Reverse-Engineering A Superior Chinese Product

It makes an Arduino look like a 555.  A 364 Mhz, 32 bit processor. 8 MB RAM. GSM. Bluetooth. LCD controller. PWM. USB and dozens more. Smaller than a Zippo and thinner than corrugated cardboard. And here is the kicker: $3. So why isn’t everyone using it? They can’t.

Adoption would mandate tier after tier of hacks just to figure out what exact hardware is there. Try to buy one and find that suppliers close their doors to foreigners. Try to use one, and only hints of incomplete documentation will be found. Is the problem patents? No, not really.

[Bunnie] has dubbed the phenomenon “Gongkai”, a type of institutionalized, collaborative, infringementesque knowledge-exchange that occupies an IP equivalent of bartering. Not quite open source, not quite proprietary. Legally, this sharing is only grey-market on paper, but widespread and quasi-accepted in practice – even among the rights holders. [Bunnie] figures it is just the way business is done in the East and it is a way that is encouraging innovation by knocking down barriers to entry. Chinese startups can churn out gimmicky trash almost on whim, using hardware most of us could only dream about for a serious project.

He contrasts this with the West where only the big players like Apple and Google can step up to the plate. Everyone else is forced to use the embarrassingly obsolete hardware we are all familiar with. But [Bunnie] wants to get his foot in the door. “Can we find a way to still get ahead, yet still play nice?” he asks.

HAD - Gongkai XrayPart of his solution is reverse engineering so that hardware can simply be used – something the EFF has helped legally ensure under fair use. The other half is to make it Open Source. His philosophy is rooted in making a stand on things that matter. It is far from a solid legal foundation, but [Bunnie] and his lawyers are gambling that if it heads to a court, the courts will favor his side.

The particular board targeted is the one described above – the MT6260. Even spurred by the shreds of documentation he could gather, his company is a 2-man team and cannot hope to reverse engineer the whole board. Their goal is to approach the low-hanging fruit so that after a year, the MT6260 at least enters the conversation with ATMega. Give up trying to use it as a phone; just try to use it like the Spark Core for now.

HAD - Gongkai BeachheadHe is already much of the way there. After telling you what is on board and why we would all want to use it, [Bunnie] shows how far he has gone to reverse engineer and describes his plans for the rest. From establishing an electronic “beachhead” base of operations to further probe the device, to X-rays, photos, diagrams and the beginnings of an OS. If this type of thing interests you at all, the meticulous approach and easy-reading of this tech teardown will surely impress and inspire you. Every step of progress requires a new hack, a new solution, a new ingenious way to pry information out.

We’ve featured some awe-inspiring reverse engineering attempts in the past, but this is something that is still new and relevant. Rather than only exploit his discoveries for himself, [Bunnie] has documented and published everything he has learned. Everyone wins.

Thanks [David] for the tip.

116 thoughts on “Reverse-Engineering A Superior Chinese Product

          1. Ditto. Arduino is OK but quite frankly, I’m sick of that shit. It’s one of the reasons I quit using VB. It gets old having to deal with bugs and workarounds that wouldn’t exist if the damn developers would quit trying to reinvent the wheel. This is usually on top of whatever bugs and quirks the other languages have to deal with.

            No thanks.

    1. Because it is not produced in a large enough quantity to keep the price as low as Atmel based Arduino knockoffs coming out of China. The MT2502 is interesting hardware, and looks very powerful compared to the ATMega based Arduino boards, but until I can get a small board for < $10 US shipped to my door to experiment with, I will just stick with the Arduino platform for small projects.

        1. Arduinos came out about 10 years ago. They’ve been expensive and not as capable as you’d like ever since. At least now you can find Arduinos for fairly cheap on Aliexpress.

          1. I have made about 10 orders from aliexpress and I needed to do a claim on about halt of those orders… not a very good track record (it was for either wrong product sent, defective/malfunctioning product or wrong quantity sent)… I have gotten my refunds but it takes about a month to get your first order in… so keep that in mind.

        2. We’re talking about an existing board, not some future development. I’m sure we’ll get to a point where you can get a $3 board with 8GB of ram, but we’re not there yet.

        3. That’s $3 per *chip* not per board. To get to board level, you are more likely to see $10 minimum san connectors.

          Assuming Moore’s law of transistor doubling every 2 years still valid…
          log(8GB/8MB)/log(2)*2 years = 20 years for something in similar die size to have 8GB.

          Time horizon is too long to tell about inflation etc. to get an idea of price.

      1. Yes, like there’s pretty much nothing for the ST Discovery F3 board but tons for the F4. I got one of the freebie F3 boards, also got a freebie startKIT. Dunno what to do with them but hey, free stuff.

        With its gyro, accelerometer and compass, plus processing capability the F3 could be the core of an inertial navigation system, Wrap it up in an arm bracer like a pipboy but not so silly looking. But has anyone done anything like make an LCD display that’ll connect to it? Other peripherals? Noooo. They all go “Oooooo! Ahhhhh!” over the F4 board.

        1. This story pointed out that there isn’t even sufficient public data sheet or header files available.

          It’s also helpful to have libraries adapted to the board and a devel board.

          1. The whole thing is just a mish mash of components and probably the worst or second worst firmware.

            Why waste all that time making it usable? Why not use that time for an open hardware project using the same components?

  1. You can use them if you are willing to go the chinese way. Copying certain IDEs and datasheets.
    What they want to implement is an open, legal toolchain.
    After that you just order 10k of those boards and sell em for a 30% markup.

        1. But does it come from ARM, or was it cloned ? And if it came from ARM, are they paying all their royalties ? I would seem easy enough to produce a few million extra chips, or to obtain an illegal copy of the design files.

          1. ARM is an architecture, the British company that licenses the ARM IP does not make IC’s. So, all ARM cores are technically clones. Typically, there is an initial charge for IP which covers the core design (ARM7 in this case). Additional copyright licensing may apply, but often this amounts to $0.01 – $0.02 per IC.

            It does not appear the Chinese government tries to enforce international copyright law when the copyright holder is a foreign company.

          2. FYI: MediaTek (MT prefix for their parts) is a Taiwanese fabless company making chips for consumer electronics.

            There is no IP violation at the chip design level as MT is paying ARM for the chip licenses and whoever else’s IP they are using. The only “IP” in question is the reference designs and EDK for the chip found floating around. This is not a fake chip issue (yet).

  2. one problem with RE is if you’re doing it a few hours a day after work it’s a nightmare to document and establish direction. Them if it’s collaboration people will be annoyed by the rate of progress on your end.

    You’ll spend a week just getting your environment set up for hardware and that’s IF you already have tools like JTAG and bus loggers..

  3. It feels like the kind of thing where they don’t want to release docs, as that would make it clear what IP they’ve ripped off. More charitably, it would make it too easy for competitors to rip them off…

    1. I get the feeling from reading Bunnie’s blog that IP works somewhat differently in China than what we in the west are used to. It may well be that there is no such thing as corporate owned IP over there; it may belong to the government and they dictate that it be made available for a much more nominal fee than what is common here.

    2. Actually it sounds to me like they don’t want to be bothered with support questions … even if you release something free people will start e-mailing you. On the other hand, if you pirate it you’ll probably leave them alone.

    3. ever worked on a REAL engineering project? at the end when you total up the expenses you’ll find that documentation and training are a pretty big part of the budget. if you can bypass these steps by making the teams small and close you can save a lot of money. And trust me they could care less about how anyone else feels about how they operate.

  4. I love Bunnie’s blog and am a fan boy of his work but he’s lost me on this one.
    Why try to build an open source toolkit and build value on semi-closed “Gongkai” hardware that you’ll have trouble getting.

    It will only benefit the people that are excluding you from the supply.

      1. Thanks for pointing to those talks. I’ve started Jonesing for the Novena again and I’ve found another days worth of talks I want to watch later.
        The point that I’m not understanding is that this exclusion is being done on purpose. After all this work is done a half dozen minor changes can throw the effort back to the starting line. Wouldn’t a better use of time be to work on an open source hardware project that is modular and brings the same funtionality into the open source community. If they want to limit its use to within China, fine. But why should the community at large worry about building volume and support for that semi-closed for-those-in-the-know network. If you look at the triangle that Bunnie uses to describe the system, then all his work is going to benefit one or two of those sides of the triangle by increasing their volume. They should pay hom to do this notkeep him excluded.
        But hey, he’s Bunnie. It’s all awesome.

  5. $80 for this black box – just get a BeagleBoneBlack. For me, I’ll stay the Pic32 and Mbed supported ARM boards. No hobbyist needs the headaches of acquiring these closed source boards with no much in the way of software.

        1. of you can be reasonable about it and just buy $15 phone with free shipping like normal people?
          Yes, there are no $10 devboards for this chip yet, not surprising considering there is no SDK, and Bunnie/Xobs efforts are the first beachhead towards this goal.

          1. My wristwatch-phone does. This’ll be great, it doesn’t get much use as a phone, but being able to add new software (eventually) would be wonderful. For one thing, a half-decent video player that uses the whole screen, and not just a tiny square in the middle.

            Anyone know anything about the battery issue with those watches? After a few weeks left idle the watch refuses to charge either of it’s batteries.

          2. May be the battery level dropped too low so their undervoltage *safety* circuit kicks in? Do bare in mind that it can be unsafe to charge these batteries when they dropped too low a voltage.

            You can try to take the battery out and use a very low current say 10mA (e.g. from 5V with a large value series resistor) to slow charge it. Once it gets to a reasonable voltage like 3V, pop it back in the watch to use the internal charger.

          3. Thanks for the tip, Tekkie. As it happens Youtube of all places came to the rescue, suggesting something a bit less considered than yours.

            Basically I gave the battery 5v from a USB lead (from a wall-wart, not daft enough to risk the computer) for just a few seconds. The battery voltage must’ve recovered enough, because it then continued to charge when it was back in the phone-watch. So it probably was a protection circuit.

            It’s a useful tip, though as ever Youtube being Youtube, there’s people all over the place suggesting this as a way of fully charging if you lose your charger or whatever. Which is really asking for trouble, with Li-Ions being like little grenades if you’re not careful.

      1. That’s just the BGA packaged chip. No docs, no nothing with it. Do a hackaday write up when you put a board together and a dev kit and list it for less than a BeagleboneBlack.

        Here’s a tip even the coolest chip is only as useful as the software tool set that supports it. Right now it’s useless.

      1. Yeah, we just spent over $20K setting up to do BGA boards. That included the pre-bake oven, the optical inspection station, the hot-air station, and the jigs and stencils. We may have to spend even more on environmental controls for the area the work is being done in. It also helps to have someone with over 10 years experience working with them in a professional environment.

        BGA boards are definitely NOT DIY friendly.

        SMD components with Metcal soldering stations, optical inspection, and custom board fab is MUCH easier to contemplate doing at home.

          1. I’ve designed a few BGA boards myself and I always prefer the smallest components for my projects, not to mention I’ve replaced tons of BGA ics in mobile phones and other portable devices over the years for friends, family and other people I know. I’m not a professional (not to mention my eyesight is sheet), my job is in software 100%, I’m a DIYer and there’s plenty of DIYers who do that. I don’t see why people are afraid of BGAs and the most frightening thing is that most “DIYers” are afraid of 0805 components and QFP packages or even anything that has the word “SMD” in it. The meaning of “DIYers” these days became synonym of “newbie” or someone who likes to play with legos, or some hipster who read about this latest cool thing called “Arduino” on engadget and learned to use it so he can build some blinking lights for his tricycle.

          2. Getting the chip soldered for a prototype is not that difficult. If you apply enough flux, pre-heat the board and then hot-air the chip it will solder. Correctly, most of the time. Alignment is tricky, but the flux and solder will suck the chip in place, so you can be a little off.

            More annoying is making a board. Since the pin density is very high you quickly need 6 or 8 layers, which is expensive to produce.

    1. That’s just the technical brief. It will be hard to rewrite the document – not much you can change for basic electrical specs, GPIO tables etc. Might as well just write code from those document directly and magically have the code show up in pastebin..

      Besides, it is not enough to program the chip. What are more useful are the ones on the right hand side of the download page – the firmware document etc.

  6. This is really cool, there are a lot of cheap devices from China that are impressive, and would be great for hobby projects if there was some English documentation. $15 feature phones with an FM radios, and $12 pocket 802.11n access point on Amazon.

  7. This entire article seems to be a bit overly dramatized to me. There are solid reasons why these companies may not want to deal with the western maker market.

    First of all: Mediatek is actually a well reputed, large, taiwanese company and not some garage-style 5 person startup in Shen Zhen with pirated software. This entire “Gonkai” phenomenon does probably not apply to them.

    Generally: Many electronics companies work primarily with a key-account model and not at all with a distributor (such as digikey, mouser etc.). This means that they develop their devices specifically for a small number of high volume customers. The reason is simply that customer support in this market costs a lot of money. Revenue via a distributor could actually lead to lost money due to having to support many smaller customers. As simple as that. Same thing applies to many other companies as well. Have you ever tried sourcing the components you can find in an iphone?

    The situation with smaller chinese companies may be different. They are probably not able to offer the kind of quality certifications that are required to get into western consumers products. So they simply only support the local market and rely on word of mouth. And again: If their key customers don’t need a proper datasheet, you won’t get a proper datasheet.

    1. Gonkai is something Bunnie came up with

      Mediatek, Allwinter, Rockchip, Spreadtrum, RDA – they all operate same way in China. They offer _totally proprietary and super secret_ SDK that includes devboard, gerbers for the devboard + few variations, software stack full of magic binary blobs and software breaking licenses (modified gpl packages distributed in binary form). In order to have access to this magic bundle you need to be a 100% Chinese company with government development grants, established commercial Chinese owned partner, or in some form of official cooperation with big Chinese company that already is one of the first two. This is all Government regulated in order to bolster domestic businesses.

      Gonkai describes small fly by night Shenzhen/HK companies that obtain this secret sauce in less than legal ways (copy/ripoff/reverse engineering,friends at a company that has access to official stuff etc) and hack on top of that.

      Want to make 2 million cheap tablets? No problem, as long as
      -you subcontract, here is a list of approved Chinese companies that can design it for you
      -manufacture in china, here is a list of appro…..

      Want documentation so you can manufacture outside China using our _quad core $4 ARM_ chips? AHAHAHAHA get lost, good luck designing with Freescale $50 a pop in bulk imx6q. This is the reality of doing COTS/SOHO nowadays.

      1. I absolutely agree. Even if the MT6260 IC were $1 each in single quantity, the infrastructure is not there to allow, or support me in designing it into a new device for me to sell. It looks like you can do some programming of the MT2502 through the Arduino IDE, but I don’t know how much (considering it’s SoC makeup). I assume they would carry over some compatibility to the MT6260. But then, it was made to use in $12 cell phones, and I’m sure they sell a million of them, so I doubt they care about the relatively small hobbyist market. Perhaps in 2 or 3 years we will see low cost boards with these IC’s and I may buy one to play with. The advantage of the Arduino platform (for hobbyists at least), is the design was open sourced by a non-Chinese company. This allowed Chinese manufacturers to jump on it as soon as they noticed a market for low cost hobbyist development tools. And because anyone can design Blah-duino’s and shields, the competition brings the prices down. Not just due to increased sales volume of the IC’s used, but also by board manufacturers and distributors competing with each other.

  8. Perhaps the overlooked aspect is in Hacking as Engineering more than anything else. The IP costs of taking that SoC itself step-by-step from design to a usable, over the counter product may NOT be easily assigned neat billing centers. In plain terms- money flows in strange to us ways as this all gets born:>

    Never forget: Fabless Companies are still an evolving ecology! Fabless demands we take a step back from everything we ever thought of in economics to re-evaluate and embrace new ways. Even if it seems more like Kabuki Theater than Engineering.

    Profiting from Dev Kits seems outrageous and I openly despise it as stupid on many levels. Social Exploitation being a Cultural Shibboleth of Engineering back to Obstetrical Forceps…

  9. This sort of thing just goes to show you what Arduino did right. Even though people love to hate on Arduino these days, this is what they did right and why they became so popular and so widely imitated. They might not have the most advanced software, the cheapest boards, or the fastest 32-bit core, but this is what they did:

    – Make everything open-source.
    – Make everything simple and accessible to beginner users.
    – Use silicon that can be easily obtained from “public” small-volume distributors.
    – Make everything transparent, comprehensible, and well documented.
    – Don’t lock up stuff behind NDAs, or make it available only to the big customers deemed worth bothering with.

    1. Just about the only thing they did right are marketing towards a larger captive crowd that has more money than skills with a simple product that is easy to use. Kinda like the iProducts in a certain way in their trimming off the useful advanced feature (like code tracing, simulation and a good IDE) and heavy use of branding etc.

      The rest is really a matter of cheap Chinese knock off, cheap parts, cheap labor and cheap postal delivery. If you view it from a for profit business point of view, that’s a failure story. Thankfully, making money isn’t their primary goal.

      1. Sigh. What Arduino is this decade’s Heathkit.

        It provides a low barrier, easy to tinker Dev platform to make simple electronic/programming projects. Just like Heathkit, it was expensive because it well packaged and easy to use.

        That was the niche. Now, because of market pressure and such, Arduino has languished. They’ve seemingly quit innovating or bringing new tech as fast. Their compiler work has also stalled.

        Because of those things, it’ll be up to the next company who takes their place. MediaTek could, if they were willing.

      2. What you are suggesting is like asking Federal Reserve to do retail banking which is a totally different business. Yes they handle money, but at different scale and area.

        MediaTek is a chip manufacturer. If they like to deal with guys that mail order QTY1, they would have released full documentation already. May be someone that design board and sell them at Seed Studio, but a bit less greedy?

  10. I find it a bit ironic how we(americans) used to bemoan chinese companies of reverse-engineering, streamlining, and mass producing our technology then selling it here cheaper, ultimately undercutting the original developers. And now, we have this…

  11. I understood that idea differently. A bit od what-ifs: Mediatek sure does not bother with mailorders and such but what if they create a chip that will have all them handy interfaces built in from the start (canbus, hdmi, various serial flavors, sata and other whatnots) that are of interest to the diy/tinkering scene and couple that with enough multicores in Ghz range and enough ram/rom to power any project imaginable for the next 2, maybe 3 years? (we do not want things to go obsolete very quickly here) I mean a SoC exclusively targeted at this segment and not something from a tablet/phone world repurposed for arduino-area suff. Can someone make Mediatek do this? What is the minimum order someone needs to do for them to even consider making it? 100k, 500k, 1 M units?

    1. They are used to the consumer electronic market. The volume they deal with would probably be a per month/quarter/annual commit and not a single lifetime number.

      Product lifetime is very short. I have 3 very similar priced Lite-ON DVD burners purchased within a few week’s span from the same store that have different PCB, different MediaTek and optics inside.

      The only hope you can have is for them to open up some of the things like mt2502. Hopefully, that architecture is similar enough that future devices can be adapted.

      This particular SoC isn’t of particular use for hobbyists unless you want to play with low end phone using licensed frequencies.

      Or you can play with MediaTek-RaLink RT5350 $10-$15 WiFI router/development board with Ethernet, USB, GPIO etc which is well documented and supported for OpenWRT in the ISM band. I mentioned that near the top.

    1. I can see how you might read it that way, but I tried to specifically *not* imply that.

      If you look at the source article and the batch of phones sold with these chips, they’re clearly $10 throwaway phones.

      What I meant was, in China, you have *the option* of making cheap throwaway crap because it’s so cheap for them to do it. Meanwhile Westerners would struggle to find something affordable for a serious project (and forget trying to do something trashy and whimsical).

      If anything the point of the article is a reflection on how times have changed. Westerners are having to reverse engineer Chinese products in order to have modern hardware available.

  12. if it is open source then that is an instant green light for reverse engineering.

    in fact reverse engineering would help insure that proprietary stuff is not being re labeled as open source just to get it past legal checks and customs

      1. not necessarily for example lets take adobe photoshop where there is the so called heal tool witch the code that runs that tool is probably patented and copyrighted.

        ok lets take the open source photoshop like program gimp .

        lets say the folks of gimp reverse engineered photoshop and took the heal code and put it into gimp and called it something else.

        now gimp contains stolen code and they are calling it open source.

  13. Chinese manufacturers also have an advantage in shipping into the US. They can ship a package into the US for less than half the cost of me shipping a package of the same weight into the next city.

    The 2010 ePacket deal has lost the US Postal service millions. And makes small-scale commerce between the US and China a one way street due to shipping cost.

  14. If you have a huge manufacturing scene, like China, you will have a huge supply of surplus components, unsold stock etc. It makes sense in that environment, if you are a big Party boss, to encourage smaller businesses to ‘mop up’ these components and modules, along with a help with the engineering.
    I have no idea if this is a correct interpretation but I do know that there has been, for a very long time, a great emphasis on disseminating technical information in all manufacturing fields. This includes a very active book publishing industry and wide readership of homegrown technical journals.
    If “the west” is to match this wide dissemination of information we need a thriving manufacturing base at home. Reverse engineering interesting components and modules would be a very, very good thing but it is labour intensive.

    1. “that there has been, for a very long time, a great emphasis on disseminating technical information in all manufacturing fields.”

      Is this an april fool’s joke in advance? Oh yes the automakers spend millions and millions of dollars buylng and tearing apart each other’s automobiles. Oh yes you can find detailed information from every chemical manufacturer on how they manufacture their products. Oh yes the pharmaceuticals are just fine with disseminating the secrets to their products, after all look at how they let millions die from AIDS.

      “we need a thriving manufacturing base at home”

      Yeah we need to have air pollution just like China and we need to dump more chemicals into the rivers so that we can spend our taxes cleaning them up.

  15. It would have been nice if the HAD write-up had correctly described that the $3 part was just the chip, NOT THE ENTIRE BOARD. Probably would have collectively saved HAD readers hours of time, and lots of letdown.

  16. Chinese products like the ESP8266 and the MT2560 are overhyped. They might be cheap but come with little to no documentation, are not robust or at least proven to be reliable, and developing for them is nowhere near as easy as developing for the Arduino.and not to mention the shipping, yes its free but very slow!

    The problem is two fold;
    i)First someone or some group must dedicate a significant portion of their time to decode tha darn thing. By the time they do succeed, a new “Arduino-killer” shows up and everyone forgets all about the older kid on the block.

    ii) Second complexity. The arduino is a 16MHz 8-bit microcontroller. That’s it. Sure you can attach other shields to it but at its core it is just an 8-bit microcontroller with a basic API. Many of these new “Arduino Killers” are so much more. To start with they all contain 32-bit micrcontrollers/processors. That alone ups the complexity significantly. They also do GSM, WiFi bluetooth yadda yadda yadda.This complexity makes it so much harder to decode and write a simplified API around.

    Even when an “Arduino Killer” is decoded, someone has to not just build an API but debug it, document it, build an IDE with a compiler and a downloader around it and make all of these resources available on a website. The API/ tools must be continually maintained and a mechanism for fixing bugs developed. Someone might even have to build a nice standardized breakout board for it. And even if this is all achieved, who guarantees that these parts will be sourced for a long time? Sourcing these parts can be problematic since they’e sold exclusively in China and even more importantly are not necessarily designed for the hobbyist market.

    As a person who wants to build cool & creative projects without the pain of having to create my own compiler or dev environment or write firmware for hours based on sketchy examples and unreadable documentation, I really find these chinese products unappealing.

    Instead I prefer the $19 Spark Photon for IOT stuff. It is not out yet but March 2015 is around the corner. Also for serious IOT stuff a decent Linux SBC can do wonders. Something like the Odroid U3($65) or C1($35). Now these board can run anything under the sun, Node.Js server, Tornado(Python), Nginx, Apache, Lighttpd e.t.c. For the IO, many cheap Linux SBCs have GPIOs on them or one can use a Linux SBC with an Arduino Nano as a IO slave processor thanks to Firmata. (BTW the IO slave approach is used in the UDOO and the Intel Edison…so it’s not that bad.)
    And since hobbyists buy stuff in low volume, a bump in price even from $4 to $20 still makes for an affordable product; especially when the $20 version includes everything (IDE, compiler, API, downloader, documentation e.t.c)

    Other excellent frameworks for Hobbyists include the mbed framework/boards as well as the Tiva C Launchpad + Tivaware + Energia.

  17. I am with halherta on most things, except that for me personally the ESP8266 is interesting at it’s current state of community support. But often costs for additional parts are also overlooked. Most projects don’t use a development board standalone, right? And a lot of projects might also use multiple ADCs, want to have easy 5V-interfacing, other peripherals, etc. …

    So comparing the AVR/Arduino to a completely undocumented non-5V tolerant SoC that might not even have ADCs is like comparing Apples to Oranges. Besides, ‘Arduino’ is basically a documented high-level API for AVR controllers, the MT6260 is a chip alone without much documentation. The Raspberry Pi would be at least a closer comparison.
    When choosing any hardware, it’s important to first look at the application you want to do! It’s usually not about having a microprocessor sitting somewhere and doing number crunching for no reason, where only the general-purpose processing performance matters. These things are not just miniaturized desktop PCs…

    Microcontrollers like the Arduino/AVR are mixed-signal controllers, that both provide analog peripherals like ADC’s and digital ‘accelerators’ like PWM peripherals. They typically can run at 5V or 3.3V / 5V-tolerant. In this range you also have the MSP430 or PIC-controllers, and if you additionally need higher processing performance (FPU, DSP support functions..) and e.g. 12bit ADCs, USB OTG the STM32F4 (3.3V / 5V tolerant). Most of them are also as well documented that real industrial applications can use these components over a longer time.

    Then, of the hobbyist Systems-on-Chip’s (SoC) we mostly talk about here, most of them are only consumer-grade, with limited or non-existent documentation, more fragile 3.3V-only, etc. The Raspberry PI is also one of them. It has a dedicated graphics / DSP accelerator core and can decode H264, but doesn’t even have a single ADC, and documentation overall is still lacking. The same thing probably goes with the MT6260, it’s also relatively fast and has an EDGE/GSM modem, otherwise it’s only interesting if you need its general purpose processing performance but not much more (or does it have e.g. ADCs?).
    The ESP8266 is interesting because of its wifi module in a small package and small-$, and maybe about the processing power of the STM32F4. It seems to become more and more generally useable though (and has at least one ADC), so from my view it kind of blurs some lines.

    It only makes sense to save some extra $ if you use a higher number of them, or the limitations in documentation are not too bad for your specific application (for me that’s fine with the esp8266…).

    From the chips I named here, personally the MT6260 is the LEAST interesting for the applications I have currently in mind…

  18. Maybe this project should start up again since rumors from SZ are that the MT2502 will be discontinued and will replaced by MT2503 that are based on MT6561 thats runs the same FW as MT6560 and are out of reach for us out side China and cant do orders of 100.000+ chips

Leave a 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.