Ask Hackaday: Can We Do Better Than Phonebloks?

Our tips line is on fire with suggestions for us to cover the modular cell phone concept named Phonebloks. The phone’s designer states the problem as follows:

A phone only lasts a couple of years before it breaks or becomes obsolete. Although it’s often just one part which killed it we throw everything away since it’s almost impossible to repair or upgrade.

His solution is the above pictured phone, with modular components for each feature: wifi, camera, battery, etc. Rather than upgrade your entire phone, upgrade just the parts you need. A wave of followers have thrown their support behind this concept, and perhaps their hearts are in the right place hoping to reduce waste and cost. Behind the scenes here at Hackaday, however, the response has been a unanimous facepalm. The primary objection (other than design implausibilities) should be obvious: dividing the phone into exchangeable bits does not inherently reduce waste. Those bits have to go somewhere.

Now, don’t rush to the comments section to identify additional problems; there’s a juicy Reddit thread for that. Instead, we want to take the high road: Can we do better? Can we make a phone for the future that is less wasteful to produce, more easily recycled, and possibly upgradable? What would be included in its features, and how would we do it? Check out a video of the concept phone and tell us your alternatives after the break.

221 thoughts on “Ask Hackaday: Can We Do Better Than Phonebloks?

          1. >>>
            If I can do it so can you.
            >>>
            With that lie in mind, I want to see you do the Jordan.

            What’s the Jordan? You ask.

            That same question came to mind when I heard someone repeating that statement after a youth meeting as I was leaving. Being naturally curious about such oddities, I turned around and looked in the direction the voice was coming from, and this is what I witnessed.

            A teenage African American jumped straight up at the free throw line and proceeded horizontally toward the basket and dunked the ball before descending and disappearing into the crowd below.

            Since you can do it because he did it, I’d like to see you do it.

        1. The point is not that you should refrain from building a robot.
          The point is not expecting others to be able to do something you were able to accomplish.
          There’s no fault in pushing one’s self. There is a great deal of fault in pushing someone beyond their limitations for no other reason than thinking there is nothing wrong with them.
          They may have limitations that are not so obvious as a non-athlete trying to dunk like Micheal Jordan during his prime.
          It’s just not going to happen without the same dedication and work Jordan put into it.

    1. I would add, they don’t really undestand the technology market either.

      Take desktop PCs for example. You can upgrade them, but half the time you actually can’t because the technology has moved onwards and your old parts no longer support the new parts even if they do fit together, or they’ve become so obsolete that you’ll be replacing 95% of the machine anyways.

      1. The market is indeed rigged against making upgrading older hardware viable compared to replacing it. Often even if you can find parts that fit they end up costing more than newer, better parts do.

        1. I partially agree with ya’ll. I use old hardware all the time. The only reason I’d need new hardware is for games but I only play games on xbox. I think replaceable parts are great. The only time I replace parts is when something fails or I need a minor upgrade like ram, psu, wifi dongle, keyboard or mouse.

      2. If the rules are BS (plannes obsolescence as economy engine), change the rules (the whole market and engineering process). Better now (tomorrow/15years) than never. One more example: if your market’s beef has tons of toxic’s, find another (green) provider and eat less meat.

    2. Actually, it was designed by just one person, who was only voicing an idea with a slick video.

      The whole point of this was just to get people more interested in a modular, more PC-like smartphone, maybe entice a manufacturer to put something like that on the market. Nothing more and nothing less.

  1. Good luck writing a stable software stack able to support that modular design.. Linux is pretty cross-platform and modular in nature but in reality Linux and Android are finely tuned by the vendors for each individual phone.

    1. I would believe this comment more if there wasn’t an extremely popular open build of Android that runs on tons of phones.

      Also I’d believe it more if the vendor software for many phones didn’t suck horribly.

      1. If you’re refering to cyanogen each build has files catered for each phone. The kernel and drivers. That’s why there’s seperate downloads for each phone. If it were platform agnostic there would be a single image.

        1. Still, most, I guess, of the software is universal. Same with the little Android boxes you get for TVs. Nobody’s ever written an operating system that doesn’t need drivers, unless you stick to one rigid hardware design, you’re going to need them. That’s not a problem at all. Linux has got drivers sorted out, the tools to handle it are capable enough.

          I think the only way this might possibly work is a passive backplane. Although I believe they tend to be slower. Perhaps then a multi-lane bus system like PCI-express. Lots of individual serial channels, devices enumerate their capabilities. Works well on USB. So a hybrid between the two, or even just a USB-3 hub. Invent a quick standard to control and monitor the battery. An LCD could slip onto the front, a couple of screws would make it sturdy. Perhaps have a few size options, including screens with keypads under them.

          The GFX could plug into the USB for data and power, and output to LVDS for the LCD.

          It’s nice we have so many versatile standards. Add a couple more, for power control, low-power modes for everything. It could all be built on top of what’s already available.

          I dunno if it’d be more or less expensive than the way we do it now. I’d guess more. Also more problems, less reliability. You’d need a strong case with a reliable way of mounting the bits to not fall off.

          1. Most of the software runs on top of a virtual machine. That’s the point of the whole Android operating system.

            The problem is still the bottom layer where you need a kernel and drivers to support a myriad of different hardware configurations with backwards and forwards compatibility.

          2. Way back in the day when PCs were 16 bit, devices carried their drivers around on something called an option ROM.
            Maybe UEFI can handle this, or maybe each device could support a mass storage device supporting a standard interface (like USB) that the OS can fetch the driver from.

          3. Drivers are manageable. You just need an architecture and some standards. Even Windows, gods help us, managed to have drivers that worked over several versions of the OS. Android seems to be doing fine, and so does USB. Most modern USB stuff doesn’t even need drivers for a PC, they stick to the standard for whatever they are, and they work with the drivers the system comes with.

            Am I right in thinking Android’s driver system is modular, like the Linux it’s based on? Managing and taming drivers, and compatibility, has been solved now. Not only is it no longer a nightmare, you don’t even need driver disks. And since mobile phones are pretty much always online, it’s not hard for them to download them if they’re needed.

            Whatever these modules end up being, they’re still gonna be using the standard chips, whatever they end up being, and they’re manufactured in their millions. Think how well PC graphics cards work, extremely complex. But 9 times out of 10, you plug it in and it works.

            Driver management has been solved. It’s no longer a big problem. Worst comes to worst, take it back to the shop and have the guy there fix it. Or get your money back.

          4. Drivers are just busy work. If drivers were the barrier, we should all be out of a job. :) Thankfully, we’re talking about a pretty limited hardware set here. So the numbers and sophistication of drivers we’re talking about (maybe 20, 30 max, configurable / selectable peripherals?) are all very accessible. The wifi, bluetooth, baseband, etc. will all be done once. No need for changes there. Kernel stays the same. That stuff changes at tectonic speed.

            The things we might want to change (screen, camera, gyro, mic, speaker, etc)….All of that is academic. Talking days and weeks not months or years. And because this is on Linux / Android, you’re likely to attract an interesting community to rally around it.

            I’m not sure if it’s a money-maker but it’s a cool idea nonetheless. I’m not even sure if I care that I throw my phone away after 3 years because I can’t even give it away…But give me a phone I can build modular hardware for and connect??? Oh hell yes!

            Chinese opensource HW phone project — http://goo.gl/NhFUeq (not meant to compete with this project but rather, think of the potential for developing communities)

          1. It’s not like each phone’s source code is re-written from scratch tho is it? It’s a matter of config files, what options you add in and what you don’t. Tweaking. As in most OSes, drivers meet their obligations of exposing functions to the world, and do the sort of thing you’d expect them to when you call those functions.

            PCs are just the same, but you can build one yourself from any parts you like without needing to get your compiler out. It’s modular, and there’s a system in place to make it friendly and hygienic.

          2. @Greenaum They’re not rewritten from scratch for each device but it is a hell of a lot more involved than just select a couple options and pressing build. If you look at the various Android phones, you’ll find lots that are stuck forever at an old version because the drivers and APIs changed so much that it just isn’t possible to update.

            The main cause of this is lack of suitable developers with the right skillset to rewrite the drivers and the manufacturers not supplying source code for necessary modules. The Android OS isn’t fully open source as a large amount of it is still closed source by Google or the manufacturer of the SoC etc…

            The WiFi and camera drivers are two of the biggest roadblocks to updating an Android device as the APIs and drivers have changed so much.

            Go look at the build process for any specific phone in Cyanogen or AOKP. You’ll find that it merges in a load of binary modules together with of AOSP.

          3. Android is most definitely fully open source. Just the same like Linux is fully open source. What isn’t fully open source are the drivers for individual bits of hardware and the Google Apps which aren’t part of the AOSP build by default.

            Just because you can’t run it out of the box on your hardware without extra bits does not make the OS any less open source. Drivers and apps are NOT the OS.

          4. Yet when I download a Linux kernel lo and behold I get drivers in the tarball. How can this be? A computer wouldn’t even boot up without drivers. You wouldn’t even get to a command prompt without drivers! Only a total retard would say that drivers aren’t part of the OS in fact. The fact that some drivers are third party software doesn’t change this at all.

    2. Easy solution, since each block is exclusively independent, then simply have the code and driver software for each on modular and built into the firmware and use a standardize communications medium between them all.

      There I solved your issue for you.

      I thought you guys were smart…

        1. Except they don’t.

          The memory has a separate bus from the rest of the board, whereas in this suggestion the memory is a separate block from the processor. That’s like putting your RAM on a PCI-e link in your PC, which would introduce a horrible latency.

          1. The RAM should probably stay in the same part as the CPU. Especially since it’s often mounted on top, or in the same package. But apart from that Hojo’s right. Even laptops have used standard modules and busses for a good while now. There was a brief period when SD cards started supporting hardware (SDIO), tho that died out once everything you could need ended up integrated into one chip. The earlier, mid-90s generation of PDAs used things like PCMCIA. Modularity in tiny computers is no big deal.

      1. Interconnect in FPGA that is explicitly intended to be reprogrammed as needed was done on several systems in late 90’s. Partitioning in systems like Sun Enterprise is done essentially by reprogramming backplane FPGAs to split backplane buses into smaller independent sections. Hardware design of some entry-level Cisco modular routers strongly suggests that changing module slots pinouts to support newer modules by reprogramming FPGA was original intention (althought it seems that it was not realized for bussiness reasons).

    1. FPGAs have horrible power consumption compared to dedicated silicon. If it’s purely digital you’ll win because you don’t have to have strong drivers on/off chip if you integrate multiple chips into an FPGA, but you would still be best off creating a custom chip.

      Short version: An FPGA would be nice, but you’ll give up battery life. Not a great compromise for a phone.

      1. You nailed it. FPGAs cannot achieve the levels of power or *performance* (A8 CPU may be clocking at ~1GHZ, running multiple cores, blah blah…good luck with an FPGA) that you need for a mobile phone. Just not possible. Interconnection, as hojo is saying is not an issue. If it’s SPI, it’s just another chip select. If it’s I2C, well that’s got it’s own scheme for managing that.

        As I commented earlier, there is no reasonable technical barrier here. It’s a mech-industrial problem, not an electronics one. And a problem of retraining consumers (and this may be the biggest challenge…I can see me loving it but my wife would probably laugh me out of the house and all the way back to the store I bought it from)

        There’s no real reason that modularity would not be feasible at this level!

      2. The long-term [flash] RAM would be seperate from the running [CPU, DDR] RAM. Oh, and if someone really wants an FPGA, it becomes another module. One that sucks battery juice like a man drowning in a whiskey vat.

  2. The simple answer is standards, just like desktop PCs. Standard touchscreen interface and connector (like vga and dvi), standard battery (like atx power supply), standard peripheral interfaces (usb, firewire…), standard case sizes, etc. Anyone can make a module that can plug into any device if there is a basic set of standards. Manufacturers won’t go for it because it eliminates their lock-in.

    1. Or they’ll start making their products compatible with a particular manufacturer. Imagine being stuck with the moniker “iPhone compatible mobile device” for the next thirty years.

    2. We need more standards! The standard standards just aren’t standard enough. :)

      But seriously, solving the hardware problem is going to be tricky. The latest phones are engineered as a single unit. How could you upgrade only the CPU unit of your device if the current generation is 32-bit, then the next is 64-bit. If there were standard protocols it would be workable, but I believe that when you black-box components you will have to sacrifice overall performance and size.

      1. As I mentioned in a comment later on in the thread – there are already a lot of standards that go into almost every mobile phone to make things modular (read up who MIPI are). Trying to get concensus as to what the standard should be and getting access to the right patents is very *very* hard (I’ve been on my share of standards organisations – it’s best described as “telling people to Foff in the politest way”)

    3. That was my thought as well. A lose set of standards around form factor and compatibility exactly like IBM eventually did with PC’s would be all you need.

      Imagine walking into your local technology super store and buying the latest case, screen, main board, battery, protector, processor upgrade, power management chip etc.

      I think there is huge potential here especially as the barrier to entry for custom pcb’s continues to fall.

        1. He means, you could fit 10 notebooks inside a desktop PC because of the dead space around the components, because each component is in its alloted volume of space instead of being fit around one another as tight as possible.

    4. What lock-in? Most phone manufacturers use the same components anyway. They’re getting towards a 1-chip design now, at least, 1 major chip, plus the flash and the RAM. Everything is integrated into the CPU.

      It’s the same with Android, manufacturers loved the idea of being able to share an OS, so that apps for smartphones would be cross-compatible. Means it’s economically easy to develop smartphone software beyond the simplest games, and stil get decent performance. If people are going to write serious software for phones, then the customers need serious phones to run it on. Which means expensive phones to start with, and an upgrade cycle. Gods help us if the phone upgrade cycle gets as bad as the PC one.

      Anyway all of this means customers wanting big expensive phones, which is good. Just before Android, there was quite a bit of competition at the very bottom end of the market, some phones being cheaper than pizzas. It was the only way to go, nobody would want a phone as big as they are now, unless they could get Facebook on it, watch Youtube, and do some web stuff.

      So yeah what I’m saying is it’s the same as with IBM-compatible PCs. Standardised software led to a mass-market with lots of hardware clones, which encouraged even more software dev which encouraged even more hardware clones. Since the beginning of mobile phones, nobody’s wanted lock-in on the phones. Network lock-in is an entirely different story.

  3. First of all, it has to be as open as a PC. (I mean, a pre-“Secure” Boot PC.) Easy software changeability, on its own, will extend a computer’s life. The lack of software upgrades is a key factor in smartphone obsolescence. And, since it’s the carriers who decide if you get an upgrade or not, we have a fox-and-henhouse situation.

    1. From the technical side yes. But I think the concept is good.

      I agree with this. I don’t mind keeping my hardware if it’s still working. But the lack of OTA updates on a lot of Android devices is frustrating for users.

      Think of it this way: you say my phone is two slow for the newest OS features? What if the processor could be swapped out for a faster one that also uses less energy? That would be fantastic for consumers.

      1. It’d be fantastic, but it also very likely wouldn’t be as thin, as low power, or as robust.

        Also I have to say that vendors who say that the CPU is too slow are probably full of it. They just don’t want to support it any more.

      2. I think you’d probably run into a scenario like you do with PCs. You can upgrade the processor to a certain limit determined by the support hardware, but by that time, you’re probably going to want to upgrade to a faster bus, memory, storage interface, etc. I don’t see processor as a good place to upgrade anyway. The interconnects on any modern high performance processor are formidable. Since most people aren’t going to be soldering on their own BGAs, that leaves sockets and card edge connectors. Look at the socket on a modern upgradeable PC CPU and decide if there’s room for that in a phone. I think a better idea is to have the processor soldered to the “backplane” and have room for expandable memory. When you want to upgrade the processor, you upgrade the backplane, so you get the faster processor and upgraded bus/clock/support hardware, and you can keep all your modules and maybe your memory.

    2. think about it this way. would you want a desktop computer as a phone?

      when you do this there will be a substantial size increase, quality decrease, wireless sensitivity decrease, and speed decrease.

    3. i agree with the open software bit, i have an old hp netbook, it’s like eight years old or something now, i’ve replaced the motherboard and the keyboard, and it’s generally a peice of shit. or it was. it has 1 megabyte of ram on it, and i had it running on windows vista, because that’s what it came with. last year i got sick of it being slow and impossible to work with so i switched it to lubuntu. all of the sudden a computer i was saving up money to replace became a computer worth keeping through the rest of highschool and possibly into college. now the only thing i’m worried about on it is the plastic case becoming too damaged to work. modular phones are a possibility as long as everyone that builds something for the phone builds it to be compatible with at least the last three versions of everything. you should be able to keep your phone at it’s most basic, OR upgrade stuff. you can’t sit there and keep everything exactly the same except for your graphics card or something. at some point you’ll have to upgrade the rest of everything in order to keep up. if it was like linux, where everything’s open source and you can rewrite your own stuff if you want to, then yes, it can work. what we want from a modular phone is something that will let us change what we want, and be as customizable as feasibly possible. what we want is something that we are going to have to make because no company with the infrastructure to start this is going to make a phone for us that they don’t have control over. with a modular cellphone we should be able to pick and choose what towers we need service from. if they were all satellite phones it would be easier, but that’s not something we can expect at an affordable price. when we ask for a modular phone we want a modular phone in hardware as well as software. what we want is the ability to choose what we get.

    1. This. Most people are happy with their phones and dread buying a new one. But the new ones have a snazzier screen or a faster cpu. If a standard could be written for each component then yes i do believe that waste would be reduced. I can’t see why not. Used bits could go for sale on their own or still good bits could be assembled into one working phone in developing countries. I have to say that i disagree with the article entirely. Alone the modularity would be worth it, waste or not.

      1. I think your argument is sound. You could see this as an improvement on the current system.

        My biggest problem with the concept video is when they show the different blocks moved to different places. I can’t image the complexity the base board would need to support that kind of thing. But like I said above, even just the ability to upgrade the processor (and nothing else) would be an amazing improvement.

        1. I think that part could be accomplished by simply aligning the pins correctly or using a standardized plug of some sort. So maybe it’s not a 4 little copper pins, but it’s something akin to micro USB and the thing they plug in to would essentially just be a usb hub. So you do a grid of ports and each block would have 1 real plug and 3 dummy plugs. The “brain module” might need to be in a specific spot, but otherwise I think it’d be fairly simple.

          I think this idea is impossible with todays technology, but it’s still fun to think about.

          1. There is that tech where you can send squiggabits over short, centimetre-scale distances, with no connection at all. Partly radio, partly capacitive, I dunno much about it. But certainly another capacitive thing, where chips connect to each other thru their insulating casing, using pads just under the surface capacitively, is an up-and-coming possibly next-big-thing.

            So you might not even need a bus for more than power. And you could do away with THAT with inductive power coupling (assuming it doesn’t break the capacitive data coupling!).

            So you could end up with a “phone” that looks like a bag full of Lego, not even connected to each other! Any old mechanical support would do to hold them together, even glue! And that’d be a fantastic way of upgrading.

            Perhaps you could leave old unwanted bits around the house. Your house-net will find them and add them to it’s own computing capacity, using the futuristic grid of photons and frequencies, that permeate your living space to the point where it’s an effort to wade your way through them to the other side of the room.

      2. The issue seems more complicated to me. I’ve sold each of my old phones–including the water damaged ones–so arguably there’s zero waste on my part despite numerous upgrades and broken devices. Unless there are statistics projecting how modular phones would be used (and how upgrades would be released) as opposed to current practices, I’m not sure we can have any confidence in that format’s superiority.

        What happens if there’s a release of a better camera module every 3 months? Rather than the dread of buying an entirely new phone and incurring high costs, people would perhaps prefer to go in and buy a new camera upgrade 4 times a year. I can just as easily see the modular bits becoming an arms race of who has the best / fastest hip new add-on, leading to more waste and more cost.

        1. Your first point can be used to counter your second point (not trying to come off as a jerk, btw). If there are standards implemented then you can sell those bits back, do trade-ins, or use ebay because they can work in other people’s phones. I like the Idea, one thing to add would be at least 3 different baseboards for size of screen and additional/bigger components. Some like big screens others like compact phones.

          I agree with Mike, at least being able to just upgrade the processor would be great, but I don’t think it would be a stretch to include screen, camera, ram, and base storage. Making it modular like the video suggest seems a bit outside the realm of possibility for todays technology, though. So if they always go in the same spots then it would be possible. Plus by the time you want to upgrade the processor to the latest and greatest it should be able to fit in the same spot, anyways.

        2. 112358: You’re quite right (and I replied to myself because I think your reply is the last down in the tree before it dies off). The direction I was headed with this argument is that the more parts a phone becomes, the more disposable it may seem; like replacing the batteries in your remote control. Obviously this concept phone isn’t taking it to quite that much of an extreme, but a $20 Bluetooth module might not be worth all the trouble it takes to go to eBay and sell it, whereas even a $100 broken phone would be.

  4. Since most of the failed phones in the landfills have a cracked screen, a non-removable dead battery, or a worn out micro-usb connection, it would seem the easiest (if not complete) solution would be for manufacturers to make phones with robust mini-usb connectors, removable batteries, and screens that can be replaced without the use of acids or heat guns. Other solutions run into the same issues that desktops have…for instance, if you are replacing the cpu it is probably because it is too slow, not because it died. When was the last time you bought a new cpu for your desktop without needing to buy a new mainboard for the new socket? More importantly from a sustainability standpoint, used working phones usually get sold second hand…so if you can avoid phone failure the issue of phones in landfills virtually disappears. If you are upgrading your phone it would actually create more waste, if only from the packaging that the new parts came in.

    1. Not sure I agree with the notion that if people can somehow sell used phones, then the numbers of phones in landfills shrinks…Case in point, when’s the last time you bought someone’s used phone? Me? Never. Shame on me perhaps, but it just doesn’t seem to happen. In fact, I’ve offered to give not-so-old phones to friends or family and without fail, they’ve refused to take them… “oh, no thank you…it’s ok, really…”.

      I agree that the ability to remove and replace bits is something that’s been lost with modern phones (remember the ‘extended battery’?) but this is largely driven by size and the consumer’s demand for smaller and smaller gadgets. Or alternatively, our passion for something that looks closer and closer to “magic” (http://en.wikipedia.org/wiki/Clarke's_three_laws).

      The fact is that the electronics simply don’t fail (at least not in any statistically significant numbers…as you pointed out) so the thing missing here is divining a way to replace *perihperals*, a unit at a time. (As virtually all other functionality is wrapped up on the same chip.) So the only thing to be replaced are the actual peripherals (i.e. the lens / sensor combo, the gyro used, etc.) And this is I think where this vision of re-usability becomes possible.

    2. “When was the last time you bought a new cpu for your desktop without needing to buy a new mainboard for the new socket?”

      Everytime? By the time I get around to replacing the CPU the socket it uses has been depreciated. My i7 920 is still good enough and by the time it isnt LGA 1366 chips wont be made any more, and has been superceded by LGA 2011. The i7 970 and i7 3930K are comparable chips, but the newer LGA 2011 chip is $100. Unless youre a fanboi with deep pockets who buys a new chip whenever it gets released, you’re always pretty much stuck upgrading your motherboard too.

      1. So if the processor needs a new socket, couldn’t the. Base plate be the processor? That would give more control over how the pins were recognized for individual blocks. If you need to upgrade your processor, change out the base and re-use the old components.

        1. i thought about this. it would be much simpler to put the cpu into the backplane, and then just have everything else with the exception of the screen as a usb device (with more compact headers instead of connectors of course). if your phone is too slow, buy a new backplane and transplant all your blocks into it.

  5. Computers weren’t modular until Cromemco came out with the S100 bus and IBM Came out with the PC. Then the PCI bus and it’s variants came out. I think you’ll always have the custom throwaway electronics, but as a maker it doesn’t have to be that way. What I want is an Ubuntu Edge. Well that didn’t happen. I agree there are problems with the concept, but if yoi’re able to install Linux, shouldn’t this be within reach. The FCC will be one roadblock, do’t be another. Get your palm off your face and look for the possible.

      1. Troll, Linux was created by people that believed in true freedom and gave away more Intellectual Property striving for that ideal than most people ever have. No reason Ubuntu couldn’t live on a phone. If people want to Act like a Wookie, that’s their choice. If they want to make responsible choices, they should have the opportunity.

  6. Some kind of dedicated Hackrf chips for the wireless (mini wide-band antenna’s could be a problem though), a ir led and receiver, and a kind of screen on a motherboard and interchangebale battery, camera, soc, rom and ram should do the trick a lot better I think. Just make it like a tiny notebook or pc, those have proven their usability. I don’t think the screen should be interchangeable either because it would make a nice base in combination with a motherboard. The only problem with this solution are sockets, and the thickness it will create (and the antenna’s of course).

  7. Technical difficulties aside(the thing would be huge given all the extra connectors, there would be even more compatibility issues), the phone is part of the personal image. People want to identify with something, how are you going to sell them a better cpu/memory/screen? They want the latest X.
    And they don’t want to worry about compatibility issues.

    Now what can we do today? Well, manufacturers could:
    1. not build crappy phones with non removable battery. This is the first component that needs replaced.
    2. build them with removable storage. Current micro SD standard has you covered for a lot of years of increased capacity.
    3. Keep the core software of the phone future proof: you buy the best phone today, one year from now an OS(or whatever update) needs so much better hardware that you barely meet the minimum.
    4. provide an accurate, fast and cheap way to do a self test of the phone. No, not like today when you have a problem with the thing, send it to the service and they say they find nothing. Multiple times.
    5. make them more physically durable(let’s face it, they are quite ok given what kind of abuse they take)

    All these are for creating a market where people would reuse the phone. A 2 year old top end smart phone will likely have better features than a low end smart phone from today.

    But they will not do this… because it will decrease the sales. They will make less money…

    1. That’s a very good point — computers aren’t marketed nearly as much on the basis of personal image as phones (Macs aside). Ours is a generation that typically prioritizes features and best bang-for-your-buck (even if it’s a lot of bucks) value, so maybe that trend can be broken somewhat.

      1. The main reason you have a self test is if you can fix it. Current phones the test is “does it work”. If it does, the test completed successfully. If it doesn’t then you throw it away. The reason doesn’t matter if there is nothing you can do to fix it.

        1. Many still do have self-tests, through an obscure set of button-presses or whatever. They’re sensibly kept secret, to keep the tech-support lines free of questions from people whose mate in a pub taught what to press, and they’ve seen something they don’t understand and wonder if that means it’s broken.

          I suppose for the reason of testing phones shipped back under warranty. Expensive phones are still worth repairing.

  8. Good luck finding a processor with that few pins.
    Good luck keeping it fast with a single data bus.
    Good luck with the heat dissipation.

    there are so many problems with this I was tempted to write a whole post. It’s like it’s been designed by an art student.

      1. Yes. Someone discovered pogo pins and had a mind-blowing wet dream all over the internrt.

        The pitch is loaded with holes, I’ll give you that. SparkFun is not working on a speaker modules, and there’s no 4-pin SOM on the market, but I’d like something similar, even if it were a 7″ tablet without cellular.

    1. Cos that’d be a bit small.

      Unless you just mean have the CPU + RAM on a card that can be upgraded. In that case, it’s been done before. The early S-100 computers, some early PCs, and some industrial modular PCs have used the passive backplane idea.

      Since the RAM is tightly couple to the CPU these days in phones (often in the same package or mounted on top), the two should probably stay on the same module. Faster CPUs need faster, newer generations of RAM anyway, and they both progress in size and speed over time. They’re also the two parts that benefit from a fast, private bus.

      I think one problem might be that people don’t know what a CPU or RAM actually are. When most people’s PCs get old, they throw them away rather than upgrade. So I don’t think it’d be clear to people why they might need a new CPU / RAM, and what benefit they might get out of it.

      It’s probably worth remembering that the vast majority of home PCs are bought in one piece and not upgraded much. Same with offices. Perhaps there’s something in the psychology of purchasing and owning things that favours this way than our way, the way of people who understand what we’re doing.

      Certainly I know some people who’ll actively run away from knowing how something works. They don’t want to know. They just want it to do it’s job. Most people don’t know what a CPU is.

  9. If it can be done for assault weapons, it can be done for a phone.

    If I buy a radio module with a chip from Vendor A (which already has FCC approval) with battery and charging module (with or without UL stamp) to mount on my backplane, privided by Vendor C, what is the problem?

    Fragility and mass-market readiness aside, this is exactly what the PC market looked like through its strongest years.

      1. Exactly, nothing. This is for the people complaining about how unlike an iPhone it would be (out of necessity).

        It wouldn’t be a Dell. It wouldn’t even be an Acer. It would be for the people who want it, for whatever reason, as it is.

        Your train analogy works if you through in a Third rail, which is all you’d be able to share among components; power and USB or some other serial protocol.

          1. My point (so hard to convey) is that you can buy a backplane and select core components from a vendor, and select or manufacture the rest to suit.

            FCC and UL be damned.

            I don’t pay any attention to assault rifles, I just know there are loopholes and a willingness to use them.

            I’d be into a similar thing. I could put an SDR on the back, or a fingerprint scanner built with open hardware.

    1. If we get more stable use of components – e.g. not changing from DDR2/3/4/etc, bus connectivity to peripherals, camera interconnects, etc this might work.

      However, the assault weapons are in a market where most of the supporting parts and connections are standard- we don’t have a new type of bullet every 2 years that isn’t compatible with the old barrel, magazine and receiver ;)

  10. In my opinion, this is a short-term problem: like the PC, the phone will have to hit a knee in the technology curve, and there will come a time when the battery, processor, network speed, screen and cam just aren’t going to get much better. Yes, there may be optimized models with large lenses, “secure environment” models with no camera or external ports (my son can’t bring an ordinary iPod to work because it has USB storage), but they’ll be niche items. If you’ve got enough storage for a international flight’s movies in “retina” quality (call it 35GB x 9 for an 18-hour flight = 315, let’s say half a TB), the upper end of LTE, said retina-level screen and an 18-hours-of-movies battery life, you’re done. What’s to ever upgrade?

  11. As a bit of a “behind the curtain” comment, a lot of people are asking why there aren’t standards. The good news is that there are but they don’t go across everything.

    Have a look for something called MIPI (“Mobile Industry Processor Interface” and before you ask, it’s not just about the processor, nor its interface). The standardisation goes across lots of areas within a mobile phone and sometimes they even make it outside of the world of the phone. For the large part they are about the interfaces that may be present both inside the silicon and between large parts (think: path ot the memory, display, etc).
    Normally standardisation doesn’t happen because someone (keyword being *one*) holds a patent which stops someone from using it (without paying, sometimes prohibitively). MIPI acts as a gateway such that standards are contributed along with patent rights (but only for other MIPI members). It also stops any one company from getting ‘locked-in’ to a single supplier.

    But the long and short is that at the end of the day – making a phone is hard (look at the PCBs in the teardowns on iFixit) there is a lot of work crammed into a very small space and, when they’re broken up by ‘reusable’ connections, it’ll compromise the functionality. Not impossible to work around but it’ll be at the cost of something else (speed, battery life, etc).

    1. They’re *not* especially hard. http://hlt.media.mit.edu/?p=2182

      Meeting the demands of current consumer-grade products with trillions of R&D investment dollars and vicious patent disputes with a new, contra-traditional model ain’t gonna happen with a twitter and kickstarter campaign.

      Making a (potentially user hostile) device with discrete components that can also make a phone call is easier.

    2. If your trying to make a smartphone the same size/shape as X brand then yeah there are going to have to be tradeoffs. But what if your not?

      Personally I wouldn’t mind it if the smartphone were an inch thick along with an inch wider/higher then other smartphones as long as that “bulkiness” allowed me to customize/repair it to my hearts content.

      Though instead of trying to cram as much computing hardware into one package to the detriment of everything else (especially battery life!) why not give the Smartphone’s OS the ability to form something like a Beowulf cluster (perhaps using your home network or a specialized computing brick you carry for that one purpose) for additional computing power when needed.

    1. have a case that goes around it. like you have all the modular bits that fit between the backplane and the screen, and then a case that goes on all the edges. like over them, like an ordinary case. done and done.

  12. Making a mainstream product modular isn’t a great idea. Making the internal components reusable is. While I may buy another phone (I make it last as long as I can) I want to be able to use the old one for projects….perhaps some accessible connections to allow the repurpose of screens, cameras, etc?

    1. I own my share of bricked phone. There’s not much reusable in the current collection. Each connector is a point of potential failure so I understand the reason. It’s all just so wasteful.

      1. actually, what I envisioned wasn’t a connector to attach the devices to the main processor, but rather a connector that allowed sharing of that device to another unit. You are 100% correct, any connection is a point of failure, and extra material that they spend an extra $1 on each phone, cutting into profit margins. The boards being multi-layered also mess with the ability to just cut up the main board and wire to test points and traces.

        We live in a throwaway society…

        1. You could go back to the through-hole Z80-based phone systems, with a suitcase full of car batteries to power the whole thing, if that’s what you really want.

          As with many things in life, particularly making things, there’s compromises, and stuff like “future hack value” is really down at the bottom beneath the stuff ordinary people and the manufacturers want.

          Otherwise all those pins and PCB traces are where the miracles are hidden. If you’re in your thirties or older, you should be amazed by what people have in their pockets now, compared to the 8 and 16-bit computers we grew up with. When a single floppy drive was hi-tech and serious, businesslike! Not to mention expensive! 80 quid in 1990-money for a 720K floppy drive! That’s pushing twice that in now-money. Or the few pounds I could spend on GBs of SD-card, smaller than my fingernail.

          It’s amazing how much work engineers and scientists have done over just 20, 30, 40 years. How much thinking.

          The current state is not to make things “throwaway” on purpose, it’s to make them absolutely amazing, in a tiny size, for hardly any money at all. That’s where the future ended up! We got gypped on hover-boards and moon-bases, but we have access to huge amounts of information and media in our houses and pockets. Anything you ever wanted to know, you can ask your phone in a couple of minutes. It’s made pub quizzes impossible!

    2. Doubt that scales to the wider audience (great for this audience perhaps). if we want to create true reusability with that strategy, my wife (non-technical) would need to get value from the parts out of her phone. Dont think that’s going to happen anytime soon.

      MAY work to formalize the parts recyclers in places like Shenzhen and get them actually stripping and reselling (in an above-board way) the parts to companies that can reuse them. Think: industrial-sized desoldering. Automate it and ensure the parts are working without issue.

      1. Still, if it takes even just a few weeks to get the shipping and collection, it works out cheaper and better to make the newer, faster stuff instead. The cost overheads would also have to be really, really tight.

        I’m also not sure what the state of the art is in de-soldering. Component re-use, AFAIK, has never been done on a factory scale. I’d be surprised if there were any tools at all for it.

        I suspect if you were gonna start off, you’d have to find a niche. Goods containing expensive parts, that for some reason end up rejected, or replaced quickly. If you found your niche, you could start researching the right tools for your robots. I can imagine a high start-up cost for this, the research and the robots themselves. Probably best done as part of a large corp that already is in a similar business.

        I think a big problem is, the most expensive, hi-tech parts, tend to be those that de-value the quickest. All the way to 0 pretty quick.

  13. If a phone is usefull, you obviously have a network.
    If you have a network, you can have a server, mesh, cloud, … (Maybe only on your own servers.)
    If you have a server somewhere, you just need to have some kind of userfriendly terminal instead of a pocket sized mainframe.
    If you need more power, enlarge your server, cloud, … not your phone..
    If you only need a wireless terminal, you’ll need less battery power “on the road”…

    That’s the way I want my next generation of mobile computing, smartphone, pda etc…

    \o/ MESH THE PLANET! \o/

    NOW!

          1. if every house has an antenna on the top to help transmit… enmeshing is a possibility, as long as all of us use the same interface to mesh the world with. we could cover the globe in free wifi(if starbucks doesn’t do it for us first) but unless EVERYONE does it, or we get funded by some corporation or some big company starts doing it, or something crazy like that happens, it’s gonna take a LOOONG time. sadly.

  14. For everyone crying about form and size… I for one could go for a larger phone… bulkier… fit in the hand better… I HATE the feeling of holding a playing card in my hand as most of these “thin as possible” phones seem to be going for. Screw that! Give me something with substance and a BIG battery so I can play angry birds all day if I want… or watch those movies while traveling. There are ZERO phones on the market that can actually handle all day use. And this means minimum 18-20 hours, not 6-8 hours. Preferably 2-3 days just in case of being out in BFE where you have no ability to charge the phone. This modular idea could cater to that segment with a battery module that makes the whole phone double/triple thick. I would actually like that as I could have more confidence holding the silly thing.

    1. Completely agree! I had an old Samsung Blackjack with an extended battery and it lasted 2 days. The battery technology has *improved significantly* since then and yet my shiny new iphone shudders at the thought of me playing more than about 2 hours of video and then using gps for any length of time. How about up the power or better yet, let me decide what’s heavy! If I want sleek, I’ll change for the “suit pocket” battery when I need it.

      1. Can’t you get a third-party bigger battery?

        Failing that, there are those external power-packs, a battery with a USB port on it, so you charge / power your phone off it. They’re pretty cheap, and you can have 3 or 4 if you want to. Or make your own. Bit of a hassle but the ones that take alkaline batteries are good in emergencies when your phone’s flat and you’ve no time to charge it.

        1. 6xAA battery holder + a 5v regulator + the end of a microUSB cable. That is my magical mobile device charger and its rather effective when loaded with 6 rechargeable AA’s (albeit not the most efficient). Actually I have 2 sets of leads soldered onto that holder with the idea being I run the 5v regulator to a raspberry pi and 7.2 to my motors, but it worked for charging a phone at the time.

    2. By complaning about size i am being realistic. if it is modular, and you want the device to work the all of the radio modules will have to be huge.not only that but the seperation between the modules would have to be gigantic. the way that phones have become small enough to fit in your pocket is by sharing space. Open one up, there is no extra space for theses walls, now in your mind separate all of these parts, and remember you are not allowed to stack things,so you now have 2d space rather than 3d space. The problem gets crazy, not to mention that at the clock speeds we use today the concept of a peg board is impossible as your signal integrity will fail.

      IMHO this is not an increase of .5″ or so, it is an increase to 10″x12″ or more This would be more practical as a tablet or laptop concept, but as a phone concept… impossible

      1. with current technology maybe, but look at how fast it’s improving? i’d be happy to have a fairly basic phone(access to wifi, basic web capabilities, texting and calling and like… clockworkings, basic stuff) that’s the size of a kindle, and 1 1/2″ thick, if i could upgrade it and hook it to my computer… that’s what i want. i want to be able to connect my laptop and my phone and use them as ONE or separate, in true modular fashion.

    3. Also As for phones that can handle a days use. Droid Razr HD, and Galaxy SII do fine. I have used both. i have no battery life issues at all. The Razr is not terrifically small, and it is tough. The with the probable size increase by this sort of technology you probably wont be able to hold it in one hand.

  15. Upgrade your wifi, screen, bluetooth? Any decent phone from the past year will never need an upgrade in most of the departments proposed. The true problem for most people is the risk of breakage/component failure, and addressing that is the true benefit of this idea.

    The biggest obstacle to phone serviceability is that the touchscreen glass, which breaks frequently, is never designed to be easily swapped without replacing the LCD, which is usually fine. They are glued together and the glass is on top, with little protection around the edges. Both parts are specific to the shape of the phone, so the cost of replacing the them is inversely proportional to the number of units in circulation. A replacement Nexus 7 screen, for example, cost more than half the cost of the device, and is a nightmare to replace.

    A smarter design would be standardized sizes for the screen/glass, and a frame holds the glass and screen in place separately via screws instead of glue/taping them together and in place. This would involve wrapping the frame around the perimeter of the glass so it can grip it to lock it in place, via pressure, when the screws are cranked down. This would also protect it, instead of leaving the glass floating on top which promotes breakage anyway. Since the contact with the frame is going to effect the touchscreen, they have to be designed with a deadzone perimeter.

    1. Agree, though sadly so many of the suits are demanding compelling product differentiation. Standard for factors kill the only differentiation left, if you consider that most are running the same OS, and are in the hair-splitting stages of technological differentiation (really, 1080p on a phone? for god’s sake what would I need that for on a screen I can barely stand to look at!)

  16. Uh… do none of you understand the point of a concept video? It’s designed to get people interested in the idea of building a modular device. They’re not showing you what the final product is.

    They’re looking for social approval so that they can then go to manufactures with a concept and then figure out how to do it with hardware.

    This idea reminded me of what BugLabs were doing.

  17. I’d done teardowns on probably 20 Chinese smartphones while living there recently to try and determine just how they build a $10 phone for the market in Africa (http://goo.gl/UIx92m). What was incredible to me was not how complex or sophisticated most phones are, but rather, how simple! I used to be impressed when an engineer interviewed with me and said “I build phones” but today, it’s really a turnoff, largely because I know just how modular and self-contained each element is…And the fact is that to design them today is largely more a mechanical / layout challenge, than a complex engineering task.

    Hats off to the guys designing the chips inside of them! Phones are (with few exceptions) all simply a few modules that fit together like pieces to a puzzle! That said, the dream of interchangeable elements seems to really rest with the peripherals… (read on!)

    An example, the chip at the center of most Chinese phones is from Spreadtrum. You can buy them on taobao (www.taobao.com) for about $6USD. This includes your ARM A8 core, transceiver, baseband, display / touch interface, SDIO interface, memory interface, standard buses (SPI, I2C, etc) and it’s all right there for the taking.

    The variation comes at the level of the peripherals you attach or hang off of those interfaces, and the memory you put on…also the SW at the heart of it. And even the memory is modular, with ICs that include both Flash and SRAM on the same chip. Even interchangeable packages for more / less memory. Truly astounding!

    So here’s what I’d propose…

    Modularize the peripherals and make a standard set of drivers for each that include standard interfaces and such, all supported by the same Linux / Android distro you’re targeting. You can obfuscate the peripherals behind APIs and make them swappable at the SW level easy enough. (Afterall, this is what the Chinese phone manufacturers are doing; the problems are all moved to SW, and even there, they are using the same calls to achieve different results. A few changes to some variables in a struct somewhere and it’s a completely new display.)

    Now the problem moves to layout and mech-industrial design, and also people’s willingness to carry something larger in exchange for interchangeability. I disagree with comments about the video that say that the creator is trivializing a complex problem. Horse$h*t! Half of those peripherals are running on SPI or I2C anyway. The only things that probably aren’t are the camera and the display because of bandwidth issues (or SDIO or memory but those may be part of the main module or could be a module by themselves).

    So now when the latest peripheral comes out, so long as the CPU can support its improved capability (and I don’t aim to trivialize this!) then we can hot swap it. If the peripheral is *not* supported by the current CPU, then swap the CPU module and again, the system just “works”. The fact that things aren’t designed this way is not a function of whether or not we can, but rather, whether or not we *have*! There is no reasonable technical limitation that cannot be overcome for this design challenge.

    Couple all of that with the completely awesome biodegradable circuit technology (ok, maybe a few years out…http://goo.gl/XD82lS) and you are golden!

    1. The article you linked to wasn’t very informative, in fact it was almost empty of information, and pretty short at that.

      Have you got a link to your teardown stuff? I’d be interested to read it.

    1. At the end of the day this is nothing more than wishful thinking. All physical/technical problems aside, the software for such a thing would be insane. None of that really matters though. Phone manufacturers would rather you go buy a new phone every X years. Supporting a single unified device is easier than the nightmare a modularized system would be, and I think the vast majority of phone consumers would rather have a single device they can identify with that offers easy support vs something modular.

    2. I think David Mellis and his crew just took an Arduino and it’s cellphone shield, plugged ’em together and that was that. They wrote a rudimentary OS for it, too, but I think the module used on the shield is 99% of a cellphone anyway. Does all the hard stuff, just needs a mic, speaker, and a microcontroller to feed it AT commands to make it dial and text etc.

      Pretty tempting actually, modules like that. There’s been a few phone builds on here that I think have used it.

      1. They did a version two of the DIY cellphone with an SMD Atmega32. the PCB is worth looking at.. for some reason it has an odd antenna taking up lots of board space… that display could be abstracted away… I wonder how far SPI could be pushed to handle display… many fewer buttons would be necessary if the tablet digitizer and screen could be abstracted, but what protocols on the backplane? PS/2 and digital VGA? USB? FlexCAN? OC-3? LED Optical Ethernet? it may a while until any kind of standard could stabilize… a charger would be a whole separate module… the ability to fit many sizes and shapes of battery in there… the number of combinations boggles the mind… and then you have to have decent CAD… i’m curious about wood as a material due to appropriate density and ability to take shocks and could be remilled every few years…

  18. seems to me like the modular design is good, after all that’s what the PC uses, that’s why i can replace my video card if i want or add a new keyboard or replace my hard drive. but you need a fixed architecture (using PCs as an example this would be the x86 architecture) which, for phones, tends to be ARM so good there.
    the parts should be self contained and use an open protocol for communication.
    course then you get into efficiency concerns (most phones are optimized to use the least battery power, not to be the most upgradable.)

  19. oh god i just watched the video…should do that BEFORE making a comment.
    this guy has no idea what he’s talking about “we just connect them with a couple pins” “just replace the block that affects speed” “and it’s all held together with two screws” my god…so much facepalm. might as well have said “it’s all running on unicorns and rainbows”
    to clarify on my last post, i think A modular design is good, i don’t think guy has a good modular design. anyone who thinks there is a part that controls speed,or that you can even MAKE such a block, really needs to learn how computers work.

  20. the only problem I see here is the lack of optimism by people.

    I think this is a great idea.

    One problem would be how every block would communicate with the motherboard.

    2 pins or 4 pins each would be a problem for lets say the CPU and the screen as well as other many-pin IC’s and such, so a new bus protocol could fix that.

    On the plu side, a new bus protocol using only 2 pins that can deliver several hundreds of MB per second would only be an advantage for other areas in electronics.

    1. Agree about the lack of optimism. I2C and SPI would both work for this with no new bus req’d. Only place you’d have to reconcile the pinouts would be with the display and that would be pretty easy to sort out. Could be the only thing that uses a different connectivity schema.

          1. You can’t do USB over I2C either, or run SD card, or WiFi. So you’re going to need a faster bus. And if you have a fast bus, why bother with I2C or SPI at all, even for the peripherals that can use it ?

            So, you’re going to have a really fast serial bus, that can be routed to any pair of pins on the backplane, which can also double as power pins. And everything will be properly impedance matched, and EC/FCC compliant in every configuration. It’s just insane.

          2. Yes and no (and I wasn’t sure you needed a complete system diagram, I was making a comment not a spec, but apparently we’re going to call out anything that might sound at a little wonky so let’s go…). Of course you can’t do USB over I2C (though you probably could it may just not be practical or fast)…and regarding the other bits:

            a) would you actually want to change the USB I/F? location? connector size? ok. perhaps. but it could just as well be static. there’s no major performance gain there unless the bus standard changes.

            b) SD card is SDIO. you have one interface. it’s standard on the CPU. again, a location change is an industrial design constraint as much as anything else. so if you leave it in a static location, then it’s really just a SW constraint. again, nothing to change there.

            c) wifi is embedded on-chip. otherwise it’s a separate chip and again, how frequently does that have to change?

            all of the other bits like ambient light sensor, gyro, etc, can all run on standard interfaces .

            not sure if you’ve ever seen the layout inside most non-iphone, non-motorola mobile phones but I’d encourage you read my longer post below and then suggest doing a teardown on 3 or 4 chinese android smartphones. the variation is almost uniformly in SW, memory, and screen / camera *quality*. the basic hardware and interfaces are pretty much static.

    1. And in comes the nightmare of making an OS that will work with devices (and drivers, oh god the drivers…) made by underpaid Chinese engineers :P

      In my unhumble opinion, the problem isn’t hardware, but software. There are tons of various parts that could easily be made into removable modules*, but somebody has to write a driver and do the friggin’ debugging, on top of that somebody else has to solve why module X works fine alone, but keeps crashing everything when module Y is present…(just like with a PC)

      *Or even screw the FRUs, you could simply sell modules in BGA (or similar) packages and those without the skills & tools would simply pay/beg somebody to install them…

      1. Though I can relate to your experience with drivers (I’ve struggled to get some and had some that hadn’t worked before, surely), in this case the drivers are abundant and are being distributed with the linux distro that comes with the boards and a nice walk through to get them up and running. (Easier to get drivers from these guys than it is to get a datasheet from Marvell or Broadcom). It’s unfair to assume that ALL drivers stink or don’t exist just as it’s unfair to assume the engineers are underpaid or somehow inferior. I managed 30 engineers in china, paid them well, and they could all hold a candle to anyone I hired in the US or elsewhere.

        I agree with you completely that this is a SW problem. And that is something that CAN be solved, with the right dose of patience and ingenuity. The fact that a driver might not exist is not a reason to *not* build something. Write the bloody driver! :) And if debugging is a challenge, it’s why we get paid big bucks. Debug it! The fact that something is hard is not an excuse for why it can’t or shouldn’t be done. Great innovation comes when someone dares to close the gap on something everyone else is saying is too hard to do.

  21. I think the discreet components need a wireless interlink (like bluetooth). One little component would have the Cellular Radio, Wifi Radio, Dect Radio, etc. Another device would be a portable storage device (backpack HDD, Solidstate mem device, etc). Then you get a camera of some type, (goPro, SLR, misc. webcam, etc) which also connects. You could also add in various headsets/speakers for audio. Finally you have various displays that you could use to access the devices in your personal network.

  22. My vision of the future of portable electronics is the personal area peer-to-peer network. You have a processor and power source in your pocket and the phone is simply a dummy terminal for that processor plus has the feature of being able to make calls. You could wirelessly attach any number of ‘smart’ devices to it with various bits of functionality.

    @Dave: I have never once purchased an all-in-one computer. Still build them… they’re better that way.

    1. I hate to through a wrench in your logic. But if you have never purchased an all in one, how do you know that yours is better? I have used devices that were better on paper, but catastrophically worse in reality .

  23. There’s some folks who always want cutting edge. Tech in this area is moving fast, and nothing will prevent them from getting a new phone on a regular cycle. Not even a modular phone.

    But I am not one of those people, and here’s what would make me happy:

    * Put a few extra bucks into construction so that it will last, not merely survive the warranty period.
    * Stop the lead-free solder nonsense. I firmly believe the overall environmental impact of throwing away devices due to failed lead-free solder joints, greatly exceeds the benefit of removing lead from their construction.
    * Make components subject to unavoidable wear and tear (screen, battery) easily replaceable.
    * Remove artificial, non-hardware related restrictions. I should not have to root my phone to add software functionality. There should not be deliberately disabled hardware features (like GPS). Nor should a phone be locked to a specific carrier.
    * Improve hardware hackability. Bring out all unused pins on the CPU to pads, whenever space allows; rather than leaving them inaccessible under a BGA. This way, I (or someone else) may find a new purpose for a phone when it’s no longer useful as a phone.
    * Carriers should introduce additional and affordable low-cap data plans. If I want to attach an MCU to an old phone in order to relay a few KB/month over the cellular network, why should it cost $10/month to do so?

  24. Obsolescence is a function of marketing. Treat your phone with respect, replace the battery every 18-20 months or so (phone makers are well aware that if they keep changing the format of the battery and making replacements hard to get, it encourages consumers to buy a new phone instead of tracking down a new battery… we have to stand up that nonesense!), and don’t give in to advertising hype.The only time you’ll need to replace your phone is if the technology of the network is abandoned (most recent example, the analog system being sunsetted…). I’m using a phone that still uses 1x data… guess what? It works just fine. Calls and txt’s come and go. Does what a phone is supposed to do. 1x will be with us a while longer. 3G will be with us for another decade or so before it’s phased out. 1% of the population actually needs the latest/greatest, the rest of us don’t. Want? Sure. Need? No.

    When we learn to seperate desire from necessity, we’ll discover that we can do a lot more on a lot less.

    Keep buying into the market hype though, and you keep increasing the waste and excess. Have you seen the conditions people work in to mine the materials of which our phones and circuitry are made??? Have you seen the conditions people work in to relaim materials from discarded electronics??? Ignoring the realities of where our stuff comes from and goes to doesn’t make them any less real.

    (rant mode: off)

    1. Thus, in answer to the original article, the number one thing we can do is insist on battery support for existing phones. The batteries are the wedge by which manufacturers/networks split us away from our existing phones. Very little else about a phone needs any improvement.

      1. However wierdly-shaped a battery is, there must be a standard-sized Li cell that’ll fit into it’s gutted case.

        That said, no-brand batteries for pretty much any phone are available on Ebay. I suppose all you need is a supply of cells and something to mould the plastic, so some Chinese companies must be taking care of that market. I tend to keep my phones for around 5 years each. Only one of them needed a replacement battery, but it was really cheap.

    2. +1

      We have amazingly powerful phones today, what do people use them for? I’m certain that 99% of use cases can be covered by standard hardware and great software.

      It’s difficult to design an OS with minimal overhead while still allowing 3rd party apps and integration, but I think it can be done a lot better than it’s done today, especially if long time support is a focus.

      Banning vendor-specific bloatware is probably a big step in the right direction. What about auto-updating of drivers and the OS itself? Most people I know would not know how to update their OS on their phones.

      I believe obsolescence is mostly driven by marketing and bad software. Waste is only reduced by lowering consumption.

  25. I would call how they want to implement this a fail but I wouldn’t call the idea bad. There is a precedent for this with the IBM format PC. laptops like phones though have space constraints which modular bus’s impinge on. each modular bus will cost space in the device. not to mention that the variations in size formats would result in multiple standards. this phone block idea turns one third of the phone into a bus, it dosen’t even state how this bus could transfer a lot of data at low power.

    why not standardize two things:
    1) Packet controlled common medium (internal in phone) communication for each module to minimize any wasted space by being modular (imagine half duplex Ethernet).
    2) CPU can have as many lanes as wanted, each lane can have as many peripherals as wanted and any peripheral can take multiple lanes. (example 4 lane display, 5th lane Wifi and 6th lane is bluetooth, GPS, cellular)
    3) phone internal peripheral power voltage (e.g. 5v like USB)

    Then there could be two types of Mobile internet devices:
    A) small ones without breakouts that you could strip out a module (like a camera) swap for any modal that fits (or with a dremel if it dont) and solder it where the old one was
    B) large ones with breakouts that you can upgrade or repair easily later.

    they are right that there is no call for modularity for several reasons in phone history:
    1) phone companies like selling you phones over and over
    2) phone network providers like selling you phones in contracts
    3) PC started when people who owned them, most likely built and wanted to repair them. Phones started from people who owned them for a single function, designed inexpensively (time to repair had high relative cost) and have no interest in understanding them (little will to repair).

  26. I’m sure all of these things have been said before but my perspective is that phones are thrown away because of these reasons:
    1. Fashion – People want the latest and greatest. Hopefully most of these upgrades get resold.
    2. Software – If you’re lucky you will still get new versions of the OS ported to your phone once you get a generation or two behind. After that it’s a forced upgrade if you want the new features.
    3. damage/wear – While it is possible to fix parts of your phone. Manufacturers certainly don’t make it easy.

    Currently phone manufacturers want your phone to have a limited lifespan. It makes them more money. That is the first hurdle to overcome.
    After that the easy solutions I think would be to require manufacturers make replacement parts available for a length of time after they stop making the phone. Even if it is just the screen and battery that would help a lot. Next, get the manufacturers to provide software updates for models longer. Even if it cost some money to get the latest version of the OS for your model its better than no option.

    Going a step further having the CPU and memory on a removable card could allow an upgrade or two to increase speed/capabilities.

  27. While bashing the idea is fun and all I see very few suggestions about how to do better. We have modular computers why is a modular phone so far out the realm of possibilities? When did damned hard but worth it start being the quit point? Do ISPs require us to have a specific type of Desktop, tablet, or Mac?

    We went to the moon in a space ship with less computing power than an TI-83 calculator. Don’t be so quick to shoot an idea in the head just because you see so many problems.

    1. …After thinking about this a bit more perhaps there is a way to accomplish the goal here.
      The complete Lego concept is impossible however you could perhaps come close to modularizing things similar to current computers with a few caveats.

      In my mind the overall phone industrial design and motherboard would be tied together. your IO connectors (usb, SD, SIM) would likely be tied to the motherboard as well.

      What if the modules were more like custom chip scale packages. Ideally being able to carry multiple dies. For example, a CPU, RAM and ROM integrated into a single package. This would allow you to attempt to standardize the packages. The connector part is tricky. It will need to be extremely thin high density and controlled impedance. I suspect the answer will be similar to current fuzz button or conductive polymer connectors.

      Many key parts of the phone are on standard busses at least on the physical layer. MIPI, SDIO, LDS/DSI, IIC/SPI. This could allow you to change your camera, display, add ROM. I don’t see the base-band processor being something you would upgrade often. I think if it isn’t part of the CPU they talk SDIO generally which would lend it to being a module.

      The end concept would have pockets that these modules fit into. and the case holds in place once it is all together. So you might have a large spot which carries the processor, two smaller SDIO spots for bluetooth/wifi/baseband or other high data rate type thing, and several smaller IIC spots for things like low power wireless or sensors. Things like the display and camera are already connectorized as it is.

      It is definitely not friendly for your average person, but building your own computer isn’t exactly something your average person does either. I think it would require it being cheaper to take the phone to a repair shop than to just buy a new one for this to gain any traction.

      I don’t think there is any way around the software being a nightmare…

      1. Exactly! THIS is the kind of response we should have to this sort of idea!

        A buddy of mine at Gulf Stream mentioned treating the plug and play hardware like a physical app. Treat the new hardware like you would any upgrade on a computer. Install an app just like you would drivers or GUIs, then reboot the phone with the new gadget on board. Though, I don’t know how difficult that would be with the main processor.

        It may take the companies that produce the devices making their own proprietary modular units first and the market catching up, much like the aftermarket auto parts industry.

  28. If making everything modular is fraught with difficulty (as many have pointed out), why not just focus on the parts that are easier to make modular? I’m thinking just standardize the housing so that you can swap out all the electronics and keep the case. Perhaps something similar could be done with power / batteries. If I had to replace all the electronics but got to keep the housing and battery, that’s already a massive improvement on what we have now.

  29. Here’s a strange thought.
    Instead of connectors, why not fiber optic windows. Get them aligned and they talk to each other. Only electrical connectors between modules would be for power.

  30. So what upgrades would you need on say a Galaxy S2? NFC? Bluetooth 4.0? Seems like a software radio or a tiny additional radiomodule could handle that need.
    Batteries should of course be standardized for all phones just like chargers was, It was unfortunate that the charging standard was set on microusb since its not the most rugged connector but atleast it is a standard. If fact the mini-connectors is part of the problem if phones had fullsized sd-slots and usb-host connectors, it would solve most of the problem making it easy to add storage and additional new perhiphals.
    Of course partically you want that usb connector enclosed with an empty space for the radio/storage-module, it could be put on the top of the phone like the CompactFlash card slots on PDAs of old. Multiple storage cards would be even better.
    The breakage of screens is a big issue but could be migated by putting a dedicated video output on it. If the screen breaks plug in mouse, keyboard and screen and repurpose as something else, media player, network camera, etc Adding two layers of plastic shell between mainboard and screen for a average user replaceable screen with some kind of slide on mechanism would add to much bulk.
    Nobody is impressed with ultrathin shiny shit anymore, but if somebody takes out a phone with a x2 or better capacity battery everybody starts asking questions. “You only has to charge it twice a week instead of twice a day” “The grip feels more secure” etc

    Of course everybody else in the industry is envious of Apples 2-year planned obsolescence profits just look at “ultrabooks” very few of which has any right to exist.
    Lighter with the same capacity is always good. Shiny and thin? Not so much, connectors and physical ergonomics sets limits. If something becomes lighter and thinner I prefer to have the choice of adding capacity(storage, battery) if I want.
    It could be done rather easily, a few standard expansion cardslots, a few standard fullsize ports, and sticking to the same standard battery connector, width and length.
    The question is rather who is gonna make and market the true Anti-Iphone?

  31. We’ve tried integrating everything into the same box, maybe it is time to separate it.

    Have a “candybar” hotspot plus phone (with a jack and/or bluetooth headset), maybe with a keypad to dial. So it can stay in the pocket.

    Come up with a display unit (with touch and/or keyboard – mouse – digitizer).

    Camera (with or without microphone) should be universal so Nikon would just connect to the master hotspot.

    Somewhere have a multiple SD or microSD network server.

    Come up with a universal lithium rechargeable battery (there are some, but they are bigger or are too small). Perhaps a series like AAAA through D or greater.

    So there would be a wireless personal network where the functions would be split out, or could even be transferred (at home or in the hotel room to the TV, or to a real keyboard if nearby, the storage stick is synced when on external Wifi…).

    1. FWIW I’m not really bothered about a camera. I think the idea is, because they’re cheap enough, nobody wants to be the company that doesn’t put a camera in. But how much photography did I ever do before? I do about the same now.

  32. “The primary objection (other than design implausibilities) should be obvious: dividing the phone into exchangeable bits does not inherently reduce waste.”

    Waste? That’s your big worry? You want to try and re-use that IBM XT that’s holding open the door in the basement while you’re at it?

    The CPUs and memory double in capability every two or three years. First gen iPads are too slow to run OS7. You’re not going to modularize your way out of that.

    Fixing consumer-grade phones? You’re living in a technological singularity. Stuff is obsolete six months after it hits the streets, replaced by something twice as good. And its -cheap-. As in, too cheap to be worth stocking spares for. If you could fix your Motorola Razr, would you? No way.

    You guys need to get over this hipster greenie kick. Its -stupid-.

    1. Actually technology-fetishism and a need to outdo one’s peers with the latest cliquey gadget is hipster. The green thing is a bit hippyish, maybe.

      I get about one new phone every 5 years, usually cos the old one wears out. Why? IT’S A PHONE! Exactly the same model was good enough from the invention by Bell, up to the takeoff of mobiles in the mid 1990s. Before mobiles, tone-dialling was the single advance in over a century of phones. You talk into one end and your friend can hear you.

      I’m fine with that! The advantage mobiles bring, is mostly never having to miss meeting someone, or to hang around waiting. If you want to meet someone now, you call them and go exactly where they are. In the past you’d wait by some ambiguous landmark in the cold for hours. The end of this is great, thankyou mobile phones.

      But apart from that, so what? We can talk over a distance. When I was a kid we just used to shout or use phoneboxes, did the trick.

      Why would anyone need a new mobile phone once a year? It won’t make your wang longer, although it *does* vibrate.

      1. Because it’s *not* “just a phone” any more. You may think it’s just for voice communication (and so do I) but that’s not what (entirely speculative figure follows) 95% of smartphone users think. In fact using the damn thing for vocal communication is the *last* thing they would do. Message, surf, take photos (mostly of themselves), squint at TV programs that absolutely have to have been made only yesterday, that’s about it. Faster! Brighter! Clearer! Shinier! And they couldn’t give a toss about the environment or 3rd world working conditions. They are the fetishist hipsters indeed. I think the real problem with a phoneblok/whatever-is-better is the lack of a big enough market. Cynical, moi?

  33. I think the big issue with a modular phone is that everyone wants to go too far with the modular concept. You only need to make the minor functions modular.

    Why not just make the camera and screen modular. That way you can buy the phone at dirt cheap prices, and add the new features that you would desire. Also you could add extension ports for a slide down keyboard, new flash, extra memory, extra simm, add thunderbolt.

    There will always be limitations. If you build the base to accommodate these ideas, then you will have a phone that can hit more markets.

    Some people want to buy a phone just for calls, so they don’t need a camera, they don’t need tilt functions, they don’t need an hd screen, they can pick the bare minimum. If you are a gamer, then maybe you would like to get a controller attachment that doesn’t need Bluetooth, or a USB cable. Something that will fit in a pocket, and use tilt functions. There will be parts to make the phone be able to handle this.

    Just imagine, a phone that works the way you want it to work.

  34. Texas Instruments had a plug-n-play system in 1979 with the expansion hardware for the 99-4 and 99-4A computers. Devices which plugged into the side port or into the Peripheral Expansion Box (PEB), which itself connected to the side port, all had all the driver software on a ROM.

    That was called the Device Service Routine or DSR. The OS in the chips built into the computer had no built in capability to communicate with serial or parallel ports, or floppy disk drives or any hardware other than what was built into the computer console.

    What it did have was a place to look for any new commands and functions contained in a DSR. Plug in a peripheral and new commands and functions are available.

    While some 1980’s micros were limited to only what the designers could envision and provide support for in the computer’s firmware, TI’s open ended approach to operating system design has allowed peripherals to be built for the 99 with technology that didn’t exist in the early 80’s. A typical way to do a new device for the 99 is to use an EEPROM and build in the ability for the firmware to be installed using the computer. Some older 3rd party peripherals used battery backed SRAM to store their DSRs. Most people these days remove the batteries and rig up a small wall-wart to keep those old cards going.

    Software on cartridges (TI called it Solid State Software or Command Modules) could also automatically patch in new commands and functions to the core OS. One example is the Terminal Emulator 2 cartridge (which was actually a terrible terminal emulator) which added a powerful command line speech synthesis capability to TI BASIC, if the Speech Synthesizer was also connected to the side port. Completely replacing TI BASIC or anything else in the console’s firmware wasn’t required. Just plug in the hardware and it’s “Oh hai! I can haz speech!” or read and write floppy disks or use a printer or a modem. A DSR could contain anywhere from zero to dozens of new commands, available either to TI BASIC or only usable with Extended BASIC, Assembly language or other specific software, on disk or cartridge.

    If only computer software worked so cleanly now! Very few software packages will remove 100% of the alterations their installation creates when they’re uninstalled.

    Applying this concept to a modular cell phone could have the phone’s core component with a minimal system that only supports the core’s built in hardware – plus a large “box” to contain access points to the support software so any new device that’s connected to the phone core has its software transparently available.

    The extra cool bit about such a system is the additions to the core software are truly modular. Disconnect the peripheral and it’s *gone*, leaving no trace it was ever connected because the core software was never altered, no configuration files or logs left behind, no Registry littered with leftovers.

    For me, the only way I’d want one of these modular phones is if it has a 5 row slider QWERTY keyboard module to connect.

Leave a Reply to GreenaumCancel reply

Please be kind and respectful to help make the comments section excellent. (Comment Policy)

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