Cubieboard: A small, high-performance ARM box
cubieboard.org
cubieboard.org
- The website. Jesus. With all the templates and content management you can pluck out of thin air, the result is very shoddy. Apparently that isn't a warning sign enough for some people.
- No corporate history. The company has another site but their blog there is empty. No news. Dodgy RSS listings. The whole thing is even shoddier than the product site.
- The payment is through Paypal. Address was visible yesterday (now encrypted) as mr.hipboy@gmail.com. lolwut? What a handle. Very professional. Real name appears to be Luffy Wang of Hesseney Road, Hubei 150, HK. (hipboi@qq.com) This guy really likes hips.
- The product is $49. Less than the price (delivered) of two Raspis with a SATA controller and port. This was what made my eyes light up. The phrase "too good to be true" exists for a reason.
- And while I'm talking about specs, why do the pictures and words disagree so much? How many card readers are there? Are they SD or MMC?
- Oh and it's available almost without delay. Despite that there are no reviews, no tester units in the wild. No videos of on working. Not even the hint of a fart from a prototype.
+ There is a GitHub page: https://github.com/cubie-tech This is probably a detail I wouldn't have bothered with if I were running this scam.
Yesterday my acquaintances on G+ and I decided that this is most likely a scam. I really hope it's not and I'll be the first in line for the second batch, but please be careful.
EDIT: Using some different terms brings up a ton of articles about the board from popular sources, yet they're all new within the past few days. I dunno, it might actually be real, but it me a bit uncomfortable, but that may just be me.
Seriously other than the handle hipboi/hipboy in an email address that was removed from the site yesterday, there's nothing definitively linking Cubietech/Cubieboard to Tom Cubie.
And I can't see anything about the Cubieboard on Tom Cubie's Facebook or G+. That also strikes me a little odd.
He has shared the info with people who are friends on G+. But ripping in to someone for their choice of nickname? Seriously? I've seen many worse nicknames by people in this so called startup community that seems to be more and more filled with cynicism and negativity.
Apparently if you don't use a Mac/rails/js it isn't worth looking in to.
† Message board nerd pro-tip: avoid the word "lolwut" if you want to maintain the pretense of evenhandedness.
Gather round children! This is what a two-minute scam website looks like. No effort, discrepancies everywhere (specs), technical inconsistencies (domain whois). But look somebody in a forum said it's a real person, it must be real! No.
If you charge into a site like this and trust it, you're an idiot. Your idiocy is not dependent on this one being a scam. Put the card back in your wallet and wait until the product actually exists.
And the lolwut was in keeping with the style of the handle and the website. I thought that was rather obvious and certainly wouldn't have thought it needed a sidebar conversation... But here we are.
Would you like a spade at this point?
The point is: he built a business, whatever his intentions before he started, his intentions are now in selling.
The Mele A1000 http://rhombus-tech.net/allwinner_a10/hacking_the_mele_a1000... is $70, a mere $20 more: it's 2x the board real-estate, another $4 in connectors, it has a case, and it's been around for ~1 year now.
What's interesting is how lagged the development world is from the consumer electronics world. If you want an Allwinner A10, or a Marvell Kirkwood/ ARMADA 300 piece of kit, and you want it at the best possible price, possibly before dev kits show up, you need to scope out consumer electronics, not dev boards. The Mele was a product, this lack's the consumer experience of a product, is just a platform, and it, like many dev kits, show lag v.s. what the consumer boys are up to.
This is simply a small board exposing a very cheap ARM A8 SoC that does everything (just add ram+flash): hold on to your fucking pants and get used to the future. Course, may not materialize. My OpenPandora hasn't. ;)
And yeah, seems like every few years we get more and more college grads who seem to be surprised that there is, in fact, a lot of very cheap hardware out there to be had. I've been running and using micro-scale Linux hardware like these new-gen things for decades. Gumstix, BananaBoard, et al.
Okay, the <$50 pricing is pretty nice, but not entirely unexpected given the trend over the last 10 years of having these things get cookie-cut around. The whole point of ARM in the first place was to get to this position ..
Ever since the Raspberry pi shipped, credit-card sized computers have been coming out the woodwork. As expected, now there are projects on vWorker (ex RentACoder) asking for engineers to design such computers with state of the art specs with a maximum bid of $499. I can't make this shit up! $499! At least toss us a bone and add a zero so we don't laugh quite so hard!
Anyway, aside from that rant, this is a Great Thing. I predict an array of products that are now possible due to a flood of dirt cheap, physically small computers that can run Linux. One off industrial controllers that would have taken a week of development time can now be pushed out in a day, etc.
No, they've been here all along in the SBC and industrial spaces. They're just on your radar now because they're cheap.
Is Gumstix a line of "one-off industrial computers"?
I'll amend it to "cheap credit-card sized computers" if it makes you feel better.
Gumstix have been around for what, about 10 years IIRC? They are more expensive than what I'm talking about. The point I'm making is that CPU boards on the level to run flavors of Linux easily are approaching "jellybean" component status and that will enable applications that simply weren't feasible before.
[citation needed]
I'm thinking it would cost at least $150-$200 to prototype, per PCB revision. If it was design only, I wouldn't trust a random freelance designer to get it working first go.
I remember one weird case where bunch of EEG units acted weird, well noisy. The highly experienced MSEE couldn't explain it, because it wasn't the high sensitivity input amp or anything on the board. I eventually figured out it was the off-board battery, and it turned out to be the impedance would climb causing a noisy EEG, as the battery was wearing out in fairly short time from abuse (excess low discharge in SLAs will damage them). Perhaps a few extra capacitors might have help.
It would be the same for these A10/A13 clones, weird glitches can crop up, and even with an experienced engineer, the design could require a PCB revision and a new prototype.
In case it's not clear, the $499 is the price they are offering to pay an engineer to do the design, not the price they want the end product to sell for. Hence the reason they are off by close to two orders of magnitude.
I think on vWorker and some of the others, people just post projects with what they can afford -- a few hundred $$$ -- and see what sticks. There's always someone who will take them up on it, even if only to disappear forever when they realize how difficult it is.
I have received at least one personal invite to design just such a board in the last week (max bid $499), but I already deleted it.
There have been cheap, affordable, fun eval boards like the Rpi out there for years. Gumstix, Beagle, and so on. Prices are coming down, sure, but don't kid yourself that the RaspberryPi is something new and ground-breaking: its just a surface feature of a rather larger industry.
The point of the pi and its copies/lookalikes, etc is that they are made in production quantities not just test & evaluation quantities.
Go ask Atmel for 10,000 of a random AVR evaluation board. That's probably half of all the boards they ever intended to make.
http://olimex.wordpress.com/ or http://goo.gl/5v26a to the post.
Allwinner will reduce minimum shipping quantities for open hardware projects. Closed development OEMs have a 100k minimum order quantity.
Clicking on things gives varied results though. For example clicking the buy button brings you to a page that indicates you can't actually buy it yet.
HN has been big on launch early and get feedback. How shallow it is to have negative feedback based on look. Let's wait and see how the product turns out, rather than commenting on their web design.
The guy is working on physical hardware. Do you know how difficult that is? To ship a device like this? A lot of the issues you mentioned are the last of his worries.
Gumstix are more expensive but easily available in any quantity, so that's what I'm designing with now.
I bought one last week and it arrived no slower than amazon.com so I think it's safe to say orders are no longer taking forever.
It's not that bad
It can run a variety of linux variants. But all ARM processors that I am aware of can, even going back to the chip in the old Acorn Archimedes from the late '80s. Also, I have personally never heard of a chip that is able to run android but not able to run linux and I am not entirely sure how you would even go about making one.
We don’t claim to have all the answers. We don’t think that the Raspberry Pi is a fix to all of the world’s computing issues; we do believe that we can be a catalyst. We want to see cheap, accessible, programmable computers everywhere; we actively encourage other companies to clone what we’re doing. We want to break the paradigm where without spending hundreds of pounds on a PC, families can’t use the internet. We want owning a truly personal computer to be normal for children.
From here - http://www.raspberrypi.org/about
Then why accuse them of astroturfing?
I could quite easily imagine cubiebox being no more than an amateur scammer's effort to cash in on the uptick in interest in ARM boxes.
In particular, the grammar on the parent Company's website (http://cubietech.com/) is horrific. Plus, it has a newsfeed populating from god knows where: "Missing NY teen found safe Marriage doesn't make people happier, it just stabilizes the happiness they already have. The research shows that the married people surveyed weren't any happier than..."
Not encouraging.
I've no reason to believe they are scammer.
After a google search this "http://www.linkedin.com/pub/dir/?first=Tom&last=Cubie profile belongs to the cofounder.
edit: Some people at least are managing to run Ubuntu on a similar hardware, so I guess it'll work. http://liliputing.com/2012/07/linux-distributions-that-can-r...
AFAIK the one exception is the video decoding hardware, where the Linux libraries are apparently sub-par.
Some jumping off points for resources: http://rhombus-tech.net/allwinner_a10/ https://www.miniand.com/forums
http://lists.phcomp.co.uk/pipermail/arm-netbook/2012-March/0...
I just ordered a Mk802 II for $65. It has a 1GHz Cortex-A8 single core processor, Mali 400 graphics, 1GB of RAM, 4GB of storage, a microSDHC card slot, 802.11b/g/n WiFi, 1 full-sized USB port (usable for power), 2 micro USB ports, and a full-sized HDMI connector so you cna just plug it in to your TV.
It ships with Android 4.0 on the internal storage (with access to the Google Play Store) or you can run the linx dustro of your choise from a microsd card.
http://liliputing.com/2012/08/mk802-ii-android-4-0-mini-pc-n... http://www.rikomagic.co.uk/forum/viewtopic.php?f=2&t=305
Someone needs to draw away the popular attention from those with the will to get stuff done.
If I can re-furl my barbs & continue, if the Cubieboard were taking orders right now it would be a shit show of people coming in from Engadget and HN and all over buying these things up then being extremely aggravated they A) don't just work B) require some fairly serious willingness to chase out the good information on uboot, the sunxi platform, & to get your OS installed: there's not much in the way of authoritative, well compiled references for doing so.
I was pretty shocked seeing this board on Engadget today. There've been a steady stream of similar-ish boards, such as the $60 Hackberry, also A10 based, which too is still in "prerelease" form.
] HGR
] HPLOT 10, 10 TO 20, 20Also, the RasPi has a huge initial advantage in that it was in the space first. They have a leg-up in publicity, marketing, and simply an initial user-base. If the Cubieboard benchmarks don't absolutely trounce the Raspberry Pi for almost every type of computation, I don't see the Cubieboard overtaking the Pi anytime soon.
Still, a cool board nonetheless.
Has either of them actually good (foss) drivers? At least Mali400 has "limadriver" project going on, quick googling didn't find anything for raspi.
Overall I'd love that Allwinner A10 got more attention, it seems fairly popular chip in cheapo chinese tablets. Getting GNU/Linux instead of Android for those would be neat.
I have to ask, what is the preoccupation with FOSS GPU drivers? Not just with boards like this, but even standard desktop hardware. Does anyone seriously think a firm like NVidia is going to illicitly slip spyware or something into their binaries?
also this is no 'scam' project..
>the cubieboard (intended as BeagleBone competitor) and that aliexpress shop are both from Tom Cubie (aka hipboi). Also the author of the initial u-boot/mmc for sunxi and the software engineer at Allwinner Tech who GPLed the code released by QWare.[1]
assuming this is accurate, this puts him in a pretty good position to support it.
Fwiw, I have Linux running on my a10 tablet and it runs pretty well.. obviously a light weight desktop.. ff runs well on most sites, although there is slow down on complex sites eg. engadget
there is some linux support for the gpu, but not the video acceleration (hopefully, 'yet').
1. http://www.cnx-software.com/2012/08/31/49-cubieboard-allwinn...
Arduino is nice because I can use it to prototype something, and then if I want to mass produce it I can take the open schematics, rip out what I don't need and produce custom boards. Is that possible with any of these ARM solutions?
They've got one version running an admittedly slow i.MX processor, and another coming down the pike running an Allwinner chip. As a bonus, though, it's all reasonably easy-to-assemble TQFP chips.
The A10/A13 require memory which is only available as BGA, though.
http://olimex.wordpress.com/2012/04/24/cortex-a8-in-tqfp-sur...
http://shop.8devices.com/carambola
Anyone have some experience with these?
http://www.hardkernel.com/renewal_2011/products/prdt_info.ph...
With OpenBSD it would work well as a home firewall or similar mini network appliance.
Maybe some day! :)
UPDATE: It'd be great if ran Chrome OS, too.
That said, you're right that the board itself shouldn't pull anywhere near 2A.
I love the fact that more people are building these kind of devices at this price point. I have been looking at pc104 boards for years but they were always far too expensive for anything with some decent kick.
Give it a few more iterations and a few more people trying to build similar devices and then there is a proper marketplace for essentially a new consumer device class, in much the same way as the OLPC project was the catalyst for the emergence of dirt cheap netbooks. There will be tonnes of people working on this after witnessing the demand for the r-pis.
And anyone who thinks that the r-pi lot will do anything other than welcome things like this is off their head. They are a non-profit with the stated aim of getting as many of these kind of small cheap devices out there as possible.
Remember, you don't compete against an organisation like that in the normal sense of the word, as if you put them out of business by doing it bigger and better and cheaper, then all you have done is helped them to achieve their objective without them having to do as much work.
Personally I suspect that it is exactly the kind of development the r-pi lot were hoping for.