|
...plus, I told Klaus at the outset that the only way I
would design a micro-based ignition is if it was triply redundant with majority
voting on the outputs. And that he better be ready to do Monte-Carlo simulations
with all kinds of different things going wrong to demostrate that it would not
harm the engine or pilot when some sensor or transient failure occured. I
really like computers; I use one all the time to design my props. I've worked
with them since 1958! But some of the problems I've had have been with
software routines programmed by others, that took control or turned off
interrupts. 'Had that last happen on an Atlas flight! But I have to
consider what the software in a single-processor ignition or fuel injection
system would do if a cosmic ray/high energy particle came burrowing through
the solid-state program memory and changed the state of one bit. For those
who don't know about this, it's referred to as an SEU, a single-event upset.
That's why space hardware has to be radiation hardened. But it can happen at
sea-level too, but occurs more often the higher you go. It also is more of
problem with with the more miniature, smaller-junction size devices. Since an
ignition system is a relatively slow system, it's better to use the larger,
slower devices which have greater immunity to SEUs due to their junction size.
My rule is to not use faster gates than what are necessary!
These are some of the things you should
ask someone who wants to sell you a single-processor system. BTW, Klaus tells me
that one of his customers had a lightning strike that let the smoke out of his
avionics but the Plasma kept on ticking.
|