Rhombus Tech is working on a project to develop a computer module that supports open source software and which can be used in a variety of devices.
The idea is that you’ll have a PC-on-a-board that works a bit like a Raspberry Pi or MK802 mini PC. But instead of using this as a standalone computer, you’ll be able to slot it into a wide range of devices including tablets and notebooks.
So instead of replacing your laptop when the CPU starts to feel outdated, you’ll be able to pull out the module containing the CPU and other vital components and slide in a newer model.
The first Rhombus Tech design is called the A10 EOMA-68, and it’s expected to be a PCMCIA card-sized PC module powered by an Allwinner A10 ARM Cortex-A8 processor.
That’s the same chip used in the Mele A1000, MK802, Mini X, and a number of other inexpensive Android tablets, TV boxes, and other devices.
In fact, members of the Rhombus Tech team were responsible for the first builds of Ubuntu Linux that were able to run on the Mele A1000… which led to a number of developers porting Ubuntu, Fedora, Puppy, and other Linux-based operating systems to run on Allwinner A10 devices such as the MK802.
Right now the A10 EOMA-68 is still in the planning stages, but Rhombus Tech announced that there’s now a PCB design in place, and the next step is to produce samples that can be used for testing and demonstration purposes.
via Slashdot
Would love the ability to stick this one into my laptops PCMCIA(for that SD card reader will have to be dismantled) or ExpressCard slot and have the ability to switch from x86 to ARM on the fly. Imagine that! Browsing all day in internet while your computer barely sips power and then switch instantly to x86 and keep on doing your demanding work in Photoshop, Ableton or flight simulator! Nice!
I have an old Dell C400 (P3 866mhz) from the far 2002 that still working fine with xubuntu. Can I use this kind of PCMCIA or I need a minimum requirements?
Anyone know what happened to the CUPP PunkThis device? I was hoping to use something like that.
I’m hoping for the PunkThis board to come out too. Hopefully, they have installation instructions for Thinkpad notebooks. Right now, my Thinkpad X230 gets ~10 hours of battery life with Ubuntu. I wonder how much battery life I’d get running Ubuntu for ARM for low performance usage.
It would be nice if Lenovo or some other OEM worked with CUPP to ingetrate the PunkThis device into their notebooks. Maybe add some extra pins to the SATA connector to allow for easy installation. Kind of like with early mPCIe SATA SSDs.
It would be great to run Ubuntu on either a Core i or ARM setup depending on your requirements. Too bad I don’t think there would be a large market for it.
It’s a good idea, but it’s likely only to fly at the cheaper end of the consumer market. First, display technology is becoming a key differentiater — just look at the marketing of the Nexus 7 and the Fire HDs. Second, when it comes to speed and cooling solutions, a custom system-board based solution is likely to be maintain a significant advantage over a generic solution with a pluggable card. Third, as clkeagle implies, why would the market leaders want to sell an upgrade card for $100 when many of their customers are willing to part with $500 or a completely brand new device?
In all likelihood, you’re probably talking about a delay of up to a year or so between the introduction of a newer, faster processor and the production of an upgrade card, and even then you’re not going to get one that will match the fastest speeds produced by devices sold with the same processor.
For the enthusiast market, the approach could work, though it would have to be cheap and prove they could win a following large enough to provide economies of scale.
mike: there’s a lot in here, all of it good 🙂
1) flying at the cheaper end of the consumer market: we’re counting on that! i really look forward to being able to tell my friend, who regretted buying 2 skytone alpha 400s a few years back for her 6 and 8-year-old daughters, that there is an alternative, and it’s one that she can buy upgrade cards or low-cost replacement parts for (e.g. a new laptop case) if the kids damage them… *without* having to pay costly repair bills or replace the entire product. or, just… buy a bigger laptop chassis for her daughters, that’s more suited to when they are at school. you’ve seen the cost of the Motorola Atrix Lapdock? you can get them for $65 online! that’s just… amazing! and the BOM for an EOMA-68 laptop “chassis” (as we call it) is no different from the BOM for a Motorola Atrix Lapdock. so imagine this: $45 retail for a CPU Card; $65 retail for a 12in laptop chassis, that’s $110 for a fully-functioning *upgradeable* 12in laptop with android or ubuntu preinstalled – that’s… that’s…. amazing!
2) display technology. 24-pin RGB/TTL is pretty flexible, and covers all the way from 320×240 or even less all the way up to 2048×2048. it actually starts to depend on the CPU itself at that point, but the A10 can definitely do 1920×1080. so i’m not that concerned: it’s why we picked 24-pin RGB/TTL – it can always be converted to whatever LCD is connected, using low-cost ICs. MIPI or HDMI on the other hand, or down-converting LVDS to connect to the lower-cost 320×240 LCDs, is more problematic (read: costly). bottom line: we really really did think the set of interfaces through very very carefully. it took something like 8 months to cover all the possible products.
3) delay of a year to create a new CPU card? naah. nothing like that. nowhere *near* that 🙂 our ideal scenario is for EOMA-68 to become a de-facto standard that is recommended by e.g. Linaro that new SoCs come out *immediately* as the very first BSP from any SoC vendor. that would be absolutely awesome. the only reason why there were such large delays for the A10 CPU Card were:
a) we had to source the 5 rather unusual connectors. mid-mount HDMI was a bitch to find (at a reasonable cost. Molex’s part is $0.80 in 100k volumes, whereas Amphenol’s – the one we finally picked – is only $0.30). it took *months*. *BUT*, now that those parts have been found, we don’t need to find them again, do we? 🙂
b) we worked *with* ODMs on a “no-money-down” basis, on the strength of our main client’s committment to absolutely massive volumes. unfortunately, as we weren’t actually paying them, that meant that we had to operate at their pace. the first ODM went bust (!). the second… i don’t want to get into the details, but they let everyone down. cutting it short: we’ve done it.
c) in the future, what we might do is simply pay someone to convert e.g. the Open Schematics for IMX53QSB, Origen, PandaBoard etc. etc. into EOMA-68 form-factor. they’re pretty close already! and most of the extra space of these 3in x 3in boards is taken up by connectors that we don’t want to be on our 3.3in x 2in board. actually, we’d be delighted if someone contacted us and said that they’d be willing to do the conversion for us at no cost, in return for a cut of the revenues.
4) our client – a large factory in china – is committed to the project. it’s a series of coincidences that we happened to be in the right place at the right time, but the concept that we’ve come up with is perfect for them. so we already *have* the economy of scale. they’re also quite a big bureaucracy, so as you can imagine, it’s been sloowwww going …..
… but we’ll get there 🙂 once this first CPU Card is being sold; once there’s a first product (e.g. laptop) being sold, things will accelerate rapidly from there. breaking the ice is always the hard part. it’s just the beginning of a project that we plan to remain committed to for at least the next decade, constantly upgrading and keeping current, whilst also keeping costs and eco-waste down.
Thanks for taking the time to reply — I wish you luck even though I believe you have a steep road to climb. A couple of points, though:
1) The Motorola Atrix Lapdock was being sold off at $65, its original price was around $400. I can’t believe that you really think that a laptop chassis worth using is going to be sold for anywhere close to $65. You can’t even get a tablet keyboard docking station for that price (the Transformer docks are retailing at $150, and are still $100 when heavily discounted — and you have to add the cost of a screen to that.) Being unrealistic about this does not inspire confidence.
2) My point about the display is not that the card won’t work with higher res displays, it’s that we’re going through a period (that’s to Apple) when better displays are being integrated into newer tablet models. Yes, there will be people who don’t care about this, but in the face of a marketing blitz from the top companies, others will be persuaded to upgrade to a newer tablet/laptop because of the better displays, thus fewer people will be unwilling to only upgrade their CPUs because they want the better display too. That will probably change again at some point, but currently it’s still a factor that has to be worked against.
It’s not that I don’t want something like this to work, it’s that I find it hard to believe that it’s a sustainable business model. Modular components have been tried in the past (even by Intel and AMD) and its never caught on. Part of that may be due to it not being in the best interests of the retailers and manufacturers, but I suspect a lot of it is due to the fact that the vast majority of consumers don’t want modular bits and pieces that have to be pulled together to work (even if it might save them in the long run), they just want something that they can pull out of a box and switch on and it works.
Perhaps not everyone can afford to upgrade their iPad every year, but there’s still a big step from keeping old tech for a bit longer and starting to buy modular tablets/laptop where you have to rely on there still being a supply of affordable new modules in two, three or four years’ time.
And a more realistic approach to recycling old tech is to sell on your old devices, or donate them to charities who will refurbish them and send them on to poorer nations who can still find a use for them.
“But instead of using this as a standalone computer, you’ll be able to slot it into a wide range of devices including tablets and notebooks.”
Oh – that’s why they used a PCMCIA card! But what if you just want a ‘development card’ to use as a NAS or something?
Check out the cubieboard or hackberry a10 or gooseberry.
if you want to use it as a development, educational or R&D card, you take the case off (or order one without a case), and then you can access the DIL2-44 (same size as the 2.5in IDE connector) or for more advanced features related to TV-style products, the FPC-45. i put the pinouts onto here: https://rhombus-tech.net/allwinner_a10/orders/
as a NAS, you’d not need to take the case off: the 10/100 ethernet, USB2, USB-OTG and the SATA-II are all accessible via one (or other) end of the CPU card. only if you *really* needed the extra USB port (the 3rd one) would you need to take the case off: the 3rd USB2 port is on the DIL2-44.
Well… that’s absolutely brilliant. It would pretty much eliminate the planned obsolescence most devices are plagued with, help conserve rare-earth elements, and keep electronics out of our landfills.
In other words, this will be completely shunned by the larger companies.
yes, exactly! i’m delighted that you noticed the environmental benefits. regarding larger companies: it’s our responsibility, as humans, to vote with our wallets and our voices, and to directly influence larger companies by buying products that are clearly so much more environmentally-conscious than the alternatives, and to bung up the phone lines and communications channels of *non*-eco-conscious companies so hard that their profits they crave are adversely affected by having to pick up the phone and listen to yet another customer demanding that they respect the environment.
if you know the rules by which these large companies operate – and they operate through clinically-insane pathological adheration to profit maximisation (read “Creating a World without Poverty” and watch “The Corporation) then you *also* know how they can be manipulated and hammered into being more responsible.