Reliability Science in Audio Equipment

[antcollinet]

MTBF calculations based on statistical component lifetimes tend not (IME) to correlate well with actual lifetimes.

They don't take into account operating conditions - and most reliability issues come from those components being operated too close to, or beyond limits. In most cases in ways that are not even recognised by the design team at the point of release. This is the case for example in the Topping PA5 debacle.

Good reliability engineering is about putting processes in place that can spot these design weaknesses systematically (or better yet systematically prevent them from even being designed in in the first place), so the human "expertise" (or lack of it) factor is removed from the equation.

Reliability engineer : "Oh, so you're going to use potting compound in your design are you"

Reliability engineer : *turns to tall shelving stack, and using a ladder lifts down a 6 inch thick folder from one of the high shelves - blows the dust off it and hands it to the project manager*

Reliability engineer : "Well, in that case your team are going to have to run through every single one of these additional checks and tests, and have them reviewed by the technical board - should take no longer than about 6 months if the board accepts your results first time"

Reliability engineer : *Wanders off, whistling*

 

[sergeauckland]

In my past life in Broadcasting, whilst reliability was important, resilience, i.e. the ability to survive a failure, was even more important. That's why single points of failure were avoided as much as possible. For example, transmitters were always installed as pairs, either a parallel pair where a failure of one would reduce power but still carry on uninterrupted, or as a standby where a failure in one would result in a short loss of output whilst the automatic standby would switch over.

In the studio, there were always at least two studios with on-air capability such that a failure of one would not stop programming.

In spite of all that, there were still unavoidable single points of failure such as the antenna tower collapsing, but even then, broadcasters had mobile temporary antennas available that could restore service after a catastrophic failure with a few hours.

Sadly, with commercial broadcasters, they calculated the cost of the backup systems against the loss of advertising revenue, and often preferred to risk the loss of the advertising than invest in the backup equipment.

S.