The Open Source Hacker’s Laptop

[Tsvetan Usunov] has been Mr. Olimex for about twenty five years now, and since then, he’s been through a lot of laptops. Remember when power connectors were soldered directly to the motherboard? [Tsvetan] does, and he’s fixed his share of laptops. Sometimes, fixing a laptop doesn’t make any sense; vendors usually make laptops that are hard to repair, and things just inexplicably break. Every year, a few of [Tsvetan]’s laptops die, and the batteries of the rest lose capacity among other wear and tear. Despite some amazing progress from the major manufacturers, laptops are still throwaway devices.

Since [Tsvetan] makes ARM boards, boards with the ~duino suffix, and other electronic paraphernalia, it’s only natural that he would think about building his own laptop. It’s something he’s been working on for a while, but [Tsvetan] shared his progress on an Open Source, hacker’s laptop at the Hackaday | Belgrade conference.

As with any project, the first step is a good googling. By typing ‘Do It Yourself Laptop’ into the Mother Brain, you get back a list of projects that are ugly and certainly not as convenient as even the largest MacBook Pro. This build from 2007 features a Mini-ITX motherboard, with all the heat and power commensurate with this odd hardware choice. You could go old school, but again this is a project to build a usable laptop. Turning to the world of Raspberry Pis, these laptops are chunky, and this one is made out of a Little Caesars box.

To be fair, not all DIY, Open Source laptops are underpowered or covered in Crazy Bread crumbs. [Bunnie]’s Novena laptop is a work of electronic art, very capable, and a laptop that could conceivably make sense as a daily driver. The Novena isn’t exactly a convenient laptop, though: it’s rather chunky, the display is on the outside, and it doesn’t have the grace of a MacBook Air or Thinkpad Carbon X1.

[Tsvetan]’s laptop combines all of these features, is lightweight, has a battery big enough to run all day, is elegant, stylish, and above all, open source. It’s a laptop that will be modular, easy to repair and upgrade, and has spare parts that are cheap.

allwinner boardTo that end, [Tsvetan] spent four months building a powerful ARM board based on a 64-bit Allwinner chip. There’s HDMI, USB, output for an LCD, and this was made in KiCad, making the schematics and board files completely open source.

There’s more to it than just an ARM single board computer running Linux to this laptop. Thanks to the recent advances of an Open Source Toolchain for the Lattice iCE40 FPGA, [Tsvetan] is thinking about making this open source laptop a true hacker’s tool, with a logic analyzer, DSO, and DDFS generator.

What’s the future of [Tsvetan]’s open source laptop? He should – hopefully – have a prototype running at TuxCon, held in Plovdiv, Bulgaria in early July. There’s still a lot of work to do, but so far [Tsvetan] has contracted a plastics manufacturer for the case, found a nice eDP panel, small camera, speakers, a battery that will last at least six hours, and has a keyboard and touchpad working.

Building a laptop from scratch is among the hardest electronic manufacturing challenges we can imagine. It’s a true full-stack development job, with skillsets ranging from writing custom drivers for Linux to the intricacies of injection molding. With any luck, [Tsvetan] and Olimex will have some kits ready by this September. That’s an amazingly fast development time, and wouldn’t be possible without putting all the design files up on the Internet from the start, allowing other enthusiasts to contribute to the creation of an Open Source laptop.

71 thoughts on “The Open Source Hacker’s Laptop

  1. Having worked in computer repair for a number of years, laptops aren’t really hard to work on or really disposable. It’s like anything you take care of it (which the vast majority of people don’t) and it lasts. Yes some things do break, but usually with good cause. Beyond that they have certain failure margins just like any other engineered device, so it’s accepted and most of those issues can be easily fixed, if you research them. You learn how to work on it and it’s not hard, sometimes not cost effective, but not hard.

    1. I strongly agree with this. The only unrepairable laptops were the ones that simply weren’t worth replacing. I’ve repaired dozens of hinges, screens, keyboards, even mainboards. It’s all in how much money you want to put into them. It isn’t always cost effective though, but then again that’s subjective to each user. Nevertheless, I applaud the efforts being made, this looks like a great project :)

      1. Cost Benefit Ratio? Time is still money.
        I can buy late model HP and Dell laptops with Intel i5 CPU for under $149.00 (Brick and Mortar)
        I just about stopped doing anything except for the always dead battery and perhaps a new HD.
        You too can do the same. Lee at “Linktronics” sells from his store and ebay/Amazon.
        Yes, I used to repair everything, but time marches on guys..

    2. Yes those that pick up a Laptop with 1 hand from a the right corner and bend the motherboard and can’t work out why it randomly works SNAPPING THE BGA from the Many times it’s the on board memory or Display Chip separating just enough, By this time the laptop has a permanent bend in it and no matter how many times you reflow it nope after a day.

    3. I have to disagree. Two of my Laptops died. One because its GPU died of heat issues, the other one because the BIOS chip broke. I fail to see what “good cause” was involved there…
      Normal users and computer repair people are usually only able to “swap the mainboard”, which is time consuming and so costly that repairing disqualifies as an option. Stuff just isn’t built in a repairable way anymore.

      1. Quite frankly, the good cause on the first was poor design. I’m guessing a HP DV6000 series? The second- the chip *broke*? We’re either back to bad design or physical abuse of some weird sort. That’s an odd one.

        1. ICs can fail too. But replacing the BIOS chip on a normal notebook computer is trivial with the right tools – like a hot air rework station. Often a replacement chip can be bought pre-programmed from eBay or similar places.

  2. Awesome, you could easily make the battery life an order of magnitude better than any other laptop. The main reason I don’t have one.

    The other reason is it is so much weaker than a desktop, however giving it its own niche with the fpga will probably seal the deal.

    That and Open Source, god speed!

    1. godspeed is dated and one word and said to wish good luck before a journey.

      But I guess there probably is some kind of speed in the drugs world called ‘god speed’, so perhaps you are referring to that.

    2. Well, once the original battery dies, then the laptop’s RAM and hard drive capacity are no longer larger enough for the current bloatware from Redmond. And higher screen resolution, now with USB 3.1 or whatever on the horizon…
      I have a 1992 Toshiba laptop, but the floppy drive died, so I can’t load Linux or even drivers for a larger hard drive.
      If I get around to it, I might turn it into a digital picture frame.

    3. Totally agree that adding an FPGA to the mix will really help it carve out a market niche. The biggest iCE40 the open source tools support isn’t particularly big or fast, but it’s enough to do tasks no microprocessor can. A key bit of HDL to provide would be a simple to use configurable interface between user-space and FPGA logic. A pile of FIFO’s or a memory bus that can drive one side of a block-ram would be a great start.

  3. My only problem with ARM laptops is that you pretty much have to run Linux on it.

    Even if I like to DIY my machine, I don’t really care for DIY:ing and fixing and tinkering with the operating system as well. That’s also why Chromebooks and the like are not really interesting: you get nifty hardware, but the software is not.

    1. Here’s my 2 cents on the various ARM as a PC attempts, they fall short on the concept of platform. Intel has been very effective in killing off their competition, but there was a day when I had a choice between 3 CPU manufacturers and a number of chipsets to build a PC. I could expect a variety of OSes like Windows, varius flavors of DOS, various versions of Linux, OS2, and a few other more ‘exotic’ OSes to at lest get to a booloader and/or installer. 32bit OSes were where things for a bit tricky and device drivers were an issue but you could still have a mostly working PC. This happened because the PC is a platform.
      I’ve been waiting for the ARM world to get a clue, but they seem to be happy being embedded solutions and even promoting a lot of stovepipe style development. Software enumerable buses, firmware based hardware config reporting, a standard API for IPL source detection and image loading, and so on. These types of things allow an OS to be decoupled from the specific hardware and I hope something like this will lead to ARM based PCs that can compete with Intel and facilitate more open source designs like this.

      1. That’s because the ARM CPU itself is a mix & match solution. There isn’t going to be a reasonable standard core that everyone could depend on, because the whole point is that you can specify which bits and bobs you actually need.

        The x86 platform almost got fragmented with MMX and SSE and 3DNow! etc. etc. special extensions, and the software had to start implementing everything from everyone to make things work. Now imagine if instead of the Intel/AMD duopoly, everyone could specify what they want from an x86 processor out of a list of dozens of special feature sets.

        It would require some clever middleware and high level abstraction just to have an operating system that works on a subset of those processors. Lets’ say… Dalvik/ART.

        1. All x86 chips could still run the standard 16 bit and 32 bit instructions. 3DNow, MMX, and SSE are processor extensions, not peripherals sitting on a bus.
          If I’m an OS on a PC, I can rely on being able to, at the very least, use IO ports CF8H and CFCh to scan the PCI bus. For PCIe, I can use an MMIO config mechanism. Once I access this ‘config space’, it’s all laid out in the spec. The bus can also be reprogrammed by the OS. Platform firmware can do a partial initialization of it and let the OS handle the rest, or the OS can choose to reconfigure the bus.
          To find a device on an ARM bus, I might be able to use the flattened device tree if it’s provided, but otherwise, the device space needs to be hard coded into the OS.

          This abstraction of hardware and software is something I’ve been hoping the open hardware movement would pick up on and build an open source alternative to PCI that solves the problem of being software enumerable as well as being open source. But I guess this is a problem that only comes up in certain circles even though it is critical for allowing the PC to be open to various OSes.

      2. I think the No. 1. issue on ARM based computers is the graphics driver. Its all blobs.
        Only recently the Raspberry Pi’s dated gpu got opensource opening the doors to real X11 acceleration, and true openGL (not opengl es).

      3. Intel’s rapidly heading in the same direction with their latest generations of chips. That’s why Skylake and later don’t support any version of Windows older than 10; the power management requires tight integration between nominally seperate devices and a lot of driver support, they’re gradually dropping legacy interfaces to hardware, etc. A lot of modern laptops connect their input devices over SPI or I2C too rather than USB or PS/2, requiring special drivers and firmware tables to identify how they’re connected.

  4. I’ve been using a Chromebook C710 with custom Coreboot and various Linux distros for a few years now as my personal travel machine. Because it’s Intel Sandy Bridge based, it has a few binary blobs that are required (ME and memory init) so it’s not totally open, and of course it’s not OSHW. But being able to fix and build my own BIOS is great, and I’ve even contributed a fix for an ACPI suspend bug back to Coreboot.

    If you can live without OSHW, there are Chromebooks that are even closer to open. The venerable Samsung Series 3 Chromebook has everything buildable from scratch, including U-Boot and the EC. The only exception is the ARM Mali graphics.

    1. I could be wrong, but I believe the pile of laptops are those that have failed over several years of time. The only mention of failure rate I recall being spoken is every couple of years. That may be an affordable failure rate for many, but not for all.

      1. From the HaD post
        “….Every year, a few of [Tsvetan]’s laptops die, and the batteries of the rest lose capacity among other wear and tear….”
        About 3min into his speech, he goes into all their failure modes. The first couple sound like he’s just hard on his equipment (drops on floor, spilled drinks). Some other ones sound like he just ‘needs’ the top of the line laptops.

        Maybe if you spend a lot of time developing large code bases you need an upgradable laptop to follow software resource demand. But I think he’s using them outside of their design specs. ” I want a battery that lasts all day, is light, and is powerful” Sorry, batteries aren’t quite there yet. Learn to like a desktop, or use a netbook to log into a remote server for your process intensive operations.

          1. Providing alternate solutions to the same problem, with arguably less effort.

            ARM is great but it doesn’t compete with x86-64 and likely won’t for quite a few generations.

    1. service manuals and replacement parts exists for virtually every make and model of laptop, at one time that would all have been hidden behind the local authorized repairman or reseller but today it is fairly easy to find.

      open sourcing the hardware for something like a model of laptop would mean giving access to the design files or similar of everything in the machine that wasn’t bought off the shelf.

    2. Same. I love my M4700. Mine is 5 or 6 years old and still going strong.
      I also love how Dell give you the option of shipping you the replacement parts if you feel competent enough to replace them yourself so you don’t have to wait on a tech. (Unless you have next day repair agreements).

      Robust, powerful and so easy to work on and repair if needed.

      Now, Fujitsu on the other hand, ugh. Terrible driver support, almost no updates especially so for older models, proprietry drivers so you can’t use drivers from the OEM, low quality construction and loaded to the brim with bloatware.

      1. I must’ve lucked out, b/c my fujitsu has been pretty good to me, even w/ suspected malware. Had a little issue w/ mouse scrolling but got it resolved thankfully. I can see how it’d be cheap construction but the screen joints have held up well for me.

      2. Fujitsu’s drivers and support are a pain!

        I’ve been very happy with the hardware though. My P2110 is still running strong and my friends nh570 has been rock solid. Once I manage to get the things running they just don’t die.

    3. No most manufacturers have neither. Sometimes the service manual of some model get leaked and replacement parts may be available from disassembled machines but officially only the service technicians of the brand have access to those parts.
      Dell is better than most given the availability of a public disassembly /assembly instruction but unofficially Clevo based machines (available rebranded under a huge amounts of names all over the world) are the best: almost all models have leaked service manuals due to the distributed service model (each rebrand tend to have their own service center for simple faults while Clevo fixes the more complicated cases).

      And their manuals are true service manuals – not only are there disassembly/assembly instructions but there are schematics and other details required to do more complicated debugging.

  5. This sounds like a great project but I was really disappointing by his equating of people demanding open source GPU drivers with people demanding capacitor electrolyte fluid formulas, etc..

    If good, powerful GPUs with open source drivers just aren’t available I get that. Use what works best for your requirements in your design. However I think open source drivers are something one can make a reasonable case for wanting even if not everyone cares. I’m typing this on a 9 y/o laptop myself. I wouldn’t think of running Windows 10 on it, that would be way too slow if it works at all. It was built for XP. However.. running Linux it serves my needs well enough. Spending $100s on new equipment would have no justification for my use case. Except… the audio driver really sucks. If I try to listen to the built in sound it is raspy like there is some amplifier inside that is over driven yet the audio coming out the speakers is quiet and hard to understand unless I stay absolutely still. I’ve messed around with it quite a bit and no luck. So.. I’ve resorted to using my bluetooth earpiece that I also use with my phone. No problems that way.

    So.. anyway… if I buy this open source laptop can I expect the GPU maker to keep releasing new drivers to stay compatible with new Linux kernels, X or “Wayland/Mir”(yuck) versions? Or does that mean choosing between throwing away the motherboard every couple of years or being stuck on old software?

    If there are no good GPUs available with open drivers then I get it. Use what is available and I’ll be glad that the laptop is at least open in every other reasonable way. Comparing a desire for open drivers to a desire for electrolytic fluid formulas or chip die information, etc… He’s being kind of an Ass!

    1. Hi
      I’m Tsvetan from Olimex.
      English is not my native language, so I may didn’t express myself correctly in the slides, sorry for this.
      I didn’t want to underestimate the need for open GPU drivers at all. I know this is important, but what I wanted to say is that when one open source project is made the authors make it as open as they can. I have seen on the internet statements like: this is not open source hardware because it have closed source software drivers, which is not true.
      If there is schematic and PCB layout and all components are generic and orderable then the design is OSHW as per the definition.
      If someone demands to know the orbits of the electrons in the components used inside the project he is free to do something about it and to share with others the results.
      The GPU drivers is pain topic for FOSS community and the only pure open source GPU driver (partially working) I know and have seen was the LIMA driver few years ago presented by Libv on FOSDEM which kinda performed even better than original ARM MALI drivers. Unfortunately it seems this project is at dead end for several years. If I remember correctly ARM lawers sent him threats for doing reverse engineering of their devices and no one from FOSS community organizations backed him or made statement o support him on this issue, so he was disappointed and stopped working on it.

    2. Except you pretty much never want to do anything with the capacitors.

      Contrast the GPU, which is a great parallel compute engine, and for which the development of open-source software would be a great boon to basically everyone.

      TL;DR: apples, oranges. A capacitor is a part. A GPU is a tool.

      1. I don’t think that what Tsvetan says is so difficult to understand. What he means (or at least that’s what I think) is that all his design is open source hardware, but of course that’s not the case for the parts that he use, neither that all the functionality of all the parts is usable with open source software. Some compromise needs to be taken: not HDL code for the ICs, not technology files for the IC layouts… the thing about capacitors just means that this can go all way down to capacitor’s formulas and copper alloys.

        The important thing is that all _his_ design is open source and the parts used have a good compromise of availability in small quantities, price, documentation and open source software support. Every SOC with GPU on the market today have some open source software support issues.

  6. I’m surprised by the number of comments basically saying “my n-year-old laptop is good enough”. Seriously? Does it have USB 3.1? Thunderbolt? Do you really have no problem with spending four times as long to transfer files as your richer friends with newer laptops? The unique aspect here is upgradeability. Name ANY other laptop that has had an upgrade path for something beyond memory and hard drive size.

    Of course, it’s going to be a huge challenge to keep the same form factor. Every few years the laptop makers decree that a new size and aspect ratio of display is what the world needs, and once that happens, everything else gets the short path to obsolescence. Which means that if you can get new LCD panels in an “obsolete” size, they will be very expensive. So at the very least, the case will have to be replaced as new displays become standard.

    I’ll be keeping my eye on this one.

    1. “upgradability” means sockets all around and that means everything has to be bigger.
      A lot of the older laptops had socketed CPUs, but it usually made no sense to upgrade them, as it just wasn’t cost effective. (been there, done that).
      There was also a large number of laptops with discrete GPU cards that could be user-replaced. Again, upgrading was not cost effective and there were compatibility issues.
      I even remember the modular laptops that Intel tried to standardize. Obviously, it did not catch on.
      It will be interesting to see if the external GPU interface ever becomes standardised.

      But doe to the fact that it’s becoming impossible to maintain Moore’s law, we may see a lot of older laptops still having the crunching power to keep up with modern demands, so there may be another large attempt at a modular laptop.

      1. Upgrading by use of sockets is old school. When a whole motherboard is under $100, you don’t use sockets. Sockets are unreliable and bulky. What I’m talking about is, how many consumer laptop computers have an upgrade path for the motherboard? We’ve had that for desktop machines forever, why not laptops?

      2. I replaced my hdd to an ssd.
        I replaced my wifi card (which was a utter crap) to a better wifi card.
        I upgraded my ram to be 16GB (instead of 8).

        And this was a 1500+ USD laptop back in 2014.

        On some today’s laptop the ssd is soldered, the wifi is soldered, even the ram is soldered onboard.
        And yes I’m talking about highend laptops…
        (I change my laptops 4+ years, I use it for works. It is really hard to get used to a new laptop…)

    2. Do the newer laptops have support for open source disk encryption? Also on a pure dev PC I have no use for microphone, speaker, or camera usually; much easier to remove on older laptops. Also a trend I *really* hate is not having CDROM’s, I like booting live, I do it for vast majority of web surfing.

      1. Loading a live distro to a USB drive instead of a CD is too difficult eh? If you picked up a sandisk cruzer fit or similar you can just leave it plugged in all the time and modify grub on the USB to boot the hard disk as well.

        1. No it’s not, I actually use a USB-CDROM ASUS thing for a PC that lost it’s harddrive and it’s internal CDROM lol. No driver issues, I lucked out. Doing that on a laptop is very clunky and gross. And it’s adding more things to go wrong (conversion of comms etc.), and badUSB was/is a big concern for me (I have had quite a few bad experiences w/ USB malware).

    3. My daily workhorse is a Lenovo T61 from 2009.

      The hard drive has been replaced with a SSD, and as the backlit died in the screen, I replaced it with a new screen, and I could without any modifications increase the resolution from 1680 x 1050 to 1920 x 1200.

      Of course the battery time is low (approx 2 hours for a 6 cell battery) compared to newer computers, but the batteries are dirt cheap.

      But the most import thing is, that the performance is, despite the age of the computer, still absolutely good – I am using Linux Mint 17.3, and am having no problems with 8 concurrent desktops, on which I have scattered:

      2-3 browsers, 2-3 development ide’s, 1 or 2 texteditors, SQL admin, mailprogram, 2-3 terminal windows, and what ever is needed.

      AND the interesting thing is, that the machine is much faster than a brand new Toshiba I got my hands on, even though the specifications on the paper claims that the Toshiba should have 2-3 times more horsepowers.

      Also all repair manuals are free to download from Lenovo.

    4. Yes, seriously. I’m typing this on a ten year old single-core ACER laptop and it’s doing what it’s supposed to just fine. I use it for eight hours daily. Yes the battery is pretty much dead, but then again this doesn’t get turned on on-the-go and it’s still good enough for a graceful shutdown on power loss. I could buy a new battery at any time, there’s just no point. USB3 would be nice, but inconsequential – I can’t remember when was the last time I plugged an actual pen drive into it; the Ethernet jack on the back works just fine than you (so does the WiFi). Doesn’t have a camera, I don’t have a use for one either – I’m not Pewdiepie. The lube in the hinges seized up at some point and the case broke at the screw posts; so I got a used bottom and topcase, drilled out half of the stiffeners from the hinges and moved everything over into the new case – yes, it still works just fine and yes, it actually still looks almost like new….

    5. My daily use computer is an IBM ThinkPad X31 from 2004. 512MB of RAM and 40G harddrive is enough for me. Ethernet works great but internal WiFi doesn’t connect to my school’s network. I simply don’t need USB3 or gigabit ethernet. And I have a Nokia 3310 and no smartphone. I also have an even older X31 which I used before but it has two dead capacitors in the internal power supply and I haven’t yet got the replacements.

  7. No need for Thunderbolt, no need for USB 3. I’m organised enough to have no need to copy files around other than for backup. My current main project builds in about 3 seconds, what’s the point of bringing this down to 1 second? Backup takes 10 minutes unattended, what’s the point of bringing this down to 2 minutes if nobody has to wait for anything anyways?

    Over the years I developed a simple rule of thumb: something (disk, screen, whatever) has to be ten (10!) times better than before to be worth upgrading. 128 MB disks were replaced with 1 GB ones, then 10 GB, then 250 GB, and so on. Doubling performance gives just a “it feels a bit snappier”, which isn’t enough to justify replacing an already well working system.

    Replacing a system built up and fine tuned over many years is easily a weeks or even a months work, so one has to think twice before upgrading.

    1. Everybody has different requirements. I need something I can stay up way late with, doing mostly text work but also schematic and PCB layout and some graphic design, with no power available, AND I need to be able to transfer video from full 32 GB SD cards to SSD when on a shoot. But this is where open source comes in: if enough people want a feature that Tsvetan chooses not to incorporate, then somebody can take the initiative to spin that into an alternative main board. I realize that this rarely happens in real life, but it’s still a possibility. Someone on Olimex’s blog asked how long Olimex plans to support this. Wrong question — how long will the COMMUNITY support it, because that’s what really matters, and THAT depends on how useful this machine is to a large number of people. This to me is Olimex’s greatest challenge: to make a laptop what will meet most people’s needs without being bloated to try to satisfy everybody.

        1. Since this is Hackaday, I would be remiss if I didn’t mention Arduino. Arduino (both of them) could close shop tomorrow and the Arduino platform would not die. Most of the accessory hardware that works with Arduino was not developed by Arduino, and that’s hardware, which unavoidably costs money. I see the possibility of a similar situation with the Olimex laptop.

  8. found the original article here: https://olimex.wordpress.com/2015/12/01/help-day-recap/

    The physical size is 29 x 20 x 1 cm, less then 1kg. It means 13″ screen size.

    For me a laptop is a 15″ machine and below 2kg. Otherwise I have a smartphone which is 5.5″.
    Do not see the need a yet another screen size between 5.5″ and 15″.
    (The company I work for gave me an ipad 10.1″. I used twice maybe…)

    On the video he says the price is about 200EUR, and the motherboard is about 30EUR.
    Upgradability is like changing the motherboard. No mentioning about sockets at all.

    1. 13″ has always been my ideal laptop screen size, and lighter is always better, as long as battery life is sufficient. I’ve been forced into 15″ screens by the industry, and they just make the overall machine too big. I’ve long suspected that they bumped up the screen size just to have more area for the motherboard. What Tsvetan describes is what I’ve always wanted. I almost had it with an old, old Toshiba 13″, but that was in the USB 1.2 days, and of course the motherboard wasn’t upgradeable, so an external hard drive was not even worth trying. The only other real limit there was that battery life was still only about 4 hours with heavy use, but that’s the most I’ve ever had.

      I’m sure Tsvetan has had to both temper his own requirements and also make decisions based on conflicting feedback from commenters on his initial design. You can’t make everybody happy. Some people seem to need optical drives even though they’re big. Some people need lots of battery, even though it’s big and heavy. Some people need maximum CPU power because they’re using applications (like video editors) that just can’t get enough, even though this reduces battery life. Some need a big screen, even though it’s too big for others. Some think (like Microsoft) that touchscreens are essential. It’s hard to imagine making a laptop that’s going to make everybody happy, and I predict he’ll have to make multiple models to cover the range, hopefully while maintaining a motherboard form factor that will work for all.

  9. also found the original renderings: https://olimex.wordpress.com/2015/11/05/a64-olinuxino-oshw-64-bit-arm-diy-laptop-idea-update/

    It looks pretty much like a samsung ATIV book 9 lite laptop:
    http://www.notebookcheck.net/fileadmin/Notebooks/Samsung/ATIV_Book_9_Lite_905S3G-K01DE/ATIV_9_Lite_Teaser.png

    What I do not get is, why the hack is an SD card reader is needed? I had sd card readers in my laptops, never ever was a reliable reader. Alsways driver problems under linux. Failed to mount many times.

    I would really like to see dedicated volume buttons, and preferably a scroll wheel on the front edge, which can be operated when the lid is closed. Also same goes to its leds.
    Be able to see if the laptop is in suspended state when the lid is closed (led on the edge).
    (ie. no need to open it just to see if it is sleeping).

    Don’t know about charging ports, but a magsafe type would be awesome.
    Otherwise just make sure you can throw anything to it between 15-25V.
    And have many type of adapters. I really hate carrying laptop chargers around (just kills the weight, and usually bulky).

    If it could be charged (really slowly, overnight, camping, etc) from an USB port, that would be a definite plus.

    The touchpad seems small, and the dedicated buttons are oldscool. I prefer a big-ass touchpad, macbook style.

    The screen seems glossy. I highly prefer matte screens.
    My current laptop (Samsung Ativ book 8, 15″, 1.98kg) is matt. I had glossy screens before, I hated very much.
    (it’s usefuly while shaving, but it is really a rare case to justify glossy screen. I only did twice with my old laptop )

    All in all, it very much looks like a netbook. I think we are after the age of netbook…

    1. Using Ubuntu from 10.xx versions I never had any problems with SD readers in three different laptops, and I actually use them – to get pictures from SDs, format them, or create a RPI image.

      Reading further, all in all it seems that you do not want a hacker laptop. So really stay with your mac…

      1. my main problem with the current sd card reader (apart of being unreliable), that microsd card do not fit without adapter. The normal sd card is sticking out when plugged in. So if I really want to use it, I need to carry
        a normal sd card adapter. Carrying an sd card reader is just as much trouble, and not more.
        Both are tiny.

        > So really stay with your mac…

        I try to stay a civilian, but your reading skill faces some serious challenges.

        The comment to which you replied explicitly stated that I own a !apple hardware.
        In fact I never owned an apple product, I only used them at work,
        I have an iPad, but it is company owned, not mine.
        (although I do not really use it at all)

        Buying laptop hardware is a bit hit and miss, even if you buy (research) for linux in mind explicitly,
        and you do not care for the price either.
        I only used linux, never installed on my hardware either windows or macos.
        Although I used them both (at uni and work).

        Still builtin SD card readers are more hassle then good imho.

  10. This laptop might just lack horse power and without that it will be only another toy. I feel that if I had to I can add tools to my laptop, even if I’ve to create them, more than that that: these tools will probably work on any new popular Linux distribution I install wherever I go. But I can not imagine programming, or designing anything on weak laptop.

  11. I bought an OlinuXino back when everybody was singing Raspberry Pi’s praises. Why? Because Olimex chose parts that could be bought in small quantities and which were available in NON-BGA versions, and therefore I knew that I could use this for development, then either buy the board or modify and build it for specific projects. Nobody in their right mind just copies a design that already works, especially when the price is reasonable, and I think Olimex understands this, and this is why they aren’t afraid of OSH. The main issue with ARM, as I see it, is that there is no basic ARM microprocessor. Everything is an ARM core with whatever interfaces and peripherals a given manufacturer chooses to add to it. If you start with a Raspberry Pi and need to make modifications, you basically have to start from scratch because you can’t GET the Broadcom SOC it uses.

    It really sucks about the closed GPU thing that seems to be the rule in the ARM world. I guess nobody wants to give away their design when they have to pay for ARM’s design in order to produce a chip. But come on now – at least ARM gives you full manuals for their designs. The least the SOC makers should do is provide an SDK for their chips.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.