Actually the requirements have pretty much converged. We have actually done work for controls in nuclear, aerospace and now
autonomous land vehicles. Designing for cars is actually the toughest since cost has to be low. Aerospace has some weight restrictions
but SW reliability and system reliability are identical. The operating conditions of a car in some sense is a lot worse. MTBF rates have to be
some an order or two of magnitude higher. You would be surprised to know that the physical electronics and the control software have pretty much converged. To the extent that in our work, we no longer design specifically for each sector. The control processor for a motorcycle for example
will be designed to be as reliable as the control processor in a fly by wire control processor in an aircraft. Think about it, why on earth
should an avionics system be more expensive other than handling g forces, lower temperatures and vibration. More ruggedized HW but the
system design remains the same. Used to design boards for Curtis-Wright almost a decade ago (yes from the Wright brothers !), more rugged
but not earth shakingly more complex.
To have a meaningful conversation on this, list the parameters which you think increases design complexity. Do you think MTBF should be lower for
automotive ? or is it environmental conditions ? So how come COTS has become so popular in aerospace. To give an example,
in the RapidIO standards, space qualification is basically an increase in error handling and toning down some features. It is
just a simple extension called part S. Now this is the difference between something used in a data center and something used in aerospace.
For auto, I will simply use the aerospace variant including the ruggedized connectors but cooling of the boards will be simpler.
You grossly underestimate the design complexity of the electronic systems that go in a high end car given the complexity
of the electronics we add nowadays. I am of course restricting my statements to avionics.
A Nasa presentation sheds light on this issue.
http://nepp.nasa.gov/workshops/etw2013/ ... 0Parts.pdf
Note that current generation processors can support radiation hardness at much lower cost. In fact these are recommended for
autonomous vehicles, the advent of which is what making the worlds converge. The rest of the concerns has to do with
qualification and testing methodologies. This is just binning and not designing specific aerospace components.
Check out NXP Qorivva or TI Cotex R5 parts.
So while you cannot just take an auto module and plunk it in a plane, it is not hard to get a high quality auto manufacturer to
supply aerospace components at lower costs. Th differential is not that great. In fact teh biggest pain is the 25 year supply
(typically 10-15 for auto) and
storage criteria. Can be done but volumes are not high enough to justify the cost. But if components are shared, it is a different matter.
nileshjr wrote:macaque wrote:One could very well make the opposite case, that the standards in auto industry are higher than in the aerospace industry. Think of the
lowly Maruti 800. It has to last 10-20 years with minimal maintenance, with no fuel quality control, no sheltered storage and driven
by mostly unqualified drivers ! The point is we do not acknowledge the reliability that auto components exhibit. If you
think of newer standards like ISO26262 (for sw and reliability) and the fact that safety critical systems in cars are approaching
aerospace levels, there is a case to be made for using auto components in aerospace.
IIT in fact is advocating using the same standards like ISO26262 and space grade SRIO interconnects in the auto and aerospace sector.
I frankly feel the distinction is arbitrary. My ABS better be as reliable as the fly by wire in the jet I fly in. There is no earthly reason
why they should have different reliability standards and no reason why aerospace grade systems should be expensive.
Time to a rethink on this issue.
You are going tangent to my point and perhaps you got mixed up in your concepts. I was referring to manufacturing specifically. I Don't agree what you say. Aerospace design requirements are far far more stringent than even the most extravagant sports car. I cannot overemphasize how much more reliable aerospace systems are as compared to cars. I know for sure as I design some. Don't be fooled by the longevity of Maruti800. A fighter which needs perhaps 20-30 man hrs of maintenance for every flight hr, doesn't mean its less reliable than a Maruti800 which runs many days without maintenance. The kind of operating conditions a jet has to face cannot be compared to anything that a car sees, ever. What would it be like to have a car which will never break down in the middle of the road for 25 yrs of rigorous use??
It would be absurd to use aerospace grade reliability standards for auto industry across the spectrum, since all it will do it make the design and manufacture too expensive. That ISO standard you refer to is much easier to apply to a car. Try using mechanical, material or manufacturing standards for auto and see what happens. Let alone airplane, just think of making a Maruti 800 as safe as a BMW or Ferrari in a crash. Would it still cost 3lakh?