Author Topic: A 32-bit processor for open source projects with longevity  (Read 1906 times)

0 Members and 1 Guest are viewing this topic.

Online SiliconWizard

  • Super Contributor
  • ***
  • Posts: 5443
  • Country: fr
Re: A 32-bit processor for open source projects with longevity
« Reply #25 on: August 25, 2019, 02:03:41 pm »
Again, and as NorthGuy and I hinted, you first have to consider whether targetting "longevity" (from the thread here, we could think of this as a minimum of 10 years I guess...) makes sense at all for open source projects. The whole point of them is that everything is open, shared and everyone in the world can contribute and make it evolve. Why want to release projects that are not meant to evolve (at least on the hardware side)? It really looks like an idea for pure software developers (and even so, why limit software development on a platform that will be completely obsolete in 10 years, even if it still is usable for many things), not for hardware ones. Target maybe a couple years just so that the project doesn't end up in the trash bin shortly. If you're still working on it, and noone has taken over, make it evolve.

IMO, for an open-source project, the mere idea of that just hinders innovation.

Longevity is something you think about for safety-critical / commercial projects, with certifications and all that stuff, or if you're using a very specific component that is hard to replace. Where it really matters. Not typically an MCU which are ubiquitous.

Design your open-source project so that people can make it evolve easily. Document it well, and try not to select too specific parts that would be very hard to replace without a complete redesign.

Just a thought. Do whatever you please. The EEVBlog forum does not endorse my opinion. Do no start a war. Etc. ;D
« Last Edit: August 25, 2019, 02:07:21 pm by SiliconWizard »
 

Offline laugensalm

  • Contributor
  • Posts: 38
  • Country: ch
Re: A 32-bit processor for open source projects with longevity
« Reply #26 on: August 25, 2019, 02:53:28 pm »
Fifteen years ago I'd have said: Blackfin. Now, I'd say: STM32.
Question is: what do you regard as most volatile:
- Companies being bought by others or going down
- Tools with fading support, new OS
- Developers no longer familiar with 'old' technology

There's plenty of tech which will not make any large steps in evolution and which might be around for another 20 years. I trust for example that some ARM Cortex series, TIs msp430 and derivatives will be around for another 20.
If you want the CPU to be opensource as well, you'd have to go do the design in an FPGA architecture that is assumed to be around for a while, plus tools that can be run from at least a virtual machine in the next 15 years.

And this is where it gets interesting: I keep running into migration problems for old systems, where customers have stocked some exotic Z80 alike in DIL for doomsday, but the tools won't work any longer.
This is where you start re-doing stuff from scratch and use a common architecture that is at least supported by GCC. And finally, since go you opensource: The community can do the porting to next generations. Problem solved?
 

Offline legacy

  • Super Contributor
  • ***
  • !
  • Posts: 4415
  • Country: ch
Re: A 32-bit processor for open source projects with longevity
« Reply #27 on: August 25, 2019, 06:56:21 pm »
If you want the CPU to be opensource as well, you'd have to go do the design in an FPGA architecture

So RISC-V?
 

Offline donotdespisethesnake

  • Super Contributor
  • ***
  • Posts: 1115
  • Country: gb
  • Embedded stuff
Re: A 32-bit processor for open source projects with longevity
« Reply #28 on: August 26, 2019, 07:15:07 pm »
Case in point : https://github.com/watterott/WebRadio. A nice little project, but based on Luminary chip which TI acquired, then dropped. There is not an obvious replacement, but either way would take a bit of rework in both hardware and software to migrate to a new chip.
Bob
"All you said is just a bunch of opinions."
 

Offline nctnico

  • Super Contributor
  • ***
  • Posts: 19839
  • Country: nl
    • NCT Developments
Re: A 32-bit processor for open source projects with longevity
« Reply #29 on: August 26, 2019, 07:21:55 pm »
AFAIK the Luminary and now TI ARM chips never got much traction so I wouldn't use these for a project which needs long term support.
There are small lies, big lies and then there is what is on the screen of your oscilloscope.
 

Offline DBecker

  • Frequent Contributor
  • **
  • Posts: 320
  • Country: us
Re: A 32-bit processor for open source projects with longevity
« Reply #30 on: August 27, 2019, 04:46:43 pm »
On a related theme, how do PLC manufacturers support their customers who expect a system to stay in service for decades? Do they stockpile components at the start of a product lifecycle, or are they dependent on recycling components from retired units that get taken out of service during routine upgrades or shutdowns?

Chip makers have the luxury of storing undiced (and often untested) entire wafers from the final production batch.  Whole wafers are easily stored, and newly packaged chips are far less troublesome during assembly than old stock.

Recovered chips, or even carefully stored old stock, are going to require cleaning and drying, test assembly runs and likely per unit hand inspection and testing.
 

Offline dot-bob

  • Newbie
  • Posts: 1
  • Country: us
Re: A 32-bit processor for open source projects with longevity
« Reply #31 on: August 27, 2019, 06:18:18 pm »
Right now unless you ultra low power standby I would be heading for the NXP i.MX RT 106x line of micro.  It is really starting to gain in popularity and cost/performance wise currently blows everything out of the water.  The teensy 4.0 platform is based on this micro.  I have many NXP processors that were originally 10 year longevity and are still being made 25 years later.

-Bob
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf