Lifespan of a Module

USMCinfinity

Hey guys, what is the normal lifespan of a microcontroller module, while on the subject what is the normal lifespan of a dip?

John R.

What do you mean by "lifespan" and "module".

When you say "module", I think of something like a STAMP where you have a number of components packaged together, as opposed to something like the SX or Prop that is a single chip.

By lifespan, do you mean the number of years from introduction to end of life, or how long it will last in a circuit? For either of these, the answer is "it depends".

In terms of from introduction to EOL, it depends on how widely the chip is used, and relavant it remains. A chip that is only used by a few items with limited life spans will be short lived. A chip that is widely used may be around for decades.

Life in a curcuit is also highly variable, affected by currents, frequency, duty cycle of the device as a whole (something constantly on, vs. used once or twice a year), etc. The closest you'll come to an answer for a give chip is the MTBF (mean time between failures) numbers on the data sheets.

As for "dips", that will vary with the family and even specific chip in question.

At any rate, across the entire spectrum, there is not going to be a "normal"...


John R.

Mike Green

There's also the issue of what you're doing with the module or other package. Are you inserting it and removing it from a socket or breadboard repeatedly? Is the socket a zero or low insertion force socket? You can easily break the pins on a module or DIP package with repeated insertion and removal from a breadboard or standard socket. They're simply not designed for that type of use.

Time to EOL is also a matter of company policy. Parallax, for example, has kept most of the Stamp modules available for many many years. The BS1 is still available and supported. The BS2, BS2sx, BS2p, BS2pe are still available and supported (the BS2px is pretty new and available and supported as well). Even though the SX processor (the heart of most of the BS2 models) is in EOL status due to legal issues beyond Parallax's control, they have stock on hand for probably 10 more years at current rates of sales. This is an eternity in the microcontroller business.

Leon

For high-reliability systems, designers calculate the Mean Time Before Failure (MTBF) for a complete system, based on the MTBFs of the individual components, operating temperature, etc. using statistical techniques.

localroger

I have stuff from the mid-1980's that is still in service because even though it is hopefully obsolete it's easier, cheaper, and less risky to fix than it would be to engineer a new and more modern solution, and they are working. I am currently in the middle of a USD$500K job to upgrade some ca. 1992 controllers whose parts haven't been made since 1998.

It is only the threat of extended downtime because parts are totally unavailable after the next failure that has prodded this customer into doing the very expensive engineering to update these systems. The old systems work and the operators and maintenance guys are familiar with them; this change is an unnecessary nuisance. But nobody is making 4 MHz 8085-based CPU cards any more (that were $2,200 each when new) and there isn't enough volume to justify engineering an exact drop-in replacement, which would require emulating a closed source program running on hte 8085.

USMCinfinity

Yeah I meant, like how much will a BS2 module "live", also since I know that is a module, a combination of many things I asked how much do I expect a dip (such as a SX Chip) to live?

Leon

As I said previously, you need the MTBFs. Parallax might have that information, but I doubt it, as they aren't high-rel parts.