ARM-based Arduino Due - where??

Graynomad:
This subject has gone very quite of late, people have either given up and bought a Pi/Beagle/whatever or they just lost interest. I suspect though that when it is finally released the interest will be rekindled.

With the number of low cost ARMs coming out, I suspect they've lost the momentum, and by the time Due finally comes out, it may be a case of too little too late. Lets see, of just the ARM projects I know of:

  • Raspbery PI, two models, $25/$35, and now you can order as many as you want, runs Linux;
  • Teensy 3.0, kickstarter project that ends in 7 days (already at 800% funding), models start at $22, uses modified Arduino s/w;
  • Galgo, kickstarter project that ends in 17 days (not yet 100% funded), models start at $19, has builtin debugger, and its own development environment;
  • Mbed, two models at $45/59, uses a development environment based in the cloud
  • Beagleboard/beaglebone, $89/$149, runs Linux

I suspect they've lost the momentum

Probably, that said I think there's plenty of scope for a "no bells and whistles" ARM board, I for one don't need or want all the crap that comes with Pi et al.

And of course anything with the "Arduino" name will have a head start.

BTW, I can't find Galgo on Kickstarter, can you provide a link?


Rob

The simplicity is what's got me sticking with (hopes on) the Due too (as with the past Arduino boards), versus RPi and Beagleboard, etc.

And the already established Arduino user community, and the body of work that has already been done -- there are a lot more users, and lots of code, project ideas, and libraries that are already existent that one can work with and tweak to create further new things.

I agree RPi is getting there but for simpler, one-thing-at-a-time applications -- that need the horsepower but not the extra HDMI output for instance -- I'm still waiting for the Due.

Plus after using the Due board as a prototype, I can then easily use the design to make my own tailored version of it when I need to adapt things to my final application.

@Graynomad:
This is Galago:
http://www.kickstarter.com/projects/kuy/galago-make-things-better?ref=category

And the Teensy 3.0 looks pretty exciting too:
http://www.kickstarter.com/projects/paulstoffregen/teensy-30-32-bit-arm-cortex-m4-usable-in-arduino-a?ref=category

Thanks for the Galago link.

Man it's getting so you can't keep up.


Rob

Graynomad:
Man it's getting so you can't keep up.

I like to think it has always been like that; but I was not aware. 8)
Best regards
Jantje

I dont even know if I want it

ARM chips are fast, but lack the grunt of the AVR, there are plenty of arm chips out there, they are not hard to work with, but when you need to nail 25ma at 5 volts you really feel their lack of micro-controller-ness

its a cpu with some embedded features, avr feels more like glue to get stuff to work

MichaelMeissner:
With the number of low cost ARMs coming out, I suspect they've lost the momentum, and by the time Due finally comes out, it may be a case of too little too late. Lets see, of just the ARM projects I know of:

  • Raspbery PI, two models, $25/$35, and now you can order as many as you want, runs Linux;
  • Teensy 3.0, kickstarter project that ends in 7 days (already at 800% funding), models start at $22, uses modified Arduino s/w;
  • Galgo, kickstarter project that ends in 17 days (not yet 100% funded), models start at $19, has builtin debugger, and its own development environment;
  • Mbed, two models at $45/59, uses a development environment based in the cloud
  • Beagleboard/beaglebone, $89/$149, runs Linux

There are the Olimex as well and Cutieboard

I dont even know if I want it

ARM chips are fast, but lack the grunt of the AVR, there are plenty of arm chips out there, they are not hard to work with, but when you need to nail 25ma at 5 volts you really feel their lack of micro-controller-ness

its a cpu with some embedded features, avr feels more like glue to get stuff to work

True, I was just looking at the Maple Mini and its limited in terms of current - which I assume makes it generally easier to kill/disturb than the AVRs.

Duane B.

rcarduino.blogspot.com

but lack the grunt of the AVR,

Yes you get used to being able to hang just about anything directly off an IO pin, even small relays. Not so with ARM chips.

Why is that so? It's a function of the IO and that's up to the chip manufacturer, nothing to do with the CPU core.

LPCs have a few high-current pins but not all of them, is it a limitation of the package re heat buildup?


Rob

Of course it depends on how it is packaged of how many pins you have for I/O. But I was curious:

  • Tweensy 3.0 - 34 digital pins (10 shared with analog pins), 10 analog input pins with 13 bit accuracy, 10 PWM outputs, UART (3), DMA (4), i2c (1), spi (1), i2s (1)
  • Galago - 25 gpio pins, 6 analog input pins with 10 bit accuracy, 10 PWM pins (6 with high speed timers), spi (1), i2c (1), UART (1)
  • Mbed NXU LPC11U24 - 40 pins, 6 analog input pins, 8 PWM pins, spi (2), i2c (1)
  • Itead Simplecortex - 40 pins, 8 analog input, i2c (3), spi (1), UART (4), PWM (6), DAC (1)

I think output drive has a lot to do with the transistor size. 8-bit CPUs tend to be made on older fabs with big transistors. ARMs tend to be made on modern fabs (the fab is probably less of a process than the fact that the chip layout is designed for a certain transistor size, but the same principle holds.)

The people that have any real knowledge aren't saying anything, but I'll point out that the World Maker Fair in NY is a little over a week away. Last year's NY WMF was when Due was announced. The spring Bay Area MF was when they solicited beta test sites...

@westfw:
Good point on the Maker Faire. I'm not gonna get my expectations up, but the timing does seem to be about right! And they do have this on the 2012 Program schedule for Sept. 29:

"1:00PM ยป The Arduino Hour with Massimo Banzi (co-founder of the Arduino Project) & Alf-Egil Bogen (Atmel Corporation)"

I think westfw is a good detective :slight_smile:

@Massimo:
Glad to hear it! Looking forward to another great presentation.

Better hurry. Both my teensy 3.0 and my Frescale Freedom also shipped this week...

Sounds promising.


Rob

westfw:
Better hurry. Both my teensy 3.0 and my Frescale Freedom also shipped this week...

My teensy 3.0 will ship in the second batch so next month sometime. Massimo better watch out or the Teeny 3.0 may indeed become the new Due. :wink:

Possibly the problem with the 'Over-Due' delay is the result of a problem identified in the book 'The Mythical Man-month' (or some such title like that), where the more people you throw at a project the more time it take to complete it. Paul working alone might just have had too large an advantage? :smiley:

Lefty

retrolefty:
Possibly the problem with the 'Over-Due' delay is the result of a problem identified in the book 'The Mythical Man-month' (or some such title like that), where the more people you throw at a project the more time it take to complete it. Paul working alone might just have had too large an advantage? :smiley:

Lefty

That is known as Brooks Law (Fred Brooks wrote the MMM book). Here is the wikipedia entry on MMM The Mythical Man-Month - Wikipedia. However, Brooks Law would only come into effect if they added more people.

I have no knowledge of the Arduino team, but I have worked in the high tech industry for 33 years now, and I have seen many late projects. I suspect the second system effect in Mythical Man-Month may be more of the cause: The second-system effect proposes that, when an architect designs a second system, it is the most dangerous system he will ever design, because he will tend to incorporate all of the additions he originated but did not add (due to inherent time constraints) to the first system. Thus, when embarking upon a second system, an engineer should be mindful that he is susceptible to over-engineering it.

If not the second system, I suspect part of the issue is just the fact that DUE is a completely different system than AVR. Do you do things like map the Arduino pins to ARM, or do you force users to have various #ifdef's in thier code. I haven't looked at the details of either ARM or AVR chips, but it may be some things like interrupts, PWM, sizes of the 'int' type, inline assembly, etc. are completely different.