AVR roadmap

Go To Last Post
36 posts / 0 new
Author
Message
#1
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Hi!

Does anyone know about the current AVR roadmap, and planned new AVR devices? Or Xmegas beeing fixed?

thanks

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Well I would suggest you contact your Atmel sales rep or FAE but are you actually going to trust anything they say anyway? They are not exactly known for the accuracy of their roadmap forecasts. So let's just take a punt and say there'll be an 8MB AVR with 512KB of SRAM, twenty UARTs and an embedded LCD panel controller - but I can't tell you exactly when! :-)

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Well I would not belive everything they say, but I wanted to know, if they have an idea what is coming. I would be interested in wheter they planning some super stuff, nice peripherials etc.
And well reliability, the UC3C for example was 3 years late, but it came finally.
So I would naturally add 2 years at least to what they tell, but if they don't even tell anything that's really bad. :) Because this means to me, that they haven't even started thinking about it yet...

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Quote:

means to me, that they haven't even started thinking about it yet...

Or they have learned there lesson and will keep quiet until they actually have something made of silicon rather than unobtanium.

But what's missing anyway? I think there could be more USB devices and perhaps some in DIP package. But in AVR8 (tiny, mega, xmega) what's missing otherwise? I guess we all have wishlists but I guess it's a question as to whether Atmel agrees and thinks those ideas would make money.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

If I were Atmel (and, fortunately for everyone, I am not), if I were contemplating something really different than current offerings (say, built-in ethernet with tcp/ip stack or 200MHz clocking for here-to-fore unknown ZMega chips), I would be very tight lipped about the whole thing. I would NOT want a single competitor to know that I was even thinking about something so big.

In part, if I let it out, my previous track record of meeting new product targets suggests that somebody else would beat me to it (it, being product in customer's hands).

So. I would not tell a single soul.

Jim

Jim Wagner Oregon Research Electronics, Consulting Div. Tangent, OR, USA http://www.orelectronics.net

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Quote:
I would be very tight lipped about the whole thing. I would NOT want a single competitor to know that I was even thinking about something so big.

No that is not the way they operate.
They sell an idea and if enough want want to buy it, they go home and try to make it :)

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

But almost every competitor now as USB, CAN, ethernet, lots of usarts, multiple SPI's and I2C's, Atmel seems stopped in time..

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Quote:
But what's missing anyway?

UARTs in the tiny chips :(

______
Rob

Scattered showers my arse -- Noah, 2348BC.
Rob Gray, old fart, nature photographer, embedded hardware/software designer, and serial motorhome builder, www.robgray.com

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Graynomad wrote:
UARTs in the tiny chips :(

Good point.

I think Atmel have been hurt by the inexpensive 32 bit offerings of their competitors. i.e. NXP.

They probably are working on this and if not they should be.

"I may make you feel but I can't make you think" - Jethro Tull - Thick As A Brick

"void transmigratus(void) {transmigratus();} // recursio infinitus" - larryvc

"It's much more practical to rely on the processing powers of the real debugger, i.e. the one between the keyboard and chair." - JW wek3

"When you arise in the morning think of what a privilege it is to be alive: to breathe, to think, to enjoy, to love." -  Marcus Aurelius

Last Edited: Fri. Apr 15, 2011 - 07:48 AM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Why do you allways want change, something "new"?
Thinking that way you will be restarting forever, throwing away experience. Yes, you allways learn something new, but never get to final results.
If a product is marked mature, doesn't that mean that it is finally error free? So if it is a proven good product one should just use (only) that!
"Mature" not exactly means "EOL/not-available-anymore", does it?

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

hi gdhospers

I was speaking about Atmel and the fact that their competitors have moved on to a different, faster business model based on newer technology. It has been very profitable for the competitors as well.

I personally don't need or necessarily want change. I have my comfort zone as do most people. We don't throw away our experience, we take it with us and add to it. Life is not static, it is dynamic, vibrant, and full of change.

gdhospers wrote:
"Mature" not exactly means "EOL/not-available-anymore", does it?

This industry is, and has been, accelerating exponentially and will continue to do so. Yes things will be dropped and not available any more. How many MCUs/MPUs have you seen "bite the dust" over the years. I have seen many during my career.

It has never been comfortable to have to learn a new chip family every few years, but it is necessary in order to remain competitive in our field of expertise.

Larry

"I may make you feel but I can't make you think" - Jethro Tull - Thick As A Brick

"void transmigratus(void) {transmigratus();} // recursio infinitus" - larryvc

"It's much more practical to rely on the processing powers of the real debugger, i.e. the one between the keyboard and chair." - JW wek3

"When you arise in the morning think of what a privilege it is to be alive: to breathe, to think, to enjoy, to love." -  Marcus Aurelius

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

If I were Atmel I'd be putting my development effort into the SAM3 range I think. Cortex M0 and M3 is where the real growth seems to be right now.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

You have to remember that Atmel started makeing AVR's to avoid paying royalty to Intel for makeing 8051's.
AVR32 is a way to avoid ARM royalty, so why should they start making SAM3's

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Quote:

so why should they start making SAM3's

Because the market want to buy ARMs and aren't interested in AVR32s as they are too "niche". I wonder if there's a "large" software development department in the world that doesn't already have some investment in ARM tools?

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Quote:
Does anyone know about the current AVR roadmap

How to compete with ARM.. I think the speed, peripherals or flash size advantage was already lost long time ago. The only thing Atmel can do with AVRs is to compete with the price and availability, although I am not sure they can win anything in a perspective of 10 years. Perhaps in a Tiny market, but on Mega, XMega or AVR32 market they have already been beaten.

Competition makes wonders.

No RSTDISBL, no fun!

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

AVR roadmap = deadend
the AVR mega will probably be relevant as long as designs
1. require 5v
2. require dip pkgs
3. can justify paying more for less performance/functionality vs ARM M series
4. require a simple to code micro like the 51 in the past
5. good for educating those just beginning with micros (just like the 51 and the MC14500 was :wink:)

Xmega and AVR32 roadmap ....WHY?...NXP's extensive line of ARM products available and shipping in the last year alone not to mention the multitude of other ARM vendors has already planted a grave marker on the Xmega and AVR32 products IMHO

Quote:
Or Xmegas beeing fixed?

it should be quite obvious by now that after a few years not even Atmel knows the answer to that although they did win a Product of the Year Award for it when it was still vaporware! :shock:

The AVR32 did just win a Product of the Year Award or something like that somewhere in China recently....it was probably some area of China that does not have internet access ! :roll:

in all honesty though..actually spent the last week doing a AVR Mega product design..eng spec required 5v...but looking around the lab all the other new product designs are ARM's

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Quote:

5. good for educating those just beginning with micros

With things like STM32 Discovery, Mbed, LPCXpresso and so on I think Cortex has that pretty much sewn up too. You can get an ARM and an in-circuit debugger for less than something like an Arduino these days. Atmel could do worse than a "SAM3xpresso" kind of entry level dev kit - preferably with a "free" CROSS-PLATFORM development IDE.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

bluegoo wrote:
NXP's extensive line of ARM products available and shipping in the last year alone not to mention the multitude of other ARM vendors has already planted a grave marker on the Xmega and AVR32 products IMHO
The new AVR32 UC3C 5 volt or 3.3 volt operation has an industrial application niche. How many ARMs have true 5 volt I/O and not just 5 volt tolerant inputs?

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

The new Nuvoton NuMicro ARM Cortex-M0 family offers true 5V operation. They are made in Taiwan and it looks like they will be very cheap. They are intended for industrial applications and Atmel will find it very difficult to compete with them.

Leon Heller G1HSM

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

clawson wrote:
If I were Atmel I'd be putting my development effort into the SAM3 range I think. Cortex M0 and M3 is where the real growth seems to be right now.

Exactly my point:

larryvc wrote:
I think Atmel have been hurt by the inexpensive 32 bit offerings of their competitors. i.e. NXP.

They probably are working on this and if not they should be.

And once the problems with the ARM M4s are fixed Atmel will have to play "catch up" even more.

Even Microchip Technology went with a licensed core for their 32 bit chips.

Atmel, wake up, stop sleeping at the wheel!

By the way, where's the OP? :D

"I may make you feel but I can't make you think" - Jethro Tull - Thick As A Brick

"void transmigratus(void) {transmigratus();} // recursio infinitus" - larryvc

"It's much more practical to rely on the processing powers of the real debugger, i.e. the one between the keyboard and chair." - JW wek3

"When you arise in the morning think of what a privilege it is to be alive: to breathe, to think, to enjoy, to love." -  Marcus Aurelius

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Leon,

Thanks for the NuMicro info. Comparing NUC140 to UC3C number of clock sources, bus matrix and the peripheral mix does not make the NuMicro an automatic slam dunk choice, but the NUC140 looks really really good.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Quote:
With things like STM32 Discovery, Mbed, LPCXpresso and so on I think Cortex has that pretty much sewn up too.

I tend to agree but based on the questions on Freaks the last few years.. the simple reduced functionality peripheal set of the AVR's are probably easier to grasp for many when starting out.
Speaking of LPCxpresso and Avr megas, just received a new NXP LPCxpresso1227 board today (like the other LPCxpresso) to start developing with...needed dual uarts without going whole hog on a lpc17xx series....the MO series is great as the new NXP LPC122x series replaces so many AVR megas we have designed with in the past, plus offering loads of new functionality the AVR megas never dreamed about.
http://ics.nxp.com/products/lpc1000/lpc12xx/

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Mike B wrote:
How many ARMs have true 5 volt I/O and not just 5 volt tolerant inputs?
Fujitsu FM3 MB9BF104 MB9BF105 MB9BF106
Cortex-M3, 2.7v to 5.5v power and I/O.
datasheet
MB9B500 : above plus 3.3v power for USB:
datasheet
Home: http://mcu.emea.fujitsu.com/mcu_product/overview_32FM3.htm
ARM list: http://www.keil.com/dd/search_parm.asp
Edit: Added home, ARM search.

"Dare to be naïve." - Buckminster Fuller

Last Edited: Sat. Apr 16, 2011 - 09:14 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

One of the big selling points for AVR is the low power.
Are there any ARM's out there that have active power lower than 1 mA ?

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

sparrow2 wrote:
Are there any ARM's out there that have active power lower than 1 mA ?
This is close at 191 micro-amps/MHz at 7MHz (an Energy Micro Cortex-M3): EFM32G200
Energy Micro's Cortex-M0 has a planned release this year.
AVR32 UC3L is a bit better w.r.t. power.

"Dare to be naïve." - Buckminster Fuller

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Quote:
One of the big selling points for AVR is the low power.

Nobody really cares about power, but the energy used per instruction or MIPS.
Most efficient are picopower AVRs running slow clock, when they go down to 190uA@1,8V@1MHz. This gives:
190e-6A*1,8V/1e6=0,342e-9J
With increasing speed AVR's efficiency dramatically falls down:
9mA*4,5V/20e6=2e-9J

Compare it to ARM Cortex-M0 (LPC111x from NXP) running 1,5mA@1,8V@12MHz:
1,5e-3A*1,8V/12e-6=0,225e-9J

For AVRs F_CPU is perhaps 0,6-0,7 MIPS. For ARM Cortex M0 this is about 0,9. (not considering the register width advantage). Even smallest ARMs beat AVRs at least twice, not taking register size into consideration (assuming you use only 8-bit variables in the application).

AVR was low power several years ago, but it is not any more.

Quote:
Are there any ARM's out there that have active power lower than 1 mA ?

Are there any AVRs having active power lower than 1mA? It depends on F_CPU.

No RSTDISBL, no fun!

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Microchip XLP parts offer much lower power levels than any AVR device - 20 nA typically in Deep Sleep mode.

Leon Heller G1HSM

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I could imagine hundreds of applications where an AVR core plus some application specific circuit would be a good idea (from my viewpont). Like they did it with Battery AVR-s. There are a lot of applications, where I don't need higher core speed, but a higher resolution AD converter, or a reliable reference voltage etc would be great.
And for example I really want an AVR with CAN bus and 100 pins, or 144 pins.

So I don't think AVR is dead, but if they don't show something new very quickly, then it might be soon. And well the pricing is absolutely mad. For example this:
AT90CAN128-16AU-ND is 11,07$ by DigiKey
LPC1768 is 11,44$

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

An AVR core in an FPGA perhaps?

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

There are some AVR cores on Open Cores. I keep meaning to try one of them on one of my FPGA boards.

Leon Heller G1HSM

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Quote:
There are a lot of applications, where I don't need higher core speed

AVRs are dying not because they are not fast enough.
As you said thare are apps where this is not a major factor. But your argument is the same as "My applications use only 8-bit variables, why to skip to 32-bitters". It is not that you actually need it, but because of the flexibility. This is especially important for hobbysts and low volume production - you have the choice to choose from multiple manufacturers. Who cares the chip costs a buck (more or less) than AVR and runs with F_CPU divided by 8, when you need only 20 chips per year?

No RSTDISBL, no fun!

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

lammelm wrote:
And for example I really want an AVR with CAN bus and 100 pins, or 144 pins.
Taking this literally at face value the UC3C series has two CAN controllers in 68, 100 or 144 pin packages. However, I kind of get the feeling you wanted an 8 bit AVR, not the 32 bit AVR.

gchapman,

Those Cortex-M3 processors look great and are even actually currently available in single chip quantity/pricing. Things have changed for the better since I looked for usable 5 volt ARMs many years ago and could not find any.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

leon_heller wrote:
There are some AVR cores on Open Cores. I keep meaning to try one of them on one of my FPGA boards.

Considering, that it is a free thing, this is wounderfull, and great respect to those who made it. Considering that they are in a half finished, alfa, beta, buggy state, and considering my very little knowledge about VHDL and co, I rather buy a chip.
I'm planning however to dig into this field, but it seems to me, that it requires enormous amount of time. Is there a qood tutorial how to set up xilinx fpga developement environment?

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

It's quite easy - download the free Xilinx Webpack ISE and install it, buy a low-cost starter kit like the Digilent Spartan-3 board ($109) with the USB JTAG adapter, and start playing with it.

Leon Heller G1HSM

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

With ARM7 and Cortex M0/M3 so close to the price of a high-end AVR, for non-product stuff, why not ARM-it?

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Mike B wrote:
Those Cortex-M3 processors look great and are even actually currently available in single chip quantity/pricing.
Comparing Fujitsu Cortex-M3 FM3 with Atmel AVR32 UC3C, the UC3C uses about half the active power and about one-quarter sleep power. But FM3 has a greater clock frequency and is available from distributors. For floating point, the only Cortex-M4 that I'm aware of are from Freescale and ST but these are 3.3v; for floating point and 5v its AVR32 UC3C.

"Dare to be naïve." - Buckminster Fuller