Tweet
Choosing a "universal" MCU/CPU
G'day everyone,
I've been reading many threads in many forums, and it seems that there are a lot of CPUs being used with different projects. Some are using different CPUs with the same project!
I was wondering if it would be worth collaborating to come up with a CPU/MCU that could be used by anyone.
I'm thinking of polling to find what the most popular and most "user cuddly" IDEs and MCUs are being used right now. Then, since most MCUs differ by only a buck or two, it might be possible to define a "universal" MCU, with a standard codeset (PWM definitions for IB, PI, and so on, sample routines, some maths for analysis) that anyone could then pretty simply "tack together" with whatever their preferred front end and display would be.
Kind of like an Arduino concept, except much simpler - here are 8 PWMs, 8 ADCs, 4 DACs, and 10 or 14 or 20 GPIOs, but with software that will drive PWMx at a definable rate, a sampler that reads ADCx at time xxnS, a simple array for ground samples, and so on and so forth.
It would take a little bit more coding at the start, to ensure the "universality" of the code, but in the short, medium, and long term, it would be a HUGE advantage. People wouldn't need to keep reinventing the wheel, like so many of us are doing - how many people have developed a sampling or averaging routine that is more or less identical to someone else's routine, just on a different architecture?
It seems that if we can get away from "my favourite controller type", and add the benefit of not having to write and debug ADC code and so on, then people could focus on the real fun - getting a fast, accurate front end, doing a great display, adding in logging that others can use, and so on.
It's not an idea about making money, it's an idea about sharing and supporting, and hopefully getting people enough time that they can forget the nitty-gritty register code, and think about the bigger picture.
I know there will be a lot of opposition to it (what if people choose an MCU I've never used before?), but the practical advantages would be enormous (it doesn't matter if I've never used that MCU before, because it's got a good IDE, the "drivers" are already written and debugged, and it's almost ready to run!).
Obviously, the choice would need to be a higher-end device, and in that case, through-hole mounting might not be an option for the MCU. But that's where the designers come in - we can create a standardised MCU board, either with a device already hand-soldered on, (with ICSP, for sure), or at least the ability to obtain a well-made board, with the MCU that everyone is using, and that's two huge steps solved before we even think about programming!
I'm not saying this will be perfect, but I would like people to consider this option, and maybe suggest better ways of doing the same thing.
I hope this interests people. So what do you think?
-PtB
I've been reading many threads in many forums, and it seems that there are a lot of CPUs being used with different projects. Some are using different CPUs with the same project!
I was wondering if it would be worth collaborating to come up with a CPU/MCU that could be used by anyone.
I'm thinking of polling to find what the most popular and most "user cuddly" IDEs and MCUs are being used right now. Then, since most MCUs differ by only a buck or two, it might be possible to define a "universal" MCU, with a standard codeset (PWM definitions for IB, PI, and so on, sample routines, some maths for analysis) that anyone could then pretty simply "tack together" with whatever their preferred front end and display would be.
Kind of like an Arduino concept, except much simpler - here are 8 PWMs, 8 ADCs, 4 DACs, and 10 or 14 or 20 GPIOs, but with software that will drive PWMx at a definable rate, a sampler that reads ADCx at time xxnS, a simple array for ground samples, and so on and so forth.
It would take a little bit more coding at the start, to ensure the "universality" of the code, but in the short, medium, and long term, it would be a HUGE advantage. People wouldn't need to keep reinventing the wheel, like so many of us are doing - how many people have developed a sampling or averaging routine that is more or less identical to someone else's routine, just on a different architecture?
It seems that if we can get away from "my favourite controller type", and add the benefit of not having to write and debug ADC code and so on, then people could focus on the real fun - getting a fast, accurate front end, doing a great display, adding in logging that others can use, and so on.
It's not an idea about making money, it's an idea about sharing and supporting, and hopefully getting people enough time that they can forget the nitty-gritty register code, and think about the bigger picture.
I know there will be a lot of opposition to it (what if people choose an MCU I've never used before?), but the practical advantages would be enormous (it doesn't matter if I've never used that MCU before, because it's got a good IDE, the "drivers" are already written and debugged, and it's almost ready to run!).
Obviously, the choice would need to be a higher-end device, and in that case, through-hole mounting might not be an option for the MCU. But that's where the designers come in - we can create a standardised MCU board, either with a device already hand-soldered on, (with ICSP, for sure), or at least the ability to obtain a well-made board, with the MCU that everyone is using, and that's two huge steps solved before we even think about programming!
I'm not saying this will be perfect, but I would like people to consider this option, and maybe suggest better ways of doing the same thing.
I hope this interests people. So what do you think?
-PtB
when I finally found this, otherwise well hidden, information for Raspberry PI.
Comment