I am new to microcontrollers, and am in the process of designing a prototype of what I hope to be a commercial device. I have narrowed down the choices of microcontroller, and so far the best choice seems to be the Freescale M68HC08 family, because of low-cost and power, and up to
4K RAM. I should mention that I am at a university.
What I don't know is the life-cycle of microcontrollers. If I design for a particular package (surface mount) and processor, what is the probability that it or some pin and code compatible equivalent will still be available some years later? I know that Freescale was recently spun off of Motorola. Is this reassuring or worrisome?
Yes, probably. ;-) They might concentrate on their higher end stuff. Or not.
Most micros are single sourced and this is a general problem for lower volume products. Using well-written software (possibly in a HLL) can ease the transition if the product has to be redesigned in the future.
Best regards, Spehro Pefhany
--
"it's the network..." "The Journey is the reward"
speff@interlog.com Info for manufacturers: http://www.trexon.com
Embedded software/hardware/analog Info for designers: http://www.speff.com
This is phrased as a rather general question and any rather general answer will probably not be very useful when you try and apply it to specific circumstances, such as the Mot part you mentioned. And eventually, everything goes away (try getting an 8087 math coproc, for example.)
Once you've selected the range of reasonable choices given your requirements and boundary conditions, you might call the few manufacturers (and distributors, as well) and ask for a detailed sense. Such opinions or statements may not be accurate either but they will be more aligned with the part than general comments might be, I think. Also, you can tell them the time period you are considering with more accuracy, I suppose, than you have here.
I don't know that anyone can tell you what the results will be. One guess might be that there will be a narrower focus with more concentration on what is successful and a willingness to shuck away that which isn't, because they may not be able to internally fund a poorly performing (marketwise) product for as long as Mot might choose to do. On the other hand, they may also be more willing to take the longer view in order to build reputation. Or it just might be that you picked the one that will be more successful, longer term.
The only one that isn't always single sourced depending on the fanciness level of the particular controller is the 8051 family. That's why we used them almost exclusively in the past. But the MSP430 seems interesting and since that series is new and expanding quickly it may be around for quite a while. That one will probably remain single sourced though and that is one downside. The other is that it won't drive most FETs too well because it can't run above 3.6V.
Micros tend to have a longer product life cycle than other stuff such as RAM, panacea chips or PC processors. The main reason is that they are used in a lot of long-lived industrial gear. One nice indicator for product longevity is defense use. But the lower end versions might disappear even then.
You don't, and nobody else does. Beyond their current planning horizon, even the actual companies that make those things don't know. Companies go bancrupt, chips drop out of use, production technologies become outdated, factories are refurbished, and whatnot. As Einstein is attributed to have put it: predictions are terribly hard, especially those that concern the future.
If you're doing this for a university project, which is probably of negligible size in terms of number of chips used for it, compared to the yearly production of whatever chip you chose, my best shot at an answer would be: stop worrying about it, and just use what suits your needs, *now*.
The rationale being that any predictions would be seriously unreliable, and if they fail, there's zero chance your small project is going to manage to convince the part maker to keep making these parts for any significant period beyond the date set by their markedroids and controlling department.
Depending on the detailed reasons why you'll still be needing parts in a couple of years, either buy your project's lifetime expected need in stock, right now, or prepare plans to throw future students at the project to re-design it if the chip does indeed vanish. At the very least, assign someone to keep an eye on the part maker and set aside funds to buy lifetime stocked supply as soon as the part threatens to be discontinued.
--
Hans-Bernhard Broeker (broeker@physik.rwth-aachen.de)
Even if all the snow were burnt, ashes would remain.
More than the military, high volume users like the car companies will drive the market and keep that part there.
In the Mot line, and in other lines, there are some parts that are more in demand than others. These also are the low cost ones because of volume. These are less likely to be quickly snipped off.
And, yes, programming in a High Level Language can help. Also keep memory sites as constants, as these will change with a new processor. And modularize things that are processor dependent.
On the contrary, *all* ARM7 and ARM9 parts are single-source. There is no generic jellybean lowest-common-denominator pin-compatible ARM part offered by two or more manufacturers, to my knowledge.
That is a good idea. Even in my days 20 years back it became harder to find a true hardware project. That, BTW, was at the same university where you are now.
aside
discontinued.
What funds? Just kidding but funds for HW parts were usually hard to come by. Unless the project had industry funding which was my favorite way to go.
You have to consider that for a commercial device "some years later" you will probably have to do a cost reduction redesign to stay competitive, also, distributors stock excess microcontrollers long after the manufacturer has discontinued them so thats an option (unless you need 100,000 of them, but 10,000 is common). Manufacturers will also give you advanced notice of products that will be discontinued, at that point you can do a lifetime buy or alternatively buy enough of them so you can have time to redesign your product.
With really, really large qties and mask ROM you may also be able to negotiate the rights to the masks should it be discontinued within the next x years. Then you can have it made somewhere on an appropriate process and go on forever.
"Forever" meaning "until the last fab still using that appropriate process is torn down". Which may last quite a bit less long than "forever" is often expected to imply.
IOW: you had better get not just the masks, but also the construction plans and rights for the fab.
--
Hans-Bernhard Broeker (broeker@physik.rwth-aachen.de)
Even if all the snow were burnt, ashes would remain.
...and commitments from all the suppliers and vendors for the FAB to supply replacement parts, training, and otherwise continue supporting their products....
It's not all that bad. We have done some chips that are now really old, on an early 80's process and still in full production. There are some smaller fabs that are pretty good about longevity of processes, for example AMI. Never had any problems there. Except when you have to go there in winter because it gets to be freaking cold in Idaho :-)
That's what I meant, not just the masks but also the rights to produce for in-house purposes. That kind of deal is often agreed to as long as the chips aren't produced for resale as parts.
Hi Jon, I'm down to my last 10 or so. I also still have a few original
8088's and ceramic 8086's. Also 8080, 8085, 8008 and even a few 4004's. The latter parts are all stashed, as 'mementoes'. The former were used until a few years ago to maintain some old semicon gear.
Al
and
might
--
Please remove capitalised letters to reply
My apologies for the inconvenience
Blame it on the morons that spam the net
If your quantities are small, you can just buy them as you need them and then, if and when they are discontinued, do a bulk buy of enough chips to last you for as many years as you wish.
If your quantities are medium, it will make more economic sense to redesign if and when they are discontinued - or earlier just to use a cheaper part.
If your quantities are large, they will keep making the chip just to keep you supplied.
ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here.
All logos and trade names are the property of their respective owners.