Atmel AVR 8-bit MCUs are expensive even compared to other MCUs with superior specifications. Suppose you can redesign Arduino today. Which MCU would you choose?
How soon does it have to ship?
Atmel SAMC chips are 5V ARMs, but not quite available.
Cypress PSoC4M chips are also 5V ARMs, and also pretty new. (PSoC5 is somewhat older, and rather more expensive.)
(Note that Cypress HAS arduino-compatible eval boards for both PSoC4M AND PSoC5lp)
I should note that the 28-pin AVRs are pretty close to as good as you can get (IMO), if you want something that can be built from parts by the target audience for Arduino (like the original Atmega8 Serial version.) There are some PIC32 chips in DIP that are "interesting" (and have arduino-compatible boards available as well.) But they seem to be rather less friendly than AVRs.
lightaiyee:
MCUs with superior specifications.
Would that mean they could flash my LEDs better or drive my servos better?
Simplicity is vastly underrated.
If it works, don't fix it.
...R
Hi,
Simplicity is vastly underrated.
If it works, don't fix it
Exactly, you want to soup up the micro-controller, it won't be a micro-controller anymore it will be a Raspberry pi.
Tom..... 
After a void of a couple of years I restarted my tinkering hobby and looked into other boards with "most power" but I figured out that most projects don't need more than an Uno, a Mega at most.
If NASA could get a rocket to the Moon and back with this kind of power, I can drive all the LED's, servos and what not as well.
So far it is working pretty well for me. It would be great to build something with an ARM CPU but I have yet to come up with something that needs that kind of power.
AVR will be enough for most tinkerers, if you surpass that level you will probably want to code it directly instead of layer upon layer of translations and interpreters.
It would be hard to fit the stuff in the image in Reply #5 inside my OO Gauge model railway locomotive.
...R
Robin2:
It would be hard to fit the stuff in the image in Reply #5 inside my OO Gauge model railway locomotive....R
damn ... those words ...
Robin2:
It would be hard to fit the stuff in the image in Reply #5 inside my OO Gauge model railway locomotive.
It looks as if it would be hard to fit inside a 7 & 1/4" gauge loco.
[mc68hc11evbu.jpg]
Really? 68hc11 cpu, 8k rom ("monitor firmware"), 8.25k ram (program or code), 8k eprom/ram (nominally program), 512bytes EEPROM, two rs232 serial ports, some gpio, and a bunch of support circuitry? That should all fit on a single chip these days...
I could get behind a coldfire-based board (http://www.netburner.com/products/core-modules/mod5213)...
(and: still waiting for x86 microcontrollers to show up...)
lightaiyee:
Which MCU would you choose?
Light the blue touch paper and stand well clear .... ? ? ?
...R
I guess modern things are too boring, especially for hobby
e.g. when a super bloated MPU with DMA, DSP, USB, FPU, etc etc etc
is used to …. blink bulb lamps on the xmas tree X___X
the above board makes blinking bulb lamps on the xmas tree like a piece of cake
not too many peripherals to care about, the assembly is comfortable
and I am feeling fine
I would like a MEGA in DIP and all pins exposed, and a board with a ZIF socket for it.
I have a UNO with ZIF like this UNO WITH ZIF and I love to work with it
robtillaart:
I would like a MEGA in DIP and all pins exposed, and a board with a ZIF socket for it.
That sounds very sensible.
And it could probably be done with a small breakout board.
...R
Robin2:
And it could probably be done with a small breakout board.
Exactly!
Rationale: If the chip is fried, you only need to buy a new MCU
I am slowly moving to ARM M4. ARM offers a lot more bang for the buck than AVR, and that's exactly were Atmel is losing space.
The thing with ARM is that you have to choose one vendor and stick with it, because even although it is all the same structure, development is very different from one another. Even within the same vendor, code is different from one processor to another. So, in short, development with ARM is a royal pain the ass, but once you master it (which I am light-years away from doing), it pays off. The ICs are cheaper, a lot more powerful, have a lot more IO, a lot more memory and so on. The downsides are: there is only one breadboardable chip on the market, the best devel tools are quite expensive, the libraries and HALs are horrible.
So, if I were starting the entire Arduino ecosystem today, I'd do it around an ARM M3 chip, possibly the SimLabs SIM3U1, which is extremely powerful.
AlxDroidDev:
I am slowly moving to ARM M4. ARM offers a lot more bang for the buck than AVR, and that's exactly were Atmel is losing space.
I am genuinely interested to know what sort of projects you work on.
My guess is that they are very different from the sort of thing an Arduino Uno or Mega is intended for.
...R
St micro :
- cheap,
- Some 8bit models are pin to pin compatible with ARM STM32 : same board can be equipped with 8 bits or 32 bits.
Some can start with the 8-bit and pass easily to 32 bits ARM . - Vdd = 3.3V with numerous 5V tolerant I/O : very useful for connecting 5V circuit with 3.3V circuits without levels adapters..
Dip package: stop dream, we are in 2016, soon most IC will be bga : .
Unfortunately, technology is moving forward without asking us if we agree.
The only ones who have influence are the big manufacturers.
Keeping in mind that fast micro have better performances with SMD package.
I'm just leaving this here.
Blinking lights on a Christmas tree. A 2 transistor oscillator was more fun than arduino and learned more; -)
xmega128A1
8bit, 32mhz, DMA, etc, etc, etc...
simple, fast, perfect for Arduino world.