Understanding electronic component failure.

[Geek-9pm]

One of the frustrating things for a electronic hobbyists is understanding when and why some component in an electronic device will fail. You seldom get warnings. Failing components do not always give smoke or make noises. Some tiny component may fail without any warning at all. What is more frustrating this might happen, even after the device has proven to be reliable for years. Engineers who work in maintenance and design are familiar with the concept called the so-called" bathtub curve". Here is a reference from Wikipedia that introduces this idea.

...
Electronic components have a wide range of failure modes. These can be classified in various ways, such as by time or cause. Failures can be caused by excess temperature, excess current or voltage, ionizing radiation, mechanical shock, stress or impact, and many other causes. In semiconductor devices, problems in the device package may cause failures due to contamination, mechanical stress of the device, or open or short circuits.
Failures most commonly occur at near the beginning and near the ending of the lifetime of the parts, resulting in the bathtub curve graph of failure rates. Burn-in procedures are used to detect early failures. In semiconductor devices, parasitic structures, irrelevant for normal operation, become important in the context of failures; they can be both a source and protection against failure.

http://en.wikipedia.org/wiki/Failure_modes_of_electronics

From the above, there are two general conclusions we can make. The first thing is suspect that a newer device may fail. Even when new devices may fail while older devices may still be working in the same hard environment. To add to the confusion, after many years the older device will eventually fail.

[patio]

There is no logic...or reason to quantify the age of a component is in fact a contributing factor in it's failure rate...

[Maleke]

Interesting discussion.

In general an IC has no moving parts, all the transistors and components internal are operated by voltages, so there shouldn't be any sort of repetitive stress failures.  Since it is packaged then external forces shouldn't effect it either.  I'm pretty sure any internal corrosion from packaging defects would manifest early in life, as would other errors, which explains the first part of the bathtub graph.

Perhaps slight corrosion from the external pins creeping its way inside the chip to the connections?  I bet if you had a nice lab you could cut the chip out of the old package, repackage it and be good to go.

-Mal

[DaveLembke]

A 1500+ page book could be written on this subject.

I have serviced electronics for about 30 years now. There are pros and cons to buying new when what you have that is older is still working. Too many to list of each, but generally people buy new when the prior device failed, or slows, or lacks features of a newer product. Others who are materialistic go through phases of always having to have new as a wealth status among their peers.

A chapter a piece could be set to each failure type with factual information from the last 70 years when electronics took off faster than ever after World War 2, with the lower cost Transistor in 1947 to that of what use to be Tubes prior, and set the start for IC's in the late 1960s early 1970s. Each chapter could state known problems with specific components and how to determine if they have this problem or not based on measurements or known bad batches or manufacturers parts that didnt last the test of time.

There was actually a time long ago that electronic devices which include light bulbs were lasting too long. People were buying them and then when they didnt fail the company that made the lightbulbs that last too long were soon saw a decline in sales not because no body wanted to buy their high quality products that didnt fail but because the product was not prone to failing and was lasting too long people were not buying what they didnt need since what they had was continuing to work.

Then some guy came up with this vision for the future called Planned Obsolescence, which kept customers happy only because they didnt know this was being implemented. Slowly products were not lasting as long and people were having to buy newer and replacing the old items because they are now being designed to have a specific life expectancy. So Light Bulbs were being designed to only last so many hours and if a batch lasted too long the company that made them would be given a fine for not keeping their products in check with other manufacturers light bulbs that were to only last a set amount of hours. There is an old antique light bulb in a fire house that has been lit non stop with exception to the occasional power outage for the last 100 years. The town this light bulb is located in actually had a 100 year birthday cake for a Light Bulb. People looking into this lightbulb as to why it hasnt failed found out that it was designed by a manufacturer that went out of business when their business failed due to its success in the bulbs lasting too long. There is a great video on youtube on this Planned Obsolescence that my one online friend in Canada shared with me.

In this day and age, we could surely design products to last many many years, but the fact of the matter is that if they did that, then the economy would crumble because sales would slow, people would lose their jobs in manufacturing, etc.

I am against planned obsolescence myself. Its extremely wasteful and the pollution that is created by it is the biggest problem. But not really a problem for the consumers because the consumers continue to consume and ship off the waste to 3rd world countries or countries with very weak laws against polution where heaps of our waste go and pollute the planet.

I generally keep computers long past the time that most people give up and buy newer. The only factor that causes me to buy newer is the fact that I am addicted to online gaming with friends and as games evolve and become more demanding on the hardware to process them, I have to keep up with the hardware required to play them. Otherwise if I wasnt a gamer, i'd probably still be running a Pentium II 450Mhz today with a Linux Distro to stretch the life of the computer.

So much to say, but dont want to ramble... 

[Geek-9pm]

Wow! This brought mope response than I imagined. 
Myself, I have about 60 years in electronic troubleshooting and design. But that does not make me an super expert.  Most of  what I know comes from reading what others have said.

For more information about semiconductor behavior, here are some general sources of information. The links are for current publications.

Motorola   https://archive.org/details/MotorolaLinearInterfaceIcDataBook1990

Texas Instruments    http://www.ti.com/corp/docs/company/history/technologyforinnovators/books.shtml

Intel  https://noggin.intel.com/technical-books

National Semiconductor   https://openlibrary.org/publishers/National_Semiconductor

RCA semiconductors http://www.chipdocs.com/manufacturers/RCA.html

Sylvania  http://semiconductormuseum.com/MuseumLibrary/HistoryOfCrystalDiodesVolume1.pdf

International Rectifier books https://archive.org/details/InternationalRectifierIgbtDesignersManual

These companies have published a lot of information over the years. They mention some of the reasons why changes were made in  IC designs.

Let me know if you want more information as to why even good package designs will eventually fail. I can not find the exact reference, but I remember some of it. I think I read it twenty years ago.






 
 

[DaveLembke]

Here is a good watch on planned obsolescence and 100 year old light bulb that still burns: http://www.youtube.com/watch?v=vfbbF3oxf-E