay Shell.
You already have a Graphics Module interposer which you can use for the Oculink module.
2) Framework 16 without the Graphics Module (AMD Radeon RX 7700S) - Click to show
You brought your Framework 16 with this: a fan-only Expansion bay shell
You need a Graphics Module Interposer.
You can put the OcuLink board into the fan-only Expansion bay shell which you already have.
3) Framework 16 with the non-GPU (fan-only) module, AND in addition the Graphics Module (AMD Radeon RX 7700S) - Click to show
You would have needed to have selected both when buying your Framework 16, and paid an additional cost. Normally, only one, or the other, is included.
You have everything you need.
You can put the OcuLink board into your fan-only Expansion bay shell, and use the interposer from your graphics module, both of which you already have.
4 Likes
useless is a very extreme word for that.
Removing the DP input would mean an ~10% performance penalty of using the eGPU with the laptop’s integrated display (since it then has to route the eGPU output through the iGPU). Still a huge uplift over not having the eGPU and any external displays connected to the eGPU will still get the full performance.
I think you switched the first two options.
1 Like
Thanks. Corrected
The interposer is not working because you need the oculink board.
1 Like
Thank you so much for this excellent reply…so I need to buy the framework expansion shell with no gpu? Which is the one I bought? Or the one that came with my framework 16? Again, my framework 16 has the original expansion, no gpu expansion, so are you saying there is a third one I can get, which I need to get so that the interposer fits?
The one you have is the one you need.
The oculink board goes inside that shell, replacing the fan controller board, that is in there now. And than you will be able to fit the GPU interposer.
1 Like
Oh so I literally need to wait for the Josh cook part…got it thank you so much for clarifying friend
Any possibility of leaving the M.2 for people who don’t need the DP input port, and are just going to run directly off of the GPU? So those who want it can use the DP input, and those who don’t can get access to another M.2?
Yes this would be definitely preferable. I plan to use the eGPU only lid closed.
I just put in my pre-order for the oculink 8x board and I ordered the empty expansion shell from framework (wish I had realized that would be useful on my initial order)
I’m going to be moving between countries in a few months so I’m hoping I’ll be able to pull my 7900xtx out of my desktop for use as an egpu and avoid bringing my desktop along with me altogether. It’d save me having to drag a 48lb case full of delicate hardware on a flight overseas.
I’ve been reading up a lot on oculink 8i and it seems like the options so far are somewhat limited as far as the actual adapters for the gpu itself.
This seems like the only consistently available adapter, but there’s no enclosure specifically made for it and would require some kind of custom solution. (it has 2x 8i ports, but it appears to be wired in such a way one should work the same as having a 16x port only wired for 8x)
There’s theoretically a full enclosure with adapter (and maybe PSU) from “OSMETA” or “GZOSMETA” in China that supports oculink 8i and claims pcie gen 5, but it seems like it’s out of stock or unavailable everywhere I’ve looked so far. I’ve only found it referenced here:
Does anyone here know what adapter they plan to use once the oculink 8i port is available?
This is what I’d expect since China has been a major adopter of OcuLink for eGPUs, but it only seems to be spreading over to the west, slowly. Depending on how things look once I’m ready to ship boards, I might see if I can source them in bulk.
6 Likes
I found this 8i Oculink to Pcie adapter, but it seems there is no redriver onboard
I haven’t known about the osmeta case, that looks more reliable than everything i found so far
I am getting that adapter. Would the lack of redriver be an issue? Based on what I have read it seems Josh would add a redriver to the expansion bay module if needed?
I hoped @Josh_Cook could have an eye on it.
I’m a bit skeptical since the site is a bit weird, but it is way cheaper than the one @AdenTD mentioned. I was planning to fit the whole thing in a small pc case where the gpu can be vertically mounted
Note that you need an oculink cable to, there is none included
1 Like
Possible to get a usb4 with power delivery and oculink version?
Would be nice to be able to pair with an oculink egpu dock that can charge the laptop, have ports/card readers and maybe a hard drive in it. And have everything plugged into the rear of the laptop.
The expansion bay only does Usb 2.0, which is already planned to be available.
charging would be possible, but for this iteration it’s a bit late.
Figured as much, I know that the expansion bay supports power. Just not sure how that is actually implemented.