…here are the Gerbers (a visualisation of the printed circuit board or PCB) for the finalised version of the Raspberry Pi. I get several messages every day asking what it can possibly be that we are still working on: I hope you will understand on looking at this why the routing, which has to be quite spectacularly complicated to minimise expensive PCB features and to keep things tiny, took as long as it did! That snarl in the middle is the signal escape for the BCM2835, the chip at the heart of the Raspi. The elves have been working overtime.
Question for the folks who are following this project closely, is the plan for the Raspberry Pi foundation to make it open-source hardware? The project page says “An ARM GNU/Linux box for $25”. The GNU/Linux part is open source of course, but we couldn’t find anything specific for the hardware. Post up in the comments if you know, that would be really cool if it was! It looks like in the Q&A they are under NDA for some of it, so maybe it can’t be for some of it, but a lot of of it – seems possible!
Get the only spam-free daily newsletter about wearables, running a "maker business", electronic tips and more! Subscribe at AdafruitDaily.com !
From a cursory glance through the comments, Liz seems to say no:
“We would have liked to be able to have a completely open platform, but until the open-source community tapes out a chip, it looks like that will be impossible. (If you have the expertise, please get working on one. I would love the open-source community to put their money where their mouths are on this one; it’d be good for all of us.)…”
The intent is for it to be as open as possible, however the Broadcomm processor there using has some proprietary interfaces which they might not be able to release.
For the time being they haven’t released the design files since you can’t easily hand assemble such a board, but this is the little that I know.
I’ve found it challenging to route all the connections out for a 44 pin QFP – I can imagine the work for dealing with a BGA with hundreds of pins. "Signal escape" indeed.
Hmm…if the Broadcom processor isn’t available to the public, then this is just a ploy for Broadcom to get free software from the community and counter TI’s OMAP on the Beagle Board/Bone. BTW, Broadcom is on the board of directors for Raspberry Pi. I suspect that Raspberry Pi’s founders have good intentions, but those intentions rely on the subsidization by Broadcom which might not have such pure motives. So a $25 subsidized Linux board is just hype and wouldn’t be economically feasible if it had to stand on it’s own.
Akiba – I agree here… Broadcom has a LONG history of paying lip service to open source. They are one of the LEAST open-source-friendly companies out there.
“We would have liked to be able to have a completely open platform, but until the open-source community tapes out a chip, it looks like that will be impossible.”
Wrong – all the open-source community has to do is purchase chips from a vendor that plays nice.
TI AM3358 – http://www.ti.com/product/am3358 – Silicon errata, full TRM for the chip (About the only thing not 100% documented is the GPU core – but that’s a lot better than the BCM2835 having zero documentation available), SDKs and examples
Now I agree that some of TI’s IC lines (the OMAP series, for example) are somewhat less open – but the BCM2835 is closer to the Sitara line than the OMAP line in capabilities.
Everyone is salivating about buying a Raspberry Pi. It is a nice project. But it is closely tied to Broadcom in terms of design. As of today (post-date) I still don’t see support for accessing the DSP capability of the Raspberry Pi’s Broadcom chip.
These non-Intel x86 boards with MMU capability (ARM etc.) can all "run" Unices quite cheaply and efficiently, but they’re all tied in a big-way to the hardware (unlike comparatively inefficient x86 compatible processor designs). Each non-x86 design typically requires a specific Board Support Package (BSP) and a highly modified board-specific tool-chain and peripheral drivers.
Buy a dirt-cheap home router (e.g., the venerable Linksys WRT54GL) and try to actually develop and (importantly) compile a new executable for it under the likes of OpenWrt. It is a steep slope to climb and there’s no guarantee the device will exist long-term.
On the other hand, try an x86 compatible "embedded" board (PC-Engines Alix comes to mind), and almost any mainstream Linux/Unix distro will be far easier to use at any level. Yes, an Alix board may be three or four times the cost of Raspberry Pi (still less than $100 at the low-end), but for non-production and/or low-volume use, they more than pay for themselves through ease of use.
There was the hope that someone like Google with Android would try to standardize some non-x86 reference designs and open them up. But this hasn’t happened and maybe never will. This may be because the non-x86 Android platforms that exist today are evolving too fast. It may also be due to Google trying to monitize the likes of restricted access to the "official" Android Apps-Store. Hmmm…
Wow, the negativity is almost overwhelming. How about we let people choose whether to buy and use the product or not instead of carping about how high a bar you have to hit before you can use the word ‘open’?
To just address one point, chosen almost at random: "So a $25 subsidized Linux board is just hype and wouldn’t be economically feasible if it had to stand on it’s own." From reading the raspberrypi.org site, any "subsidy" amounts to two things: Broadcom bending its regular business rules by dealing with a low volume customer in the Raspberry Pi foundation, and a handful of Broadcom employees helping out with a project they believe in, for no money, on their own personal time. You will note that this does not involve any sort of per-unit subsidy, so once the design work is done it certainly looks perfectly economically feasible to me.
And as a counterpoint, many manufacturers offer subsidized dev boards for their processors (e.g. TI LaunchPad) and seem to suffer no community umbrage for doing so, or for not opening up all related IP. (e.g. Atmel’s AVR debugging protocols)
tl;dr – Seeing the amount of conspiratorial abuse these people get for actually trying to find a way to make programming more available to children pisses me off.
I think for most the issue is not whether this hardware is subsidized or not monetarily. The problem is that Raspberry Pi is subsidized *intellectually* – Because Broadcom withholds technical documentation for their devices, the Raspberry Pi project cannot exist without them providing support not normally available to such a project.
Compare to, say, the TI LaunchPad, which is monetarily subsidized, but not intellectually – The datasheets for every component are available, as are the complete schematics.
Same for the BeagleBone – It uses a processor (TI AM3358) similar in capabilities to that of the Pi, yet unlike the Broadcom SoC in the Pi, a complete technical reference manual for the AM3358 is available. The only thing that may be undocumented is the PowerVR core – but that’s a lot better than the Pi’s SoC for which there is ZERO documentation available to anyone other than the Raspberry Pi team.
So Broadcom is trying to hype the Pi in “open hardware/open source” communities to try and get some free development for their products – but they have given NOTHING in return, and most importantly, the device is NOT open hardware – not when the most important chip on it has zero documentation available beyond a marketing blurb.
The Pi team can claim to be independent all they want – the fact is that they are not, if Eben were not a Broadcom employee with access to documentation normally not available, the Pi would be stillborn unless it were based on a different chipset.
Broadcom is hyping the Pi? Where?
“Your search – site:broadcom.com raspberry – did not match any documents.”
I don’t even know how it would benefit them, because they aren’t in the business of selling you chips to begin with. Corporations who build millions of devices on SoCs have no interest in DIY hobbyist-produced software, they want guarantees and support contracts, and most of them care about securing very clear licensing terms. It strains credulity to assert some “free development for their products” motivation here.
The BeagleBone looks exciting, and we should all commend those TI employees for volunteering to design it. It’s one flaw that I can see is that it will sell for 3 to 4 times the price of the Pi. The Pi’s charitable foundation has a very clear target on their web site: “We plan to develop, manufacture and distribute an ultra-low-cost computer, for use in teaching computer programming to children.” For that goal, I would say cost is the most important factor, and their approach is perfectly appropriate. You can’t deny that there are a lot of children who could be reached with a $25 board that could never touch a $89 board, and there are vanishingly few kids who could do anything with the specs for a TI or Broadcom SoC. Spec availability is irrelevant to them.
So the Pi team has access to documents that you and I don’t. I remain unconvinced of exactly why I should be offended.