Nice to see new AVRs with more features. Any insider info on plans for a mega1280PB that runs at 20MHz? Better, Cheaper, Faster?
Microchip/Atmel VPs reading this? How bout a 20MHz 1280/2560?
Author
Message
Who says it cannot run at 20MHz? THere are lots of threads about overclocking AVR's. But if you need it as a written spec in the datasheet, then I see your point.
OTOH, if you take a look at our new owners products, many of them hit wicked fast clocks. More than able to compensate for the 4 clocks/instruction .
JIm
bobgardner wrote:
Better, Cheaper, Faster?
Last Edited: Fri. Oct 14, 2016 - 06:16 PM
I've got an olimexino with a crystal in a socket, and the 20mhz 328 really books at 22.1144
bobgardner wrote:
Nice to see new AVRs with more features. Any insider info on plans for a mega1280PB that runs at 20MHz? Better, Cheaper, Faster?
Which ones do you mean ?
'runs at 20MHz' can mean many things...
There are some issues with crystal oscillators not matching the core speed, and I see on the newest Tiny817's banner 20MHz, but > 4.5V & they drop HF Crystals entirely, and support only 32.768kHz crystals, and external clocks.
Given Microchip supply external Clocks, their reply is likely to be "use one of our external clocks"
bobgardner wrote:
Nice to see new AVRs with more features.
Have I missed something in one of Atmel's newsletters? I've not noticed anything new for ages.
bobgardner wrote:
Any insider info on plans for a mega1280PB that runs at 20MHz? Better, Cheaper, Faster?
The jump from 16MHz to 20MHz just isn't worth bothering with. 16MHz to 32MHz might be worth doing.
Level: 10k+ Postman
Joined: Tue. Jan 9, 2007
Posts: 11862 View posts
Location: Arlington, Texas, u.S.A.
Brian Fairchild wrote:
I've not noticed anything new for ages.
- tiny104 (both temperatures) are in-stock
- mega324PB is
imminentETA 2016-Dec-05 - tiny817 (and similar) are "heads up" (IAR EWAVR has it)
- AVR284 AES-128 crypto on megaAVR
- AVR42788 SD bootloader for XMEGA A1U
- AVR42787: AVR Software User Guide
via "last 60 days"
http://www.mouser.com/search/ProductDetail.aspx?R=0virtualkey0virtualkeyATMEGA324PB-AU
Edit : strikethru
Last Edited: Wed. Oct 19, 2016 - 04:13 PM
gchapman wrote:
Brian Fairchild wrote:I've not noticed anything new for ages.
- tiny104 (both temperatures) are in-stock
- tiny817 (and similar) are "heads up" (IAR EWAVR has it)
The Tiny104 is surprisingly expensive for such a limited device, and no prices are showing yet on the Attiny817 series, but part codes are up for
ATTINY817-XPRO & ATTINY817-XMINI
gchapman wrote:
Brian Fairchild wrote:I've not noticed anything new for ages.
- mega324PB is imminent
Thanks. Hmmm, hardly going to set the world alight with that lot. Most of them are clearly aimed at a very specific market segment, maybe even a specific customer. The 324PB, with 3 USARTs, is probably the most interesting.
Level: 10k+ Postman
Joined: Tue. Jan 9, 2007
Posts: 11862 View posts
Location: Arlington, Texas, u.S.A.
Who-me wrote:
Approximate to a recent PIC (apples and oranges The Tiny104 is surprisingly expensive for such a limited device, ...
http://www.atmel.com/devices/ATTINY104.aspx
http://www.microchip.com/wwwproducts/en/PIC16F506
https://octopart.com/search?q=ATtiny104&avg_avail=(1__*)&start=0
https://octopart.com/search?q=PIC16F506&avg_avail=(1__*)&start=0
Level: 10k+ Postman
Joined: Tue. Jan 9, 2007
Posts: 11862 View posts
Location: Arlington, Texas, u.S.A.
Brian Fairchild wrote:
The jump from 16MHz to 20MHz just isn't worth bothering with. 16MHz to 32MHz might be worth doing.
Nah,
Why bother if you can get a 72MHz Cortex M3 (or pic32 or xmega or whatever) with loads of peripherals & ram for the same price?
I still like the 8-bitters, and they will keep on jogging on for years to come, but if performance is so critical then the person who designed that 8-bitter into your circut should probably be laughed at.
Brian Fairchild wrote:
bobgardner wrote:Nice to see new AVRs with more features.
Have I missed something in one of Atmel's newsletters? I've not noticed anything new for ages.
bobgardner wrote:Any insider info on plans for a mega1280PB that runs at 20MHz? Better, Cheaper, Faster?
The jump from 16MHz to 20MHz just isn't worth bothering with. 16MHz to 32MHz might be worth doing.
Would you really consider a +25% increase to be not worth bothering with? If my main gaming rig went from 5GHz to 6.25GHz, that would be amazing!
Also, that's 4 million additional cycles per second. Significant, I'd say! Hehe... to play the devil's advocate.
Jamison wrote:
Would you really consider a +25% increase to be not worth bothering with?
I would. If your design is that close to the edge then something is badly wrong.
If your gaming system runs out of steam it doesn't matter. If your car's brake controller runs out of steam people die.
Jamison wrote:
Would you really consider a +25% increase to be not worth bothering with? If my main gaming rig went from 5GHz to 6.25GHz, that would be amazing!
Also, that's 4 million additional cycles per second. Significant, I'd say! Hehe... to play the devil's advocate.
Yes, any speed increase is worth having, especially given other vendors are trending up, not down, in MHz.
However, Atmel/Microchip do have an embarrassing fumble around their Crystal Oscillator design, which I'd say makes the requested specific new 20MHz parts, less likely.
Brian Fairchild wrote:
I would. If your design is that close to the edge then something is badly wrong.
You've got a point there but I guess I was considering my thinking to happen in the design phase. Certainly if you've designed for 16MHz and you suddenly decide "Damn, I need 25% more!" then yes, you have an inherent design problem. I'm used to building video games and high end simulators that require heavy amounts of horsepower from the CPU (and well crafted optimization) so I guess that's where my thinking took me.
Brian Fairchild wrote:
If your gaming system runs out of steam it doesn't matter. If your car's brake controller runs out of steam people die.
But what if I need to play the latest Farming Simulator 2016??! Okay, just kidding. I'll shut up now.
Who-me wrote:
Yes, any speed increase is worth having, especially given other vendors are trending up, not down, in MHz.
To go along with some points of Brain's, a good thing to remember is speed increase is related to current draw as well so in battery operated devices, maybe not worth it. As Brain said, a design problem if added after the design phase.
All good points.
However, Atmel/Microchip do have an embarrassing fumble around their Crystal Oscillator design---why after all these years is there now a problem? Did they change something or the parametrics of the chip? If so, go back to the old way.
There is a thread here (this year or last year) that they (Atmel) removed the full swing option in the PB chips to cut costs.