Electrical Equipment Degradation

Dear all

I have experienced working in a laboratory to repair telecommunication devices. there, I learned that some characteristics of electronic devices like op-amps degrade over years of operation. for example, we take the input current of a op-amp to be zero (in scale of pA or nA). But over years of operation, the input current increases to order of mA; so the operation of the instrument is no more like the first time.
I want to expand my understandings toward this concept. Would any of you do me a favor and introduce me some readings about this concept?

Thank you in advance,
regards,

Unless the instrument (or the part concerned) is operated close to the limit (or somewhat beyond) there is no reason to suspect that electronic components will suffer serious degradations. There are several exceptions, e.g., electrolytic capacitors that tend to degrade much more rapidly. Most components have a useful life and the component may fail within its life time is Poisson distribution.

However, if an op-amp is used within the specifications, there is no reason why the input current will increase 1000 (or more) fold during its very long life time.

Hi,

I agree with c_mitra.

Especially if an OPAMP input current increases, then most likely it is caused by ESD.

ESD pulses hurt the isolation barrier within the semiconductor structure.
This may cause immediate failure or failure after years.

Is the OPAMP input invovled in I/O lines of your device, or are there any other human touchable devices (sensors, keys, push button, LCD, antenna...) ? And if so, is ESD protection installed?

Klaus

Klaus raises a good point.
While discretes such as electrolitics as mentioned by c_mitra will have a time/temperature graph
associated with them - IC's don't really need that. (The issue with electrolitics is simple drying out
of wet electrolite over time)

As Klaus says ESD does damage chips over time and the damage when seen under a microscope
looks very much like rust eating away at a car body. Once started it tends to continue also.

I seem to remember there was a fictional story some time ago about the president of the USA
demanding a chip manufacture explain why their processors didn't have a suitable lifespan...
I can't recall if it was in a movie or a book.

If the equipment is used in a high humidity/high temperature saline environment (coastal areas, for instance) and the board is no protected by a conformal coating or full potting, they will corrode.

I've seen corrosion enter IC packages (plastic packages are not hermetic) and cause some effects like the one you mention.

I was once employed as a service engineer doing board level repairs on very old or obscure equipment where there were no schematics or other technical information available. They were prepared to have a go at fixing absolutely anything. Not always successfully, but the success rate was a lot higher than you might expect.

This company specialised in this field.
Often it was just a bare circuit board removed from something, of which the function was completely unknown !

I was absolutely stunned by their approach, and initially just laughed at them when it was first explained.
But I was astounded to see how effective their weird techniques were.

The basic idea is first replace every electrolytic on the board. Replace every opto isolator, and replace any small signal transistors. It was explained to me that on a board that is old (10, 20, or 30+ years or more) all these components degrade due to age.
Amazingly that would fix more than half the faulty boards, and without even powering the board up.

The next level of repair involved very close inspection under a powerful stereo microscope of every soldered joint. Some strategic resoldering also caused many faults to go away. The interesting thing about all this, is that the whole repair technique could be learned and successfully carried out by people without any deep electronic technical knowledge, or sophisticated test equipment.

Now I initially laughed at all this, but it really did work. It not only worked, but was fast.

Obviously there were stubborn cases where all this failed to repair the equipment, and some real expertise were called for. But the whole concept just astonished me.

Another "trick" was to take a known good working identical board (supplied by the customer) and a faulty board, and use analog signature analysis to probe both boards node by node to find what was different.

This would quickly pick up faulty components, especially ESD damaged CMOS chips.
I believe this technique is also now used widely by the military to repair complex systems, as it can be carried out successfully by people with a fairly low skill level.

So I can well imagine your surprise at being told op amps degrade over time.
I did not believe it at first either, and still do not understand the mechanism how the Hfe of a transistor can gradually reduce over time.

Strange, but true.

Warpspeed said:

I did not believe it at first either, and still do not understand the mechanism how the Hfe of a transistor can gradually reduce over time.

Click to expand...

Most electronic components, when used within their specified limits, fail due to random chances. Some fail rather early in their life (infant mortality) and if they are weeded out, the remaining ones will last a rather long time.

The failure rates will be a random one but will have a characteristic half life. The half life will be depending on the quality of the manufacture and the stress of the use.

The quality of manufacture is an one time event; but the stress incurred during use must not be cumulative. If this is ensured, then you would see that the Hfe does not reduce gradually, it does suddenly.

The causes can be many (and that includes one unknown) but they all are stochastic in nature. A board fails when one single critical component on the board fails- therefore it is important to improve the survival rates (half-lives) of the least reliable components.

Thank you for the comprehensive reply and apologize for the late response.
I am glad that you have had the same experience like me.
Have you ever happened to read anything relevant to your experience in a book or paper?
I mean, would you please introduce me any readings about the above issue?

Thank you very much

Its not a commonly known or formally recognised technique of equipment repair, its more like a trick that sometimes works and sometimes does not.

I am not at all surprised that nobody is prepared to suggest something like this in a book or academic paper.

Now happily in retirement, I have had many long years working in the electronics industry on the hardware repair and design side, but my short time with that company was the weirdest and most amazing experience of my entire career.

One known problem is dendrite growth inside the OC170 series of transistors, causes shorts to the case. This could be applicable to other tin coated components.
Frank

One thing is for sure, if you have worked for any length of time repairing and restoring old equipment, the same types of problems can be expected to be found again and again.
Corroded battery terminals, noisy potentiometers, dirty and oxidised pins and contacts in plugs and sockets, cracked boards and dry joints.
All mechanical type ageing things to be expected.

I am stating to see more instances flash memory over ten years old starting to drop bits and crash software.
This is going to be a much bigger problem in the future.
Flash memory seems to be nowhere near as robust over time as the old 27Cxxx EPROMS.

Warpspeed said:

I am stating to see more instances flash memory over ten years old starting to drop bits and crash software.

Click to expand...

Pen drives have become cheap and we do not really mind to discard one that has stopped working. But the failure rates are becoming alarming these days- you can't use pen drives for any serious data storage...