# Things that to be considered while choosing electronic components to get high Durability

Creating the Electronic circuits is Ok not so much complicated for experts. But needs to think about the persistence of the electronic circuits. Its very difficult to make the circuits to work for life time. But its important to have the long run electronic circuits.

Of course it depends on individual components that we assemble on the PCB. And are there any factors that need to be keep in mind while choosing and assembling the components to have high durability.

• Are you asking specifically about the manufacturing process or component selection as well? – Kellenjb Nov 23 '10 at 17:17
• is there a question in there somewhere? – Mark Nov 23 '10 at 19:24
• Cost, reliability, performance, pick two. Making a design as "reliable as possible" is often cost prohibitive and not necessary. You need to design to the requirements of the application. – Nick T Nov 23 '10 at 22:46
• @Kellenjb - My question is about the Component selection. Sorry if I haven't conveyed that in the question. – iRobot Nov 24 '10 at 3:18
• @Mark - Sorry if there is a similar question exists in this forum. I haven't searched for this question. – iRobot Nov 24 '10 at 3:20

Don't use anything mechanical if you can avoid it - use solid state (like SSR's instead of mechanical relays, capacitive sensing/force sensing instead of switches, etc).

Use protection for every possible contingency - if it gets struck by lightning it needs to keep working.

Don't require any active cooling - fans break or get dirty too easily.

If it's in any kind of high-vibration environment, don't use connectors. Solder everything to the board.

If necessary, use redundant modules. I believe one of the Voyager probes has five independent computers any three of which are on at a time. They all work on the same problem and the solution that at least two of them agree on is used. The computer that created the bad solution is turned off and another is turned on in its place.

Have the ability to update any software on the device in the field when you find errors.

• I have to disagree with point 3, don't use active cooling. My 20 year old scope's fan is going strong - not even had to clean it out yet. – Thomas O Nov 23 '10 at 14:31
• @Thomas O - all you need to do is look at MTBF hrs for fans and see they are prone to failure. It's always best to base your judgements on evidence rather than anecdote. – Cybergibbons Nov 23 '10 at 17:09
• I have a counter-example. I rescued a Tek 454 from being thrown out. All it needed was disassembly and relube of the fan. – markrages Nov 23 '10 at 17:19
• @AngryEE - So cost is no object in your world? – Nick T Nov 23 '10 at 22:49
• @Nick - Depends how reliable/durable you want it. If you want high reliability you should be prepared to spend for it. – AngryEE Nov 24 '10 at 23:40

Run all components well inside their ratings, and avoid high temperatures.

Choose electrolytic capacitors 1.5x - 2.5x required voltage rating.

Run components at least 20 degrees C cooler than rated temperature. E.g. for 105C cap, run it at a maximum of 85C. A general rule is for each 10C drop in operating temparature lifespan is doubled.

Where possible avoid electrolytics where something like ceramics would work - although with careful consideration they can be used, they are generally pretty poor components.

Input protection on all inputs - including ESD and overvoltage protection. Ensure your project can survive voltage spikes on the power supply, if necessary.

• Good point about ESD. An ESD event may not cause a chip to fail immediately, but may degrade its performance or life. – Brian Carlton Nov 24 '10 at 16:56