| Author |
Message |
|
|
 Posted: Feb 26, 2012 - 07:36 AM |
|
Joined: Dec 31, 2005
Posts: 688
|
|
Atmel should design an AVR (tiny/mega) line with the possibility to change every feature to any i/o pin (no more hassle with pcb traces.) And build-in even more features (standard a faster 10-bit ADC and 1% DAC) and also an 8- 10- and 16-bit timer block (with more prescaler resolution) with each a seperate PWM. Perhaps there is also a possibility to build-in a digital potentiometer/temperature block?
The more AVR is becoming like 'candy' the more we 'nibble'. |
_________________
RES - http://www.attiny.com
|
| |
|
|
|
|
|
|
Posted: Feb 26, 2012 - 12:28 PM |
|
Joined: Jul 18, 2005
Posts: 62321
Location: (using avr-gcc in) Finchingfield, Essex, England
|
|
|
Quote:
with the possibility to change every feature to any i/o pin (no more hassle with pcb traces.)
To combat those who don't do PCB/system design/layout correctly in the first place? |
_________________
|
| |
|
|
|
|
|
|
Posted: Feb 26, 2012 - 12:46 PM |
|
Joined: Jan 23, 2004
Posts: 9828
Location: Trondheim, Norway
|
|
|
Quote:
Atmel should design an AVR (tiny/mega) line with the possibility to change every feature to any i/o pin (no more hassle with pcb traces.)
The newer XMEGA-B devices have a (very) limited remapping ability on some functions/pins, but nothing as close as the UC3's pin MUX or the flexibility of a FPGA. Making freely re-routable pins in a chip is expensive and hard.
- Dean  |
_________________
Atmel Studio 6.1 is now released, grab it here.
Report AS6/ASF bugs here.
|
| |
|
|
|
|
|
|
Posted: Feb 26, 2012 - 12:53 PM |
|
Joined: Jan 11, 2009
Posts: 208
Location: Stockholm Sweden
|
|
|
Quote:
with the possibility to change every feature to any i/o pin (no more hassle with pcb traces.)
FYI called remapping and offered by some suppliers NOW!
Quote:
To combat those who don't do PCB/system design/layout correctly in the first place?
NO! To make better use of tiny series where every pin counts.
Quote:
And build-in even more features (standard a faster 10-bit ADC and 1% DAC)
That is just what you need a faster clock for.
John |
_________________
If all else fails, read the instructions.
|
| |
|
|
|
|
|
|
Posted: Feb 26, 2012 - 03:42 PM |
|
Joined: Jan 11, 2009
Posts: 208
Location: Stockholm Sweden
|
|
|
Quote:
How does re-ordering the pins have any bearing on that? Surely it's down to your PCB routing software as to how it copes with a fixed layout?
Because a lot of "features" use the same pin. You have to decide which one. You can not at the moment switch "features" to available pins.
It has absolutly nothing to do with your PCB routing software.
Look at it like this. ADC & UART are on one pin SPI & DAC on another. I only want to use ADC & UART.
The argument put by RES is to make a better layout. Which also I can see.
I can not agree with Dean. Chips with remapping are no more expensive.
John |
_________________
If all else fails, read the instructions.
|
| |
|
|
|
|
|
|
Posted: Feb 26, 2012 - 04:01 PM |
|
Joined: Dec 11, 2007
Posts: 6849
Location: Cleveland, OH
|
|
Adding an I/O Mux takes a lot of transistors. Using a pricing scheme based upon transistor count, or silicon surface area, this then increases the production costs / chip.
It also increase the chip's complexity, setup, data sheets, etc. (Increased capability comes at the cost of increased cost and setup complexity.)
I would think that the Multi-pin I/O Mux would take MANY more transistors if one were trying to impliment it without a 1 - 2 clock penalty for routing and synchronization, and without switching to a core that ran on a x2 or x4 internal clock bus architecture.
But what do I know. I'm better with Carbon based structures than I am with those based on Silicon.
JC |
|
|
| |
|
|
|
|
|
|
Posted: Feb 26, 2012 - 04:53 PM |
|
Joined: Jan 11, 2009
Posts: 208
Location: Stockholm Sweden
|
|
Its easy enough to compare prices on suppliers sites.
There are two "features" that I would willingly pay more for.
That is speed and remapping.
At the moment price seams to be determined by amount of memory.
John |
_________________
If all else fails, read the instructions.
|
| |
|
|
|
|
|
|
Posted: Feb 26, 2012 - 05:24 PM |
|
Joined: Jul 18, 2005
Posts: 62321
Location: (using avr-gcc in) Finchingfield, Essex, England
|
|
| Oh then I think you expect more from pin mixing than I - if pin 7 has UART + SPI I might expect to move that to pin 13 but not split the functionality. |
_________________
|
| |
|
|
|
|
|
|
Posted: Feb 26, 2012 - 06:46 PM |
|
Joined: Jan 11, 2009
Posts: 208
Location: Stockholm Sweden
|
|
|
Quote:
if pin 7 has UART + SPI I might expect to move that to pin 13 but not split the functionality.
NO you can split the functionality. Remapping enables you to put what you feel like on any pin you feel like. If you want the UART on pin 7 put it on pin 7 and if you want SPI on pin 13 put it on pin 13.
As Dean pointed out above support is limited at the moment. It is only some pins and some features but is something that is coming in as new chips come to the market.
As DocJC pointed out the complexity goes up the size of the data sheet goes up and I spend a lot more time banging my head on the keyboard than I ever did with ATMEGA8. (It is not Atmel chip)
John |
_________________
If all else fails, read the instructions.
|
| |
|
|
|
|
|
FAQ
Search
Register
Log in to check your private messages
Log in
Maximize
RSS
