Longevity

[victorf]

Back in the early '90s I designed, build and successfully marketed a product with an embedded micro. I discontinued the product around 1997 because some of the ICs used in the product became obsolete. However, I recently found out that the micro was still being sold in its original design.

I have recently had a series of inquiries that lead me to believe that there might still be a market for my original idea. I asked myself how I would go about redesigning the original product with modern components. Since most of the effort will go into the software, it is important that the micro I select have some longevity. There are quite a few choices today for a viable platform. However, I am not knowledgeable enough to forecast the long-range availability of any of my choices. History shows that I happened to make the right choice 20 years ago, but I do not delude myself that it was a wise choice, but merely a lucky choice.

If I am still lucky, I would like to pick a platform that will remain viable for no more than 10 years. I am asking those of you who a certainly more knowledgeable of the subject to suggest what platforms might still be around in 2020! I do NOT wish to be in the microcomputer/controller design/development business. I just wish to be able to have a complete design that I can buy off-the-shelf and plug into my product today, tomorrow and throughout my timeline.

Vic

[spamiam]

I would like to pick a platform that will remain viable for no more than 10 years Vic

I am certainly no expert on this subject, but here are my thoughts:

Do you really mean NO MORE THAN (I.E. less than) ten years? If so, lots of stuff ought to fit the bill.

What do you need in the way of I/O pins, memory, processing power?

The microcontrollers I have seen seem to turn over pretty quickly these days. However, a newly released uC ought to be a viable platform for at least 5 years, and possibly up to 10. If you layed-in a stock of them anticipating future need, you could have a little stockpile that would still be useful for quite a while. Be sure to save programmers and a computer that will support the programmer too!

-Tony

[victorf]

Tony,

Thanks for the reply.

Do you really mean NO MORE THAN (I.E. less than) ten years? If so, lots of stuff ought to fit the bill.

I am currently 72 years old and doubt I will still be spry enough to want to continue past ten years.

What do you need in the way of I/O pins, memory, processing power?

If the processor had, at least, the features of e ZX type micro, I would be happy. I believe there is
enough program memory and available ports to do what I need to do to upgrade the original design.

I like your idea of preserving a computer and programming tools just in case! I am having problems updating the original devices because of the shift from serial to USB. I have been somewhat successful with using an FTDI FT232R USB/Serial converter for this work. I have NO computers with a serial port.

Vic

[dlh]

I am currently 72 years old and doubt I will still be spry enough to want to continue past ten years.

What's a decrepit old geezer doing messing around with this stuff? You should leave it to us youngsters.

BTW, I'm 3 years younger than you.

And, I keep an old PC with both PCI & ISA slots because I have a really good ISA oscilloscope card.

• http://www.gigatest.net/hungchang/protek-220.htm

[cerickson]

One thought that occurred to me (yes, it hurt) is that if you incorporate a 24 pin socket into your design, you might be able to exploit the Parallax STAMP electrical and physical package defacto-standard for many years to come. There are a LOT of different microcontroller products from many makers that conform to that standard and I don't think that will change for a very long time.

I know this does not address the problem of support chip obsolescence but maybe you can exploit the on-board hardware features of the ATmel chips and eliminate as many active external components as possible.

The unit cost will be a bit higher but I suspect that will only be important if you are doing nominally large manufacturing.

-Christopher Erickson

[spamiam]

I am currently 72 years old and doubt I will still be spry enough to want to continue past ten years.

Hey, 72 is the new 62! (or is it the other way around?) anyway, don't sell yourself short! You might be spry well into your 90's, and keeping mentally active has been shown to be a way to ward off mental deterioration.

If the processor had, at least, the features of e ZX type micro, I would be happy. I believe there is
enough program memory and available ports to do what I need to do to upgrade the original design.

Well, then you are wide open. What programming language is your preference? I love the ZX platform, but when I frequently just go straight to the plain AVR and program in C. So, you COULD just stay with the ZX platform. Being more expensive, than a plain AVR, it is harder to stockpile. However, Don does try to maintain backward compatibility as much as possible. Plus, the newer generations of AX devices ought to keep up with the needs for interfacing to a PC, so it will be less important to keep antique programmers and I/O cards on hand and functional.

If you are going to go to a PLAIN AVR and program in something like C, then the ATmega 88/168/328 family is nice in a 28 pin DIP package. They are relatively modern and are likely to remain in production for a while. 32K is a pretty big program in C. And, they are cheap enough to stockpile.

The ATmega164/324/644/1284 are nice 40 pin DIP devices. They also are reasonably modern and should be in production for a while. 128K program space ought to be more than enough for any programthat you are likely to write. Again, they are cheap enough to stockpile.

have NO computers with a serial port.

Well, you can easily and cheaply get a serial port interface card for you PC. Notebook serial cards are surprisingly plentiful too! Check Newegg at:
http://alturl.com/sqog5


-Tony