Last chance to enter The Hackaday Prize.

An Amazing DIY Single Board ARM Computer with BGA

DIY Single Board Computer ARM

Typically, you buy a single board Linux computer. [Henrik] had a better idea, build his own ARM based single board computer! How did he do it? By not being scared of ball grid array (BGA) ARM processors.

Everyone loves the Raspberry Pi and Beagle Board, but what is the fun in buying something that you can build? We have a hunch that most of our readers stay clear of BGA chips, and for good reason. Arguably, one of the most important aspects of [Henrik's] post is that you can easily solder BGAs with cheaply available tools. OSH Park provides the inexpensive high-quality PCBs, OSH Stencils provides the inexpensive stencils, and any toaster oven allows you to solder even the most difficult of components. Not only does he go over the PCB build, he also discusses the bootloader, u-boot, and how to get Linux running.

Everything worked out very well for [Henrik]. It’s a good thing too, cause we sure wouldn’t want to debug a PCB as complicated as this one. What projects have you built that use a BGA? Let us know how it went!

Comments

  1. steve says:

    Damn… I’m impressed! Not as much as the hardware, but someone willing to tackle this hardware and know enough about software to get Linux running on it! (I’m a hardware guy, if you couldn’t tell….)

  2. nixieguy says:

    So awesome, the blog it’s not avaliable just after being posted here. XD

  3. lel says:

    The blog is still down…… Excellent work Hack a Day!

  4. charlie says:

    I wonder if he actually stayed within the specifications for OSHPark. I know the specs are a little better for their 4 layer boards, but I think it’s still right on the edge of what’s needed for many BGA packages.

    • Renald says:

      Ya he might have gone out of spec on this one. They dont restrict you from uploading something out of spec but dont expect it to be bug free…. I have used OSHPark for over 150 boards and whenever I get close to the tolerances they spec my failure rate goes up a lot. (around 15% failure at 6mil traces and 6mil spaces) I have received some with missing traces or bonded to ground planes etc…. Be careful and stay away from the limits they give as they fail sometimes. I bet debugging a mistake on that non e-tested board was a roll of the dice lol. Cool project though!

    • atomsoft says:

      That’s the main reason I don’t do bga devices. . If oshpark had specs for them omg id be in heaven.

    • dizot says:

      Yeah, I’ve steered clear of BGAs on my hobby PCBs for exactly this reason. Their design specs are just outside of plausible for BGA requirements.

      Kudos to Henrik for squeezing a functional prototype out of an OSHPark PCB. Extra kudos for getting linux up-and-running. I use the older AT91SAM9263 in a production device, and I know what a nuisance it can be.

      I’ve also used OSH Stencils for several hobby projects, now. I have only good things to say about their stencils (with pitches as small as 0.4mm on some QFNs). They’re inexpensive, cut perfectly, and their turn-around time is incredible.

    • Alan Hightower says:

      I’ve tried many times to escape even a 1mm BGA on 4 layers with OSHPark design rules and always fail. I’ve even tried recently with iTead studio rules which are slightly better and fail. I’m sure if one were willing to risk tolerances, you might get lucky a fair bit. But I’d want consistency. Even .8mm DDR is tough on a drilled board w/o via in pad.

      The biggest challenge to BGA is a PCB DRC, not mounting one. I’d want 5/5/5 mil trace/clearance/rings and 10 mil drills to be happy. Would be nice if a budget board house stepped it up. I would even pay extra for blind boards if the mark-up was in-line with OSHPark/iTead pricing levels (eg. 2x for blind vs drilled).

      • Alan Hightower says:

        Wow it’s been a while since I looked at OSHPark. For 4 layers, specs are 5/5/4 width/clearance/rings with 10 mil drills. Yay for 1mm BGA viability now (primarily FPGAs!)

  5. fhkfyk says:

    This would be a DIY ARM computer if he etched the board himself and soldered it with a $20 soldering iron (FYI americans, $20 buys me food for a whole week). Right now it’s just a nice small board with some SMD and BGA components on it.

    • omegacs says:

      Are you sure about that? I thought for it to be a true DIY ARM computer he had to grow the silicon crystals, slice them, and perform lithography with a magnifying glass and the sun. Plus, he didn’t hand-forge the soldering iron tip, so it really doesn’t count.

      …………???

      • Trui says:

        Only if you carve the magnifying glass from a block of ice you’ve cut out of a glacier.

      • pcf11 says:

        Your argument is lame, and worn out, It was a waste of electrons for you to have even posted it, as many others have used it before you. That makes it no more valid today, as repeating bullshit does not make it any more true. False analogies extended into absurdity are fallacious arguments.

        The more you know…

        • omegacs says:

          HUH???

          The OP makes the claim that not doing certain elements of the project “by hand” makes it “not DIY”. My post does indeed point out how ludicrous this argument is via intentional “reductio ad absurdum”, plus it’s funny. That however does not change the fact that the original argument is itself absurd.

          The term “DIY” means that you did a reasonable fraction of the project yourself, rather than purchase it premade. Given that he:
          a) designed the schematic himself
          b) designed the board layout himself
          c) applied the solder paste (oh noes he used a stencil!) himself
          d) placed the parts himself
          e) reflowed the board himself (again, oh noes he used a store-bought toaster!)
          f) applied his own chnages to u-boot and the kernel so they boot
          …I’d say he Did a lot of It [H]imself.

          You want specious/fallacious arguments, look at the original post that claims that not only did he not “DIY” because he didn’t do the single step of etching the board himself (rather difficult since it’s a 4-layer board), but asserts that the use of a $25 toaster instead of a $20 soldering iron (on BGA, which is impossible) invalidates “DIY”. Then there’s the complete non sequitur of how much food costs in his local area.

          I’m also quite surprised to learn that my argument vis-à-vis etching his own chips is not new, and has been stated many times in the past. I also appear to have missed the depth and breadth of the hand-forged soldering-iron tip community.

          The more you know, eh?

    • HV says:

      FYI $5 will buy you someone who will pretend to care what you pay for food at fiiver, for a week.

    • pcf11 says:

      I don’t care what their iron costs but I agree that for something to be DIY one does have to do it for themselves.

    • Matthew Rogan says:

      DIY is a an acronym for Do It Yourself, he designed a BGA board barely within OSHParks Spec, reflowed it and then ported linux to it. The first and second of those is very difficult and he’s achieved something usually out of the hands of us mere mortals.

    • Tiago Dias says:

      Why did you feel the need to point out that 20$ is a lot of money to you?

  6. jam says:

    Is this a shared project on OSH? Couldn’t find it.

  7. xobmo says:

    yeah, I am interested… But site is down.

  8. ragooman says:

    getting ‘site is down’ response on hforsten.com from several places too
    hope it’s back up sometime, like to read more about his work

  9. It’s like reddit effect kind of ddos for website server

    • dsblackout says:

      Slashdotting. It’s called slashdotting. Slashdot be older than Reddit, foo.

      • pcf11 says:

        I met the original Slashdot team at a Linuxcon shortly before they sold the site. After that I stopped going there. They were massively insufferable assholes in person. I guess rapid wealth does that to a lot of folks? Who knows, they might have been born that way too. I can’t say I knew them then. That was during the PT Cruiser giveaway if memory serves me today. I wasn’t exactly taking notes. Frankly I’d like to forget the moment. There was a time though when /. was a lot of fun.

  10. Pixel Pirate says:

    I’d be all for it if I had some tips on how to align those demons.

    • Mojoe says:

      I don’t have experience at this (no reflow oven), but if I’m not mistaken, the surface tension of the molten solder pulls it into allignment.

      • Phil says:

        Yep, BGAs are some of the easiest to place. Just use minimal flux, put it on the pads the best you can and the heat does the rest. Too much flux and it will bubble making yours BGA skip across your board.

  11. thenoman says:

    “Everyone loves the Raspberry Pi and Beagle Board”
    Yes, everyone loves binary blobs and shitty usb root hubs…
    NO

  12. guan says:

    Forgive me if this is a stupid question since I’ve never worked with BGA: Do you actually need solder paste under a BGA package? I thought the balls were solder.

    • notmyfault2000 says:

      I’d imagine the paste there is more to keep it where you set it down, so it doesn’t slide everywhere while you’re moving it before reflowing.

    • dizot says:

      The paste is to promote wetting of the ball to the pad. The solder ball wouldn’t be very likely to form a decent intermetallic bond with the pad without the paste / flux.

  13. movax says:

    Nicely done! Think he got lucky with the RAM working as well as it does but great to see someone tackle a BGA project personally. :)

  14. Andy B says:

    That’s an excellent effort with some difficult components. ENIG was the right choice for the BGA as opposed to HASL. The flat surface is very important. In my opinion he did get lucky with the BGA though.

    He should not have added solder paste to the BGA pads, there is more than enough solder in the balls themselves even though their RoHS composition makes them a PITA to reflow with hobby equipment. Flux is required though. The service he used with the ball pitch on that BGA means that there will be no solder mask between pads, traces or above the pad-to-escape-via traces.

    Adding additional paste and also shrinking the land pads increases the risk of shorts occuring between balls due to flooding and also increases the risk of a ball completely wicking away down an escape via.

    It’s actually possible that if he used leaded paste that the paste reflowed and the balls didn’t, which would work but obviously creates a less than ideal joint. Very hard to spot if this has happened though.

    I hope this doesn’t come across as overly critical, any hobbyist success with a fine-pitch BGA is a great result.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

Follow

Get every new post delivered to your Inbox.

Join 91,219 other followers