Does temperature have an impact on modern electronics?

Bill DeBrooke posted:

I searched but did not find this topic.  The reason I ask is because of the previous topic covering the new train building being constructed in North Carolina I believe.  He sized a 16,000 btu AC for his 400 sq. ft. building which may be fine in NC.  In south Texas the building would have been different in two ways.  First, the pad would have been 18-24" above grade to deal with the flooding issue we face.  Second, the AC would need to be a little larger to deal with the high 90's to 100 degree days we experience from May to, I am not sure when.  All this week will be in the mid 90's.

Anyway the question is, how high can the temperature get before the electronics begin to fail?

Bill DeBrooke posted:

Anyway the question is, how high can the temperature get before the electronics begin to fail?

Anything you have will run to at least 70C, but that's obviously inside the case.  Automotive electronics are all rated to at least 85C, and aerospace stuff runs at 125C.

That being said, every 10C increase in temperature typically can cut the MTBF of a device in half.  So, if it's rated to run at 50C internally, and it's running at 60C, you can expect statistically to have failures in half the running time.

Obviously this is a gross approximation, and the specific design of the circuit has a lot to do with the MTBF and the effects of temperature.

The short answer is, there is no short answer.

Aspects of cooling like air flow and ventilation come into play here. Generally speaking, circuit resistance increases with an increase in external temperature, thus generating more heat. When I place my home theatre components in an entertainment center, I am careful to allow for adequate air flow from component vents. Heat dissipation is the key to prolonging the useful life of any electronics.

While John has correctly stated the temperature ratings (commercial, industrial, military) that semiconductors are rated at, it must be noted that these components can generate heat of their own. Now if you operate them in a closed environment , especially with other items that generate heat, and place them in a warm location, rising temperatures could cause degraded performance and possible failure. There are certainly many variables that determine, the life of a semiconductor integrated circuit performance.  On model trains the PC board components are likely rated to 70*C (158*F)  but being of Chinese manufacture there is no way to really verify the rating.

Apply some common sense. You are not going to sit in a room 90 to 100 degree F and play with your trains. The electronics are good to a much higher thermal level if they are not running and generating heat. Most of the current electronics are tested in a climate controlled room (70F) and do not fail.

When I was seriously searching for postwar Lionel in the 60's and 70's I was never concerned about how they had been stored unless it was some damp basement and that problem was always obvious.  Today the problem is very different.  I regularly see new in the box items for sale that are from the early 90's as newer electronics were beginning to spread into every aspect of engines.  There is no way to know what you are going to get when you open the box and put it on the track.  Phrases like, 'smoke free home' and 'stored in climate controlled space' always add to that unknown.

Where I live in south Texas you are used to the dangers of buying a really good deal on a used car.  Flooding damage can not always be detected without careful inspection.  It seems to me that the more electronics that are added into trains and the more time that passes the more difficult it is going to be to find working equipment.  The days of one on one transactions are fast disappearing as a larger percentage of sales move on line.

You will have GREAT difficulty killing the electronics with any normal storage temperature, even an unheated or conditioned space.  Also, the maximum storage temperatures for most semiconductor parts is greater than their maximum operating temperatures.  Since the operating temperature is raised by internal heating, the effect is greater than it appears.

I always admired/respected the durability of pre and post-war trains from operators who had/have attic layouts, where it can get quite cold, yet those old trains would sit and operate just fine when called-upon for duty.

As I think about it, that state of metals and electronics being very cold, and then suddenly heating-up from operating, would that be an electronics killer, ostensibly due to contraction and sudden expansion? I recall from my auto-mechanics days that engine damage can occur within the first thirty-seconds of a cold motor being started. But, we're talking a different animal here, with many moving parts and thermal expansion of a much higher temp.

Paul, there is an issue with what is called infant mortality in the industry.  In the first few hours of operation, statistically you'll have more failures of semiconductor devices than once they have accumulated some operating hours.  However, once they have been burned in, temperature excursions within their operating temperature range should have minimal impact.  In the aerospace world, we actually tested equipment with rapid heating and cooling, starting up at minimum temperatures (typically either -40C or -55C), etc.

The packaging of modern integrated circuits is designed to withstand normal temperature cycling without issue.

A far more common failure is the aluminum electrolytic capacitors, those will likely fail far before any semiconductor devices.  They're also more sensitive to high temperatures, the electrolyte dries up.

John, this discussion reminds me of a fellow that use to post on the forum quite a lot back in the old days, early 2000s, and while I don't recall his name I do remember a conversation I had with him - and that he left his Z-4000 transformers on all the time - 24 hours/7 days a week - as he believed that would make them last longer. He was under the impression that going from cold-static state to an energized state was harmful. I really had a hard time believing his theory.

The effect of heat and moisture on trains is usually negative as well as large, rapid temperature changes.  However these effects would be best answered with far more factual information than has been supplied thus far.

The exact design, expected desin life/failure rates, the operating  environment, storage conditions, and the components involved have a large impact on the answer.  Electrolytic caps will have a high failure rate in time vs tantalum caps.  And the price reflects that as well.

Some of the main problems with prewar and postwar devices (ignoring usage) is degradation of the lubricants and decals, discoloration,  zinc pest, and warpage.  For the most part they were well designed and manufactured.  I doubt their designers were expecting an EOL of 100+ years but that looks very likely with reasonable care.   Very few had integrated circuits or circuit boards.  Comparing them to current train products is fruitless.  

Today's trains could be designed to outlive their owners, but that doesn't appear to be the design criteria.  

Having the trucks stay on when the train is removed from the box would rank high on my list.

Well actually NOTHING is designed to last any more, quality control is now handled by the buyer, this is because the company's have done away with quality control now the buyer is the quality control inspector. This is how big business have saved the wages and healthcare cost not having to hire a quality control inspector at all! just look at everything tv's radio's cell phones all consumer electronics, a lot of job for repair tech is just going away little by little. look how many tv radio appliance repair shops have closed over the years. Most company's would just rather sell then throw away anything that breaks  save money on repair technicians.my Uncle use to work for IBM company and trouble shot down to the bad component. now a days the same company uses software to tell the company which part is bad and anyone can change the part with out all the schooling n training! so why would you expect  a train company to design boards not to fail especially when made in china Korea. everything is made to sell for a price its all about sales only! 

Alan Mancus posted:

 Most company's would just rather sell then throw away anything that breaks  save money on repair technicians.my Uncle use to work for IBM company and trouble shot down to the bad component. now a days the same company uses software to tell the company which part is bad and anyone can change the part with out all the schooling n training! so why would you expect  a train company to design boards not to fail especially when made in china Korea. everything is made to sell for a price its all about sales only! 

How long do you allow a technician to isolate and replace a component that costs pennies?  15 minutes? Half hour? Hour?  Diagnostic software is only as good as the software engineer and doesn't catch everything.

Even the company  I retired from (industrial controls made in the USA, and some of out units operated in hostile environments,) unless there was obvious component damage, it was less expensive to replace a defective board than troubleshoot down to the component level.  Plus the repaired board would have to be tested individually and again in the customer's unit.

Rusty

Alan Mancus posted:

Well actually NOTHING is designed to last any more, quality control is now handled by the buyer, this is because the company's have done away with quality control now the buyer is the quality control inspector. 

Actually many electronics are designed for long life and high reliability.  Heart pacemakers, most medical equipment, military hardware, most aspects of air travel and ground control,  automobile control electronics, shipping, oil well and well logging, automobile traffic control, industrial surveillance equip., internet servers, mainframe computers, to name just a few categories.  And most phones have fairly high reliability rates considering the operators and their environments.

Very little of the electronics world is repaired at the component level.  However, most of the electronics groups listed above do have their defective units returned to the source for repair at a component level.  The cost to discard the unit is cost prohibitive.  What is usually of greater importance, are the results of the "autopsy" on the field returned part.   Lord help the unlucky entity that is the source of the failure.  Most customers have the ability to track the life history of the electronic products to the minute station by station as they are being built new, or in for repair.

My employer builds a lot of those categories of electronics worldwide.   A separate autonomous division of our company does the autopsies. 

It's not a pleasant experience to be on a conference call with the "customer" to discuss the autopsy findings.  There are usually hi magnification photos and product cross sections that pin point the exact source of the problem.

There is also financial responsibility to consider for units that fail during the negotiated warranty period.

It's not nearly as casual as you may think.  We're not talking your big screen TV,  microwave oven, or most consumer electronics (like trains).  

Just ask what the longest extended warranty period is available for purchase on your new electronic toy and you'll know when to start shopping for a replacement.

As far as I know, there aren't any extended warranties offered for toy trains, so that should give you a clue.

Alan Mancus posted:

Well actually NOTHING is designed to last any more, quality control is now handled by the buyer, this is because the company's have done away with quality control now the buyer is the quality control inspector.

You're talking about consumer electronics that are generally considered "throw-away" items.  There are a lot of electronics you interface with in a daily basis that was designed for longevity, consider the millions of cars that have extensive electronic controls.  While they do have failures, statistically they're few and far between.  Of course, my usual example is avionics, a topic I'm pretty familiar with.  We designed for longevity and reliability.  Consumer "toys", regardless of price, are simply not subject to the same rigorous quality standards.

There was, may still be, an ongoing debate about leaving a computer on 24 hours a day or turning it off when you aren't using it. Some folks thought that the sudden "jolt" of turning the power on wasn't good for the electronics. They felt that leaving it on all the time was much better for it. When I worked at Cook County Jail in Illinois, we left our computers on 24 hours a day. Didn't seem to have any negative effect on them. I suspect that the electronics on most modern trains can survive heat that most humans wouldn't want to play around in. I also suspect (or hope) that Lionel and other manufacturers would test and require that their products be able to survive temperatures that normal human beings would be using them in. Storing them...that would be a much greater variable. But Lionel must realize that people store this stuff in attics and basements that don't have "optimum" conditions.