This update lays out the roadmap for the EOMA68 project and what the future holds for modular, eco-conscious, libre computing.
As I’ve previously mentioned, the EOMA68 project has five-year history even before this crowdfunding campaign. It’s almost overwhelming for me to describe to people the multi-faceted aims of this project. That there are (at the time of writing) over 1,700 pledges is amazing: I’m so proud to be backed by so many people who believe in this project.
Remember, the goal is not just to sell a few thousand gadgets and go home with the profits. We aim to reach mass-volume with an eco-conscious strategy that at the same time saves users money in the long-term, puts people back in control of their hardware, reduces spying opportunities for governments and corporations, and reduces people’s stress levels due to incompatibility, viruses, irreparable accidental damage, and the like.
Central to this long-term strategy is building up a healthy ecosystem of products beyond those offered in this campaign. Already, there are other EOMA68 housings and computer cards in the concept phase or actually under development. Here’s a summary of the current landscape, some paths already followed, and the roadmap ahead.
There have been several of these attempted - literally dozens of SoCs have been evaluated: we showed a small glimpse into this world with the processor update.
A20, is the CPU we’re running with. Its news page is here.
computer card got a long way before it was realised that the power consumption of an older quad-core Cortex A9 would simply be too high.
series - used in the original Beaglebone - would be a perfect candidate as there are even full PCB CAD files available which can be adapted… except that this SoC is now so old that it’s no longer attractive.
A33](http://rhombus-tech.net/allwinner/a33/news/) Computer Card only to find late in the day (due to incomplete datasheets) that the LCD output was hard-restricted to 1280x800 where the EOMA68 Standard "Type II" 5 mm specification requires a minimum of 1366x768.
Ingenic to answer that the reason why they state that their SoCs are restricted to 1280x720 is not down to the LCD interface (which can easily handle 1920x1080 resolutions), it’s down to the internal memory bandwidth alloted to the hardware-accelerated VPU. The VPU can only handle 1280720 @ 30fps. So this was why the jz4775 went ahead, only to find that the OS support is at a bare minimum and quickly eroding for MIPS32. All SoCs using PowerVR are automatically blacklisted, so the jz4780 will never be considered until ImgTec provide full source under libre licenses.
is a quad-core 400 MHz "Unified Processor" design was also attempted: samples are actually functional. However as they use open64 not gcc, chances are extremely remote that the IC1T will ever have GNU/Linux distro support.
concept is also under consideration: now that the Zynq 7020 is commonly available with full PCB reference designs, if those reference designs could be converted to useable formats it would be easy to do without overwhelming financial investment required.
have Intel spyware back-door co-processors. We have no intention of endorsing that kind of behaviour.
This is a partial list - the amount of personal funds spent on these over the past five years is somewhere around $16,000 - that’s excluding the $20,000 spent on the EOMA68-A20. The SoCs currently being evaluated are the Allwinner R40 (due out very soon) and the same Exynos Octa-Core 28 nm SoC used in the NanoPi3. This is going to take time and resources to evaluate. They’re not going to be available immediately. We need your backing for this project and the current Computer Card before we can be in an established financial position to properly evaluate and bring you these faster Computer Cards.
In parallel with the Computer Card evaluation, over the years people have been suggesting projects which could be converted to modular designs: these are maintained on the community ideas page. The best ones
Here are a few:
was an early concept that had its beginnings in the Micro Engineering Board. Its "bigger brother" is the Mini Desktop which includes on-board Ethernet and on-board audio (USB CM108AH).
tablet](http://rhombus-tech.net/community_ideas/kde_tablet/news/) has a long history behind it, where we now feel that the better option would be to create a 9" 4:3 aspect ratio (1024x768) tablet with associate keyboard-dock similar to the Always Innovating "Touchbook." Space is extremely tight in the 7in form-factor, and there are significant difficulties finding matching components. It’ll be doable… but from experience we know that it will take at least a year’s worth of full-time development effort, the costs of which would of course have to be covered by actual market demand.
was an early "dumping ground" and research exploration page for component sourcing. ThinkPenguin helped narrow this down to a very specific and popular segment, the 15.6" Laptop market, with full-sized keyboard including numberpad being a known attractive feature for businesses, as well as the larger screen size making for easier viewing in office environments.
Console](http://rhombus-tech.net/community_ideas/games_console/) is being run by an independent team that loves the fact EOMA68 allows them to not have to deal with DDR3 RAM and other issues, that they can consider doing a 2-layer or 4-layer PCB at a fraction of the cost, and that they can upgrade in the future. They’ve stuck with us for many years and I have been helping them quietly in the background on PCB design and component sourcing.
There are so many possibilities here - Digital Cameras, Camcorders, Hand-held video and DVD players, GPS Sat-Nav players, In-car Stereos (where you’d plug the Computer Card in, instead of a USB or Memory Card), etc. We’ve chosen to focus on the Micro-Desktop, the A20, the Pass-through Card and the 15" Laptop for this crowdfunding campaign, because it’s what can be managed. We invite you to help out, especially if you have a background in electronics. We need you - the makers: it’s why we created the Breakout Board, so that you can consider getting involved and link up to parts from Adafruit, just as I did with the prototyping of the LCD+CTP for the Laptop’s touchpanel. Albert then helped out and did AZERTY actually during this campaign!
What we are quite likely to do next, however, (unless someone comes forward and offers to sponsor a project, just like ThinkPenguin did) is adapt the Laptop’s PCBs to create an LCD monitor that can’t be used to spy on you. It will basically be the Laptop’s "PCB1" (main PCB) and a Pass-through Card. It’ll be upgradeable to an all-in-one Computer simply by popping out the Passthrough Card and popping in an EOMA68 Computer Card. The other project will be an all-in-one Keyboard + Trackpad (similar to the Logitech K400r except without the bluetooth). This (again, surprise!) will be based on the Laptop PCB2, which is the STM32F072 PCB. A different PCB1 will be needed (similar to the Micro Desktop… in fact we might be able to use the Micro Desktop PCB) and in a relatively short amount of time we will have a Keyboard+Trackpad (again with a Pass-through Card) that you can use as a standard USB Keyboard+Trackpad… but that can be converted to an all-in-one Keyboard Computer simply by popping out the Passthrough Card and putting in an EOMA68 Computer! It’s so simple and amazing when you envisage these things, it’s just incredibly hard to understand why this hasn’t been done before.
So yes: there is a roadmap. We have a lot to do. Way too much, in fact, so we need your help. So if you believe in what we’re doing, back the project, join the mailing list, and get in touch. Let’s help each other out, take back control of our hardware, and help others to do the same.