How is the Aging of Electronic Equipment determined?

[Muhammad Usman]

I am working in telecom company and they told me that they installed new system and then they told me the equipment will be EOS (End of Service) after 5 years. I was wondering that how the electronics equipment's age is calculated. If I strictly categorize my questions and explain the confusion so it will be

1) How the transistors/Diodes age is determined and what factors are counted on the single transistor/Diode age ?
2) How we can calculate the aging of the other basic electronics equipment such as inductor, capacitor, resistor to publish it ?
3) How the reliability of these components is calculated ?
4) The final age is the least age of components in the system.

I know folks this topic is Broad topic therefore the pin point answer may not be possible so if anyone can give me some heads up to start or guide me to some book/links or any other content or search engine key word where I can study and have this knowledge and information.

Many Thanks

 

[anorlunda]

Calculated is not the best word.

I think you are asking about this
https://en.wikipedia.org/wiki/Bathtub_curve

 

[Svein]

There is a handbook for calculating the estimated lifetime of an electronic device: MIL-HDBK-217F (http://www.reliabilityeducation.com/intro_mil217.html). It may very well be somewhat out of date, but it is the (unofficial) standard for such calculations.

 

[berkeman]

I was wondering that how the electronics equipment's age is calculated.

https://en.wikipedia.org/wiki/Mean_time_between_failures

 

[Klystron]

https://en.wikipedia.org/wiki/Mean_time_between_failures

And to dig down to component level "duty cycle" concept could be relevant: https://en.wikipedia.org/wiki/Duty_cycle

Edit: the formulas, quite useful for maintaining pulsed systems such as radar, have been applied to many other devices.

 

[jrmichler]

In addition to to reliability and MTBF, electronic equipment is rendered obsolete as better equipment is developed. They may have determined that it is best to just replace everything after five years.

 

[phyzguy]

I think jrmichler is on the right track. I doubt if the equipment has worn out after 5 years. It will probably just be obsolete and will need to be replaced by the better equipment that will become available, that will have high throughput, lower cost of operation, ... When I replace my cell phone or laptop, it is rarely because it has stopped working; it is usually because a better model with more features is now available.

 

[f95toli]

Agreed, there is a huge difference between financial and technical lifetime. The former is usually set by rules related to taxation, asset management etc and will usually be the same for all equipment irrespective of how quickly it actually ages.

Good quality "expensive" electronic equipment will usually last quite a while as long as it is maintained (clean out dust, replace any worn out cooling fansetc ) and is not not abused, We use a fair amount of expensive microwave kit and since we can't really replace them every few years (since they cost too much) we often end up using it 10-20 years. This is kit that is often on 24/7.
We have some generators from 1987, they are very heavy and noisy, but they still still sort-of works if you feed them an external 10 MHz ref signal).

 

[CWatters]

Some types of capacitor have a life expectancy that is very short just 10,000 hours at Max rated temperature. That's about 18 months. However it doubles for every 10C reduction in temperature so getting to 5 years is possible if temperatures are controlled. Getting to 25 years as desired for a solar panel inverter is more of a problem.

 

[Rive]

4) The final age is the least age of components in the system.

You better take every soldered connection as a component, then.

By my experience the expected lifetime of an electronic device depends only on surprisingly few things. (The initial part of the bathtub curve has no part in this.)
- Manufacturing. Every soldering point ages: the actual aging process will depend on the specifics of the manufacturer and the temperature 'wear' of the equipment. Sometimes mechanical 'wear' will count too.
- Capacitors. Most components aging well, but (electrolytic) capacitors are always a problem. This part was already mentioned before.
- Underdesigned 'key' components. This can be calculated as a simple design error, but sometimes even I have doubts. Resistors or semiconductors on the cheapest PCB, with expected dissipation or voltage at the exact design limit and so on...

Textbooks might say otherwise, but (again: by my experience) the 'age' (as matter of expected lifetime) will depend only a few key components (and that includes the manufacturing process of the assembly too).

 

[trurle]

I am working in telecom company and they told me that they installed new system and then they told me the equipment will be EOS (End of Service) after 5 years. I was wondering that how the electronics equipment's age is calculated. If I strictly categorize my questions and explain the confusion so it will be

1) How the transistors/Diodes age is determined and what factors are counted on the single transistor/Diode age ?
2) How we can calculate the aging of the other basic electronics equipment such as inductor, capacitor, resistor to publish it ?
3) How the reliability of these components is calculated ?
4) The final age is the least age of components in the system.

I know folks this topic is Broad topic therefore the pin point answer may not be possible so if anyone can give me some heads up to start or guide me to some book/links or any other content or search engine key word where I can study and have this knowledge and information.

Many Thanks

Advice from foundry engineer:
Electronic device degradation is 2-way.

1) TDDB (Time-dependent dielectric breakdown). If the transistor is made perfect, it can work for centuries. On the other hand, if you have single errant atom (so called charge trap) buried in the gate oxide, MOSFET degradation cascade is started if electric field is high enough.
Historically, transistors were qualified to tolerate a up to ~10 defects per 1um³ of gate insulator. Recently that policy changed. Improved production resulted defects density, and transistors are so small what majority do not have gate dielectric defects at all. Therefore, chip makers recently stress their chips to ~70% of defect-free breakdown fields (to improve speed and reduce costs mostly). In this case, initial atomic-scale defects located in the middle of gate dielectric layer will actually pass enough gate current to destroy the transistor in few years. The process is stochastic though.
2) HCI/NBTI (Hot-carrier injection and Negative-bias-temperature-instability) - this degradation mechanisms are relying on energetic electron flow through transistor channel actually doing the damage on transistor. Roughly the effect is slow degradation of transistor gain and output power, as the interface between semiconductor and dielectric is gradually altered. Power amplifiers in the radio equipment or line drivers are most susceptible to this form of degradation, and telecom EOS is actually mostly determined as the point then antenna/line power at equipment output drops below acceptable level. Most likely it is the case of 5 years lifetime specified for you.

Resistors/inductors are recently quite reliable to be not taken into account. Capacitors are susceptible to same TDDB degradation as transistors, but not to the HCI/NBTI degradation.

Duty cycle affects HCI/NBTI, but not TDDB (unless you power down block instead of idling)
Temperature rise increase TDDB degradation rate very rapidly (2-3 times per 10C of temperature rise), but have less effect (or sometimes in case of SOI CMOS wafers even negative effect) on HCI/NBTI degradation levels. Actually some of HCI/NBTI damage can be healed by "annealing" at high temperature (~150-200C), although it is rarely used because of likely collateral damage to more heat-sensitive components.

 

[skeptic2]

It's been years since I've done reliability testing but as I recall the algorithms look at the voltage, current, and power maximums in a device compared to the limits of the device. It uses exponential curves so the life of the device deteriorates much more rapidly as the limits are approached. For longevity it's best to keep the parameters well below the limits for the device.