inventor wrote:I don't understand why anyone would want to build a single point of failure into their electrical system
As an IT professional I spend a lot of time thinking about redundancy and points of failure. I see the sentiment quoted above on the Sonex forums a fair amount and it baffles me because there are *lots* of "single points of failure" throughout the entire electrical system for most aircraft; at least until you get up to Biz-jets and Airliners.
- Replacing the contactor does nothing to solve the fact that there's probably a single wire between your battery and your main bus. So that's a "single point of failure".
- Do you have multiple parallel wires running from each fuse or CB to each item that it powers? If not, then all of those "single string" wires are single points of failure.
- Do you have multiple fuses or CBs providing power to each item? If not, then each of your circuit-protection devices is a single point of failure.
- Do you have fully redundant instruments? If not, then each one of those is a single point of failure (whether electrical or driven by pitot/static inputs).
- Do you have multiple wires feeding each ignition coil? How about grounding each magnetron?
Each one of these items, when installed and maintained correctly, is
highly unlikely to fail. If you add redundancy for these items you're adding extra parts, complexity, and new maintenance items (& failure points) that you will likely never need - yet you carry around the cost (in money, in weight, and in labor) on every single flight.
And if you only add redundancy to one of two of these things, how much have you really improved your odds of never having a failure; given that each scenario is so very unlikely to happen? Yet by adding that redundancy have you lulled yourself into a false sense of security? And have you fully taken all of the additional necessary maintenance into account, and are you performing it diligently?
In the IT world we focus a lot on being "fault tolerant" and having "risk mitigation" factors. That is to say, rather than just buying 2 of everything and trying to run with perfect redundancy we architect our systems so that a small failure does not cause a major problem, and we try to make sure that any failure scenario merely degrades services in as safe and graceful a manner as possible.
There are lots of things that can fail in an VFR aircraft and you can still have a perfectly safe conclusion to your flight. For example: as long as you're not flying in the clouds, losing your artificial horizon is mitigated by the fact that you can visually orient yourself. Heck, even losing your voltage regulator (assuming you can isolate it from your electrical bus) is a manageable fault that simply results in landing sooner than you might've intended.
Adding parts and redundancy does not necessarily raise safety. Most of us by now have been exposed to the topic of safety in light twins versus single-engine GA aircraft, and that example can serve us well in this instance.
If you want to add things to your experimental aircraft, that's totally fine and is the right of any builder. But as Bob N (author of "The AeroElectric Connection") has said in his interviews on the SonexFlight podcast: its important to think through the positive and negative effects, and understand all of the failure modes of each item you add to your aircraft.
--Noel