Industrial Automation

Don't large portions of the railroad still operate on relays? Don't most safety critical systems have redundant backups or failsafe modes? Inquiring minds want to know.

Echoing rcapper, safety critical systems are built so failure of the circuitry doesn't make the system unsafe.

"Nobody seems concerned about them..." Not true, the writer just doesn't realise that there are many people who are deeply concerned with such things.

Anyway a "use by date" wouldn't help

When commercialization rate is far ahead than the life of product no one care. I had Computer 386 and practically every year end up changing the computer. All dielectric breakdown because of ion diffusion is calculated for 20 years basis and the life is not more than few years for any electronic device does it matter.

for firming up 20 years we do accelerated life test which end up for 30 years in life to protect ourself.

This is sign of how young electronic market is and still we are in terabyte of memory and grow every quarter.

We have not done yet layer memory.

Next we are looking is spring memory which will be few billion times more than what we have. This development is in fancy and we will be eventually do it as product. Our computer is still slow and runs hot and we have miles to go to make human life little more easier.

We still have a pain of when I send fax in English to my dad in India it does not print in Hindi there. When I talk to customer in Japan and china my voice is not going in Chines or Japanese as I speak English and Hindi.

This is our growth path to make communication open and for that we need newer electronic and more rapid development and commercialization and that means the product will not matured few more decay before it is absolute and we will not have failure

I have not heard of freshness dating for semiconductors except for newer plastic encased components that may require a period of "baking" prior to installation by hot air soldering if they had not been stored in the required dry environment.

The entire assembly containing plastic encased components may require a period of "baking" prior to being put into service if they had not been stored in the required dry environment.

A few semiconductor components have a predictable life expectancy such as a certain electrically alterable read only memory used for storing characterization and test failure data. It was said to be good for greater than 10000 rewrite cycles.

I'm not flying anymore until this gets straightend out. Agood scary movie doesn't scare me anymore but that kind of thing is always first on my mind. Money hungry corporations in this world that we are in, will only continue to overlook these things as long as they meet a reasonable time of dependability. By the time they break down, they figure it can be blamed on anything that can be salvaged from the wreck. I'm sure it has been brought up but, a mass petition would have to be generated or something. I wonder if pace makers are made with parts from the lowest bidding manufacturer that has the minimal capacity microprocessors left over from the first computer ever built, LOL. As long as the chip can meet the needs, who knows how long it has been kept in stock in a warehouse because they over produced them by a million or so. Big brother helps only if you ask with a loud enough voice and they probably don't mind the benefit of population reduction due to a little deafness.

Guest,

In Commercial Aviation redundancy and self test is built in.

The manufacturers and date codes of components are part of the specifications of components and assemblies used.

What is a bother is when a manufacturer of a reliable part changes a process and can't maintain their standard and the fallout from parameter testing becomes so great that you have to use the alternate manufacturer.

Jon

The cost to replace a chip would be more than the entire circuit board. When a failure is isolated to a particular circuit board, you no longer have to go any further to pinpoint the fault. You just replace the board. TV technicians do just that. This is what is done with military electronics. It is designed so a fault can be quickly found. The faulty board will go to a "depot" level repair facility where the fault can be pinpointed and repaired, if economical to do so. I'm sure systems in commercial aircraft are handled the same. If we are talking about toaster ovens, trash it.

As a TV Technician for over 40 years, I only replaced an entire board if absolutely neccesary. If you can replace an IC chip that costs (Wholesale $1.97) Why replace a board for $125.00? I always provided a full year warranty (Labor and Parts) on all repaired boards which was far better than the manufacturers warranty on replacement boards. Besides most replacement boards are were rebuilt, not new.

Ronseto,

"I'm sure systems in commercial aircraft are handled the same"

Yes, there are centers of excellence in the Avionics field where Assemblies and subassemblies are processed and recertified.

Jon

This just looks like sensationalist reporting to me, and a rather moot point in most situations I can think of considering three very important points.

1) The life expectancy of any electrical part is HIGHLY dependent on environmental conditions and usage (heat, humidity, vibration, electrical abuse (eg overloading, overvoltage, etc), read/write cycles, static discharge degrading the life, etc, etc). When a bunch are used together the life of the end product becomes a complex formula of conditional estimates and failure analysis (eg- part and system MTBF (mean time between failure)), and then you have to take in the variables of part assembly on the PCB which can degrade the life of parts dramatically (ESD damage, overheating during soldering, etc, etc).

2) Semiconductors are very reliable and generally only fail if they are abused by operating them outside their electrical limits. The life expectancy of the semiconductor is very likely to be one of the best out of all the electrical components (if not the best). Electrolytic capacitors and even the solder joints connecting the semiconductor to the PCB will likely fail long before the semiconductors themselves do.

and

3) Generally the expected and practical life of the product containing said semiconductors is shorter than either of the two points mentioned above. Additionally for critical systems, failure analysis is performed and redundant backup electronics are employed to counter possible component failures.

There are a number of situations where this is not the case (and I haven't mentioned part repair cost exceeds replacement module or PCB cost), but the majority of products I can think of are covered by the above points. Products like computers, cell phones, cars, satellites, etc fail or are replaced long before any semiconductor age specific issues crop up.

Anyone actually SEEN a semiconductor failure that was due to age and not due to something else (like electrical abuse, ESD damage, solder joint failure). I know I haven't.

Jack,

Anyone actually SEEN a semiconductor failure that was due to age and not due to something else (like electrical abuse, ESD damage, solder joint failure)?

See the end of entry 4 predictable life expectancy of an IC.

Jon

Anyone actually SEEN a semiconductor failure that was due to age and not due to something else (like electrical abuse, ESD damage, solder joint failure). I know I haven't.

See the end of entry 4 predictable life expectancy of an IC.

Excluding specialised parts which by their very design have a known life expectancy based on usage and are already marked with a "lasts for" expected lifetime date (such as memory chips).

I see what you mean.

Those baddies are caught during manufacturing burn in. So in the field we didn't see failures that were not induced.

Jon