I'll have to do some research on this display driver board thing. Since there are so many permutations, it might be sensible to do multiple versions, instead of having one giant board with every kind of input and output on it. The point about OLED and LCD panels is true...that suggests that it would even be worth having an eDP-to-eDP permutation, allowing you to use a display with a different (configurable) pinout.
It should at least be useful to release the board designs so that anyone can manufacture them, or flash custom firmware for them. The supply of these proprietary boards is completely unreliable.
Anyway yes, I agree on making the ecosystem more accessible. By "ecosystem", I mean "custom/DIY laptop ecosystem", broadly, without reference to a particular motherboard. Which is I guess what you're getting at with the separate motherboard tray idea. Focusing on a specific motherboard standard does help narrow down the problem space, but even if not, we do have some common problems in this space, similar to those listed in the OP. I mentioned some of these in my last post, but without a very clear focus on "what keeps people away from this".
...I have many thoughts on every component of this problem, but this post started to get very long. For now I'll provide my thoughts for discussion on what I consider to be the overarching top-level question, which is the motherboard, as it influences everything else. (I'm also interested in hearing your thoughts on making a DIY heatsink, in terms of difficulty and effectiveness. I planned on doing one for the Devotion as well...)
Every motherboard option here has its flaws, which propagate issues down to each component.
Other people [anyon_e] [MNT Reform] [and of course Framework] have approached this decision by making their own motherboards, which disregards existing ecosystems, and places the entire ecosystem maintenance burden on the creator (among other downsides like performance usually), but this approach does offer maximal control.
As for my chosen option, thin-ITX, I'll say its single biggest flaw is that it's a space built around industry, not individuals (retail). This means that even if motherboards exist with the specifications you're looking for, it's often difficult or impossible to buy them retail; for example the IMB-1240-WV (which I am using), or even worse, the eDP variant of the PH12ADI, which has a large MOQ totally unreasonable for an individual. So the entire space is geared towards large orders for business, and this issue propagates down to, for example, eDP cables, which I have never been able to find retail in the right connectors and pinouts, meaning a custom order of several from an Alibaba supplier.
This is the one thing which troubles me the most about my choice of thin-ITX; even though I don't want to be the single point of failure, the best solution I have to offer other people right now is to simply become one of the first businesses to sell some of these thin-ITX components (rare motherboards, eDP cables, Devotion parts...) at retail. Despite all this, I still think it's the best single option.
I think that general-ish, modular solutions to the Devotion's problems will ultimately be a lot more useful and interesting to people than the Devotion project itself, which is just one configuration of those solutions. Maybe "standards" is a bit much to invent from scratch, but "any kind of solution at all that other people can use" would be good!
I would say its very feasible at this point, you can buy heatpipes, copper heatsink blocks, heatclamps, etc, readily off aliexpress, reliability is another question entirely however. Breaking up the cost of the heatsink I made featured in my earlier pictures:(I'm also interested in hearing your thoughts on making a DIY heatsink, in terms of difficulty and effectiveness. I planned on doing one for the Devotion as well...)
100x50x15 heatsink - $32.62
150mm heatpipes - $17.33
50x50x6mm heatsink clamps - $11.04
lga 1700 heatpipe clamp - $19.97
50mm blower fans - $16.24
$97.20 not including tools, which for me was just a drill and some m4 tapping bits, however designs with bends in the heatpipes would require a tool for that, so another $10. Not exactly the most cost effective, however the heatsink in the second photo of my first post cost just $25.23 and likely provides similar performance. Finding cheaper suppliers off of alibaba could help but in general I would only recommend building custom in specialized builds.
I would definitely agree that thin-itx is most viable at this point, however alot of the systems you are working on could certainly be adapted to other motherboards such as sbcs, etc from what It sounds like. If new/better motherboards with varying standards are made available they can be supported within reason. Thats what I mean by communal standard, outlining general solutions for projects such as these, guidelines, systems, and more.As for my chosen option, thin-ITX, I'll say its single biggest flaw is that it's a space built around industry, not individuals (retail). This means that even if motherboards exist with the specifications you're looking for, it's often difficult or impossible to buy them retail; for example the IMB-1240-WV (which I am using), or even worse, the eDP variant of the PH12ADI, which has a large MOQ totally unreasonable for an individual. So the entire space is geared towards large orders for business, and this issue propagates down to, for example, eDP cables, which I have never been able to find retail in the right connectors and pinouts, meaning a custom order of several from an Alibaba supplier.
This is the one thing which troubles me the most about my choice of thin-ITX; even though I don't want to be the single point of failure, the best solution I have to offer other people right now is to simply become one of the first businesses to sell some of these thin-ITX components (rare motherboards, eDP cables, Devotion parts...) at retail. Despite all this, I still think it's the best single option.
There might be a better way to communicate what I mean for sure, as of right now I plan to finish my proof of concept build along with documenting it, in the hopes of attracting attention and getting others interested.Maybe "standards" is a bit much to invent from scratch, but "any kind of solution at all that other people can use" would be good!