Slashdot is powered by your submissions, so send in your scoop

 



Forgot your password?
typodupeerror
×
Hardware

Linux Microcontroller Board 28

WillWare writes "Here's a nifty microcontroller project being done by Ryerson Amateur Radio Club in Canada. They are building a SIMM board with a Motorola Dragonball (same processor as the Palm Pilot), 4 meg of flash, 8 meg of DRAM, some digital I/O pins, ports for Ethernet and RS-232, and able to drive a 320x240 LCD panel. This board is intended as a target for their MMU-less Linux port, which has previously been running on Palm Pilots. There has been mention on the mailing list of the possibility of running a Python interpreter on this board. This would be a huge win for rapid app development on embedded controllers. "
This discussion has been archived. No new comments can be posted.

Linux Microcontroller Board

Comments Filter:
  • Give me Linux console on a Mac Plus booted from floppy ;)
    Nice clear tiny screen big enough for a terminal, cute form factor, only draws 40W or so including the CRT (hell, some pentia draw more than that all by themselves!), up to 4M of RAM (more likely one, but you never know) and once they make it past the first ten years they're pretty much good for a hundred... why boot from a floppy? (________) meaning SILENCE. It's weird to be running a computer that makes NO noise, no fan (convection cooled!) no drive (floppy spins up as needed) no anything... very peaceful, now imagine instead of System 6 MacOS and teachtext or something, Linux console and vi... perhaps telnet... you have 800K of floppy to work with, and of course if you can switch them like you can with MacOS you could just boot it and then leave it on, only draws as much power as a dim bulb (I'd seriously favor white-on-black text for this too, retro charm and anti-Mac-ness and lower CRT power consumption...)
  • A bulkier, _amber_screen power sucking antique PC _just_ because you want to call Apple 'crapple'? ;P and give you 'those' what? Somebody has never seen the apple Tech Info Library archive for stuff over eight years old. Pinouts for all those funky cables etc. ad infinitum. Beating on their historical closedness is just FUD now, half the PC parts from certain vendors are every bit as closed.
    Sorry, I just had to squawk there- a Plus would be every bit as good a Linux machine as any six microcontrollers. None of them is gonna run X! Enlightenment is RIGHT OUT ;) besides, what I was talking about _was_ a glorified term. Just a glorified term running Linux, with a cute form factor and a nice clear little screen and peripherals that are better than you can get today and last longer...
  • The pages for the flux oskit indicates that currently it only works on the x86, which is not an ideal target for a system this small.

    -Peter

  • Money cost probably isn't the ussue there. Board real estate and power consumption are more likely issues. Plus, with a max of 8M, there's not really much to manage. Finally, swapping should never happen on an embedded microcontroller anyway.

    It might be an interesting exercise to equip the thing with bank switching persistant memory (with non selected memory being powered down) at some point though.

  • It is possible to produce good code. Code that does not crash itself or the system. All MMUs do in this respect is provide a convenience for the developer so that he does not have to reboot every time that his still buggy code crashes itself or something else. MMUs do have other redeeming qualities, but in reference to protecting the system from buggy code, it is just a convenience.

    With this system residing in flash, it is practically an instant-on or instant-reboot setup anyway.

    Besides, nowadays you can write with modern languages that protect you from pointer errors and missing free() calls, see gcj.

  • The real question is why half the readers can remember that, but the posters can't?

  • I can understand why you would want a MMU-less version of Linux, but when you are building the machine yourself I would think the MMU would be worth any extra cost.

  • This doesn't seem like that portable of a solution. How about this:

    Remember the old, original, amber screen Compaq luggables? About the same size (maybe a little more power consumption ;), better keyboard IMHO.

    There's already a Linux/86 port that should run. You'll need to manually port the MMU-less version to the Mac Plus. (Have fun getting (cr)apple to give you those)

    Don't get me wrong -- I like Macs, but this isn't a good linux machine. Now, it might work as a serial console...
  • Probably because they spend all their time combing through 100-200 summissions per day, and tweaking Perl, and never actually get to read anything. Poor guys. :-)
  • ..turn an old 30-pin simm memory expansion board
    into a parallel processor...
  • Add the proper transceiver(s) (or stick to 10bT)
    and this little thing would be great for network
    analysis.. tcpdump, netwatch, etc.. nice little
    interface for them. It'd be great for network
    engineers who need to look at ethernet frames
    away from their desk at low cost..

    Using this for wearables interests me, too, but
    I don't think it has enough horsepower to do
    the kind that I'd prefer..
  • Add the proper transceiver(s) (or stick to 10bT)
    and this little thing would be great for network
    analysis.. tcpdump, netwatch, etc.. nice little
    interface for them. It'd be great for network
    engineers who need to look at ethernet frames
    away from their desk at low cost..

    Using this for wearables interests me, too, but
    I don't think it has enough horsepower to do
    the kind that I'd prefer.. (primarily audio input and output, palm pilot serial console for everything else.. "Computer, read email"..)
  • Linux is fine, but I think Flux [utah.edu] OSKit [utah.edu] would be more appropriate for this project.
  • Oops, sorry. But from the mailing list activity, it looks like they are a very few weeks away from actually having boards done. So it's an interesting time for this project, and probably worth reposting (even if some people get pissed off).

    The idea of scripting language interpreters running on embedded controller boards is interesting, and that didn't get mentioned last time. People used to put Forth interpreters on controllers to do in-system debug and testing, but with faster CPUs and more memory, we can use more convenient and capable languages.

    There was another reply to your post, from somebody who seemed to be really irate that something would be re-posted. I confess to being mystified, but I also can't understand how people can flame every tiny tweak in Slashdot administration.

  • A USB interface would be a wonderful thing, as you point out. But what the board designers appear to be doing is relying on the strengths of the dragonball controller, to minimize chip count and cost. The dragonball has an on-chip dram controller, so they can use DRAM, and it interfaces pleasantly to flash, so they can use flash. I have a dragonball product brief handy, but not the description of the Ryerson board at the moment. The dragonball doesn't appear to have an Ethernet port, so they must be doing something special for that, probably catering to some need of their own. (Can't fault them for that! They're the ones bothering to design the board in the first place.) Maybe some future rev of the board will have a USB port. The dragonball is old enough that the future of USB may not have appeared stable at the time of its design, otherwise Motorola would have probably thrown it in. USB's rise has mostly been in the past couple of years.
  • Hrmm.... Seems another good Microcontroller project would be one using a ColdFire CPU... like the MCF5206E or the MCF5307. They support up to (either) 256MB or 512MB of RAM, and have tons of features that would be perfect not only for a simple Linux project, but maybe even for embedded wearable computers. Any thoughts?

    If it comes from man, it will fail.
    If it comes from god, It will succeed.

  • I've seen PC-on-a-stick and PIC-based SIMM products before. Has anyone checked to see if any of these products are even remotely pin-compatible with each other?

    Wouldn't that be nice?

    Vik :v)
  • With Linux already chewing at the Windows NT market on the server end, this is going to start nibbling away at the WinCE market at the embedded/palm top end.

    Cool.

    (I'm going to get me some of these. (Too bad the Ethernet is 10bT and not 10b2 -- the hub is going to be just as big as the rest of the cluster. Wearable supercomputer, anyone? :) Now, how about a small cam interface...)
  • This would be perfect for wearable computers... it could be placed on one's belt. And it draws *microamps*. Amazing!
  • How about us people with the Amiga or Atari's which don't have MMU's installed unless you shelled out loads to upgrade them years ago?

    Edge
  • While their hardware is really cool looking (the 10Base-T really gets my attention). I'm thinking that you could get a lot of development done (in terms of I/O and stuff) ironed out with regular PalmPilots. Does anyone know if somebody is selling used pilots? Even the origional non-backlit ones could be used if they had TRG flash upgrades. I could see shelling out $50 for one of those for prototyping. Possibly one of the cheepest Linux boxes ever :-)

    - Mike

  • If you take a look at the MMU-less Linux link in the posting (http://ryeham.ee.ryerson.ca/uClinux/) you will see that they are also working on porting Clinux to the Coldfire. They are using the eval boards from Motorola right now, but maybe someone will get inventive & create something like the SIMM project using Coldfire too. There is always hope.
  • If the MMU is missing, how do you stop buggy programs interfering with each other or the kernel? How do you stop them accessing hardware they're not supposed to? I just don't see how a modern OS like Linux is going to work without memory management hardware. Someone care to explain?

If a thing's worth having, it's worth cheating for. -- W.C. Fields

Working...