out of curiosity, i'd be interested to hear if the design that you are creating is intended for anything approaching a decade-long support and end-user lifespan.
First of all, I want to make it clear that my area is closer to industrial and sometimes automotive than consumer. Also, I am sure that I don't have your experience and I don't see all the details that you have seen - and large customers that expect the CPU cards.
And, the usual disclaimer, this post will not bring any creative ideas, it will explain my position, so feel free to stop reading here...
But what I have learned in recent years is that there is no such thing like "one size fits all". Even if something fits well at the current time, this is temporal and will change in several months or an year. In our devices, in past 10 years, we had passed over several form factors of "processing boards" - PC104, ISA, ETX, and some other, our own too. And what we got like benefits is limited to: - possible second source - I hope the EOMA will be able to attract more that one designer and producer of CPU cards - possible upgrade of the CPU card - this works sometimes, but never for 10 years - maybe 2-3-5 years is realistic ( I mean in industrial, in consumer 1 year top); with current CAD tools it is not so hard to redesign to other form factor (you know it, you have made many desing recently) - save us some investment in development and gives us faster time to market - this is true for something like 100-5000 pcs per year, but not in mass market
Nope, I cannot even imagine that something (commercial) will live for so long. It might run fine but customers are always demanding new features, better perfomance (even just better design). And even if the lives than it will not be upgraded. Maybe it will be glued to the fridge for recipes in the kitchen, or will become garage door opener with twitter notifications ...
Components are getting obsolete, sometimes looking around the whole world to get some more pieces to build next batch is harder than to redesign (and at the end you have to redesign). Let's imagine that I get a product with EOMA slot - a good one with FHD, 1000:1 IPS and so on. I could imagine that in 2 years (or, more likely, 12 months) I would like to get a UHD display, probably AMOLED, probably one that consumes 50% of the power of the old one, probably better colors, maybe something like e-ink, maybe flexible... Probably a new wave of special batteries will be out there in 2 years... My point is (and probably I am biased by my lifestyle) that I will prefer to put the old one in the basement, or on the e-bay, and just go for the newest. Or, if you prefer, recycle it. My USB connectors will be worn out (yes, I use tham a lot, 2xJTAG, logic analyzer, serial converters and so on), my DC jack too. The electrolitic capacitors in the device will reach their 5000hour lifespan. Flash erase cycles will pass the limit. The keyboard will go bad, the screen will be scratched. (now I recognize that I can overcome this on my current laptop if I use a docking station - some 200USD for some plastic and connectors? But quite close to "base product" around a EOMA card?...) There are 2 type of user - one that would upgrade and others that will not. First one will like to get 4G RAM instead of 2 - so with EOMA they have to drop the complete card, istead of swapping a SODIMM? Others already have 8G of RAM, but want a faster CPU or 3D - they will have to dump (or re-buy) the 8G at the old board. It is similar to desktop-vs-laptop trends. One gives you limitless upgrade, other gives you mobility. The other players like smartphones and tablets are on the extreme side of "hey, remove the microSD slot to save some space, solder the battery to make it 6mm". So I am wondering were does the EOMA product go? Not in smartphones (hey, google, modular smartphone? worst idea I have seen from them since long time). Hard in tablets, where thickness, weigth and design are important. And that usually finish their life with cracked screen when someone drops them or sits on. Laptops? I like my i5 with 12GB RAM, don't count me in, next thing will be i7 with SATA6G SSD (at least, maybe something newer and faster then 6GBits/s). Desktops? No. NAS devices and home servers? Maybe, but CNX is full of news for android tv devices, now with RK3288, 4K UHD... DIY market? yes, why not, but do I need the PCMCIA housing and (limiting) connector when I put it on my breadboard, beside the arduino, to create a fancy LED flasher with "social services"? What's wrong with Olimex boards for this project? Aha, one is discontinued and the next one is not pin-compatible? Ok, redesign and be more careful. But this is nice opportunity to fix those hair wires below, or to add a header for second UART. There is nothing to fix or improve on this board anymore, the whole redesign it just to use another cpu card? Dream on, never on this planet.
It is not only the CPU and the memory, everythings get old and our of time.
I also use some old devices. e.g. one thinkpad. I liked it a lot - it is almost 10 years old. The keyboard was fine until I started using it a lot - obviously the previous owner had used a docking and external keyboard that saved it through the years. But 2 years in my hands and it is nothing special any more. And dirty, too. Display is slow, low res, not as bright as current ones.
What is probably a moving force behind this project, I believe, is the usage of the Linux, and huge market of Android devices. This means that a lot of people work hard to design cores, make chips, make hardware, port software. And this efforts come back as cheap SoCs with ready made android support.
You have given one example to support EOMA - everybody can push the switch and plug out the PCMCIA card, that put the other one. But I don't see my father, for example, doing this. For the same reason why only geeks (and normal engineers like us) are replacing SODIMMs, HDD, SSD on their laptops. So, what happens if somebody gets a CPU card and it does not work in Miguel's product? The same what happens when I buy a normal product and it does not run in my house / car / pc - I send it back (RMA). In current scenario, I don't see why his console will not bring a message - hey, I know this cheap CPU cards, but I cannot work with it! Please, open a web page and get the list of certified boards, or just order through our store, or contact your seller for replacement. That's all, simple and easy. Because if he does not refuse this card, he will start getting other question - hey, my wifi is damn slow, my audio does not work, my USB transfer to external HDD is slow? What? It is because of the new cpu card? That I want a replacement! I checked your web site, you have removed the list of compatible EOMAs? What? You had to do this because the standard says that it should run with all? But now you will pay the return shipment! You had to warn me that this is not going to work! If I only knew this issues I wouldn't order this, so that is now your problem, pay for return and find me a new one - and I am not going to pay anything above the original price!
I hope that there is enough user base that can benefit of the EOMA. In industrial area similar solution might work - if only it was designed to support CAN, Profibus and other industrial communication standards (more pins!). It it was "board level" product that 200 pins are there. Yes, non technical people cannot replace it, but hey, some people cannot work with computers either! And in fact similar standard are used very often here.
Again, saying "my point", I still believe what I said long ago, even before this list existed - open source (HW and SW) system-on-module using those cheap ARM SoCs is a good idea, but I see better marker for regular sodimm / qseven or even SMD soldered "cpu boards". That simplify the design of the end used products and don't add restrictions like "any board on any product". Maybe closer to the Olimex things - not much noise but several well selling boards. And more to come. Designed to allow entusiastic people to make their "startup".
2014-08-10 15:41 GMT+03:00 Luke Kenneth Casson Leighton lkcl@lkcl.net:
On Sun, Aug 10, 2014 at 12:04 PM, Miguel Garcia gacuest@gmail.com wrote:
2014-08-10 10:52 GMT+02:00 Luke Kenneth Casson Leighton lkcl@lkcl.net:
if someone is buying a product with a $2 SoC then they can expect to get what they pay for. the software will have to be configured to take into account that one of the interfaces is slower and e.g. ramp down the WIFI speed. or disable it.
So we have 2 options:
Option A:
1st USB + 3.0 ----> USB Host 3.0
plus 900mA power requirements (0.9A) which the AXP209 can't handle, so you'll need to replace that (which means finding a suitable replacement plus example circuit)
2nd USB ----> USB HUB -----> USB WIFI -----> USB BT -----> USB Audio -----> USB STM32F072
Option B:
1st USB ---> FE 1.1 --> USB Host 2.0 --> USB WIFI --> USB BT --> USB Audio
2nd USB ---> STM32F072
... where you can use the AXP209, it can handle the 0.5A load for the USB 2.0 Host, you have plenty of example circuits to use.
What should we use?
Option A is the most complete, but if option B is the most compatible...
take power requirements into consideration i think it's fairly clear which option is most favourable.
Thanks.
arm-netbook mailing list arm-netbook@lists.phcomp.co.uk http://lists.phcomp.co.uk/mailman/listinfo/arm-netbook Send large attachments to arm-netbook@files.phcomp.co.uk
arm-netbook mailing list arm-netbook@lists.phcomp.co.uk http://lists.phcomp.co.uk/mailman/listinfo/arm-netbook Send large attachments to arm-netbook@files.phcomp.co.uk