Consumer reports shows the repair history for different electronic devices such as digital cameras, laptops, PDAs, TVs, ect...

It seems 5%-10% electronics will fail in 2 or 3 years and will need fixing or replacement.

But for those 90-95% of electronics that does fail in 2 or 3 years, say, a memory card or lcd screen, how long can they expect to last? if we use them properly?

Do they degrade at all? Or do they just suddenly die eventually? I have a projection TV from the 90s that still works fine, looks just like the day I bought it. Will it last forever?

I mean, a car eventually wears out. A shoe or clothing would eventually wear out. But how do an electronics, with very few moving parts, "die"?

 

most electronics die from heat or moisture.

 

and from lack of use

 

I took a break before responding to this, because I wanted to think back on all of the electronic failures I have seen in twenty plus years.

BAD ENVIRONMENT (moisture, vibration, corrosion, etc...) accounts for a lot of what USED to happen, but modern electronics design take that into account in the design phase. They are HARDENED against damage, because anything less would be inviting ruinous losses as things are returned for service. They are built to last as long as possible.

Within limits...

What I have seen is that "INACTIVE" components last almost forever. Resistors, wires, etc....

"ACTIVE" parts, such as transistors (mainly), diodes, tubes, capacitors, and inductors (including transformers) don't last as long.

TUBES tend to have the shortest life expectancy, and transistors can pop at any time.

TRANSFORMERS only tend to blow when something is driven much harder and longer than it was intended to. Pity they don't plan for young adults getting hold of Daddies stereo when they build them...

Electronics die when they have too great a current draw on them for too long. They "FRY" when they are overworked.

Well designed and cared for they can last god only knows how long.

 

The type of electronics failures you see on, say a PC main board, are things like:
1. The green board has thin strips of metal that acts as wires/conductors for the components. If you look close you can see the paths. If the metal strip cracks it will cause a failure.
2. Large components (anything that doesn't look like a chip) can short out/burn out because of high heat, high voltage, etc.
3. The chips suffer the same failures as 1 and 2, except the components are very, very small, burned onto silicon wafers. Conduit paths on the chips break and components fail.

 

They die when you let the smoke out ....




Jeff

52 F3

 

I agree. The magic smoke is all important

It stinks too - so you can tell!!!!

 

I'm not too familiar with transistors. Why are they "active' and wires are not? Aren't they just wires etched into silicon?

But a chip has many transistors, one or two die only matter right?

 

I would guess heat is the main cause of failure, It is usually caused by a high resistance, a weak trace( a trace is the thin metal path on a circuit board) In battery operated electronics the main cause of failure is low batteries, the low batteries creates a higher resistance, which in turn creates a higher amperage, amps of course results in the smoked resister,led,trace,wire from heat.

Excessive amounts of moisture can be bad, and is usually noticed by rust of course either inside, or outside on the screws.

The first thing I usually find when I'm repairing a board, is a burnt resistor, I would say a trace is way up there, or in some instances a bad relay on the board. One other cause for failure is a rather easy fix, the first place the cord enters the electronics in an AC circuit is the transformer a lot of heat here, and actually fail quite often. then from there it is usually (not always) rectified to lower usable DC voltage.

 

"Active" devices require power, and usually have gain, either voltage or current, ie, more out than was put in.

Semiconductors, devices made with silicon, etc, such as diodes, transistors, chips, can die from physical damage, heat, wrong voltage, or (and a big one) static discharge. "Passive" devices, resistors, capacitors, wire, etc, usually from physical damage, or too high voltage, or excess current. Chips today, such as a Pentium processor, can have transistor counts in the millions. And they all have to work!

 

Most large component electronic devices should last 40 or 50 years, under extreme normal use. Things like caps will dry out and fuses go bad, ect, and are kind of like replacing a tire, but even caps have lasted 40 years. (By large component, think of that stereo from the early seventies or earlier, but post tube.)

Newer things have ribben cable and aren't really engineered to last, either electronically or physically. Lots of cheap plastic and not many, if any safeguards built in for current limiting, reverse polarity, etc. Your Walkman or flip phone from the early nineties probably isn't as good as new any more, but the surface mount components are still good. They just can't take much physical abuse or current. (Flex the board, they will crack.

Now we move into the new millinium. A motherboard can have seven+ layers, with trace wires running along and between these layers. The wires can be as small as a micron in diameter. There are millions of transistors on a single chip. They are so small that gamma radiation, (that travels through us and the earth all the time), literally "take out" a few of these millions of transistors with every strike. So they are limited by time.

Engineering is more cost benefit now. For years IC engineering was rock solid with common engineering practices. One of these practices, (rules), was something along the lines of, "You must be an idiot or lazy if a chip you design needs external cooling." Now it's "cool" to have a nice fan or liquid cooled processor chip.

That's kind of why things don't last, while you spend more on them.

 

I was an engineer on one of the first commercial computers, the Univac Model 1. The government (FAA) used it 16 hours a day and we did 8 hours of maintenance, 7 days a week. There were 7 of us rotating to provide repair on a 24 hour basis. We repaired everything to the component level. We were trained on the entire system but each man had a specialty, mine was high speed paper tape and line printers. I worked through all the evolution of computers from tubes to solid state. It was great walking into a computer room with a system that had 2000 diodes and one of them was bad and you had to find it. When PC's began to appear, with great foresight I proclaimed that they were just toys and wouldn't go far. Today's components are getting more reliable, smaller and resistant to shock, age, and atmosphere, and we ain't seen nothing yet.
Dono

 

I've been working in the electronics arena for the past 22yrs and have designed and seen my share of products in both consumer and non-consumer markets. The reliability of devices has been studied extensively and can almost be considered a separate topic for a graduate program.

I'm sure others will chime in, and for what my credibility is worth, there are several factors that determine the 'robustness' of the product, most of what you guys have already mentioned. Proper design around known issues can work around those problems.

Heat is probably the biggest killer. For every 10C device temperature rise, the MTBF falls in half. The other is the design itself. The biggest assumption is the designer hasn't pushed the devices beyond their rated specs in terms of thermal dissipation, device timing, current handling, etc.

Semiconductors don't generally wear out, but these days as geometries shrink to .65nm or less, and things are separated by a few atoms apart, people are discovering new forms of device failures.

Ultimately cost constrains what the designers can do. Consumer-grade products are severely cost-driven, and often compromises in design margins, component selection *must* be made to meet those cost targets. That could mean, for example, allowing a device to run at the high end of the maximum operating temperature spec to save the cost of a heatsink.

I'm quite impressed with TV designs and what is done to cut the cost to the bone. It also doesn't surprise me they are disposable items as a result.

I'm even more impressed with automotive-grade electronics. Think about those COPs in each of your truck's engine. Those are power transistor devices running above water boiling heat most of the time but they also have to work in the dead of winter cold in the other extreme. And I see in my truck the computer is out in the engine compartment instead of traditionally behind an environmentally more friendly glove box. That's a several hundred MHz computer running in the same environment with no reliable cooling. Period. You better believe there's quite a bit of cost and engineering sunk into those designs to make them bullet proof!

But a product is more likely to fail due to a mechanical component or mechanical failure than a semiconductor... connectors, sockets, fans, switch contacts, hard drives, etc are notoriously unreliable

 

In most cases, an "ACTIVE" component can be thought of as one that changes state. Transistors, diodes, and tubes change their rate of conduction like small switches that operate when different signals are applied to them.

I include transformers and capacitors because they have different charge levels or voltage outputs.

In other words they WORK harder, they actually do something (unlike me)....

Resistors and wiring have fixed values.

 

I never saw a transformer or capacitor collide with a truck..........