Why Not Use PC To Make Glass Cockpit?

wrote:
Are you ready to bet your life on Windows XP and some $200 disk drives made
in China and designed to operate in an office?


Actually yes. Of course, I would take into the account of
altitude/pressure/temp problems. As a matter of fact, I would be more
comfortable with commodity components than something that's customized,
unless the customization were very very trivial.

Regards,

-Chaud Lapin-

Nathan Young wrote:
On Sat, 18 Jun 2005 23:02:25 GMT, john smith wrote:

Le Chaud Lapin wrote:
Being a software developer, I am very suprised to discover that not
every aicraft costing over $30,000 has a full-featured glass cockpit.
Unless I am missing somethnig, it appears that everything that a pilot
needs can be made with very very cheap hardware.

Already been done.
10-15 years ago Burt Rutan had one of his aircraft (Catbird?) completely
controlled by an Apple laptop computer with custom coded software.

His asymetric twin, Boomerang used a laptop for much of the
instrumentation.

I think this is just wonderful. I've always dreamed of making an
ultr-lightweight aircraft using advance controls like this.

Hopefully in a few years I will be able to jettison my current
occupation and devote my effort to it full-time.

-Chaud Lapin

wrote:
"'Vejita' S. Cousin" wrote:
Not the original poster, but actually yes I am A good stable system
that you do not touch/play with should work/last without problems. And I
don't need to run winXP I could run linux (in fact given the costs and
licenses I would) which is rock solid stable.

You still need sensors for attitude and heading. You also need sensors for
engine and fuel parameters, and so forth. Somehow, you need all those to
interface with your PC's flight instrument display (displays?).

USB. I think it would be highly appropriate for this type of
application. Theoretically, you can have up to 127 USB devices
pluggged into the same USB card on a PC. If you put multiple cards in
the PC, then you can have even more USB devices. Since the required
bandwidth for temperature sensors is very low, It is conceivable that
one USB PCI card could drive the entire control infrastructure of the
aircraft.

The design would probably entail a redundancy model where one PC
motherboard acts as a secondary to the primary controller. Windows
supports multiple displays, so you could have 4 displays in the
aircraft. One for each passenger, with radio, music, DVD's.

I think this model is the same thing that Boeing/Airbus & friends use,
only they probably get charged 10-20 times the cost of what they should
be paying.

-Chaud Lapin-

What about attitude and heading?

Le Chaud Lapin wrote:

wrote:
"'Vejita' S. Cousin" wrote:
Not the original poster, but actually yes I am A good stable system
that you do not touch/play with should work/last without problems. And I
don't need to run winXP I could run linux (in fact given the costs and
licenses I would) which is rock solid stable.

You still need sensors for attitude and heading. You also need sensors for
engine and fuel parameters, and so forth. Somehow, you need all those to
interface with your PC's flight instrument display (displays?).

USB. I think it would be highly appropriate for this type of
application. Theoretically, you can have up to 127 USB devices
pluggged into the same USB card on a PC. If you put multiple cards in
the PC, then you can have even more USB devices. Since the required
bandwidth for temperature sensors is very low, It is conceivable that
one USB PCI card could drive the entire control infrastructure of the
aircraft.

The design would probably entail a redundancy model where one PC
motherboard acts as a secondary to the primary controller. Windows
supports multiple displays, so you could have 4 displays in the
aircraft. One for each passenger, with radio, music, DVD's.

I think this model is the same thing that Boeing/Airbus & friends use,
only they probably get charged 10-20 times the cost of what they should
be paying.

-Chaud Lapin-

Le Chaud Lapin wrote:
USB. I think it would be highly appropriate for this type of
application.

It's way too easy to trip over the cord with USB -- or to have it
vibrate loose. My experience with USB indicates that it isn't reliable
enough for my servers at work, it's not good enough for any airplane I fly.

Maybe if you replaced the connectors or soldered everything well....
But, I'm reluctant to suggest USB to my users for anything more
permenant than a mouse.

-Luke

Le Chaud Lapin wrote:
wrote:

Are you ready to bet your life on Windows XP and some $200 disk drives made
in China and designed to operate in an office?

Actually yes. Of course, I would take into the account of
altitude/pressure/temp problems. As a matter of fact, I would be more
comfortable with commodity components than something that's customized,
unless the customization were very very trivial.

If you do that, don't ever plug the machine into the Internet. Also
remove IE.

I take care of about 250 computers, and there's no way I'd bet my safety
on a desktop Windows installation. A customized/embedded/trimmed
version that runs off a ROM might be acceptable if it's been through a
*very* rigorous testing process -- but after you've seen the difference
in performance between a clean machine and a machine after it's been out
in the real world (even with Mozilla Firefox, AdAware, Spybot S&D, *and*
Symantec Corporate Anti-Virus 10 running on it), you won't want to trust
your @$$ to what most people think of as Windows either. A
trimmed/embedded/customized Linux would work too, although I think I'd
be most comfortable with a device running something like VxWorks[0].

As for the disk drives, I recommend a pair CompactFlash storage modules
wired up as IDE devices[1] set up in a RAID mirror setup. That way, the
system is more resistant to things that damage moving parts, and with
the RAID setup, it will continue to run if one of the CompactFlash
devices fails. Use three or four if you don't feel lucky. Of course,
you couldn't run desktop Windows on this because the swapfile would
quickly wear out the CompactFlash (they can only be written a finite
number of times[2]).

-Luke

[0] In one of my earlier jobs, my task was to make VxWorks run on a
single-board-VME Sun clone. I got to read some of the source code, and
it's some of the most carefully documented/commented and sanely designed
C code I've ever seen.

[1] CompactFlash is electrically similar enough to IDE that they can be
wired into an IDE bus without the need for a chip.

[2] I don't remember exactly what the number is -- I seem to remember
that a USB keydrive could take about 500k writes, but it wouldn't
surprise me if that has been extended to a couple of million. Also,
most Flash devices will remap data internally to avoid repeatedly
writing a particular block -- so that the device will last longer.

Luke Scharf wrote:
Maybe if you replaced the connectors or soldered everything well....

Er, I meant to say that it might work if you soldered everything or
replaced the connectors. My post-writing editing made what I was trying
to say unclear (though better organized!)

-Luke

Luke Scharf wrote:
Le Chaud Lapin wrote:

USB. I think it would be highly appropriate for this type of
application.


It's way too easy to trip over the cord with USB -- or to have it
vibrate loose. My experience with USB indicates that it isn't reliable
enough for my servers at work, it's not good enough for any airplane I fly.

Maybe if you replaced the connectors or soldered everything well....
But, I'm reluctant to suggest USB to my users for anything more
permenant than a mouse.

There is already an adequate standard buss for interconnecting avionics
devices. I'm sure you can purchase a PC card to interface to that.

But you're on the right track. Non-stop, fault-tolerant computing is
the issue.

Stubby wrote in message ...
Luke Scharf wrote:
Le Chaud Lapin wrote:

USB. I think it would be highly appropriate for this type of
application.


It's way too easy to trip over the cord with USB -- or to have it
vibrate loose. My experience with USB indicates that it isn't reliable
enough for my servers at work, it's not good enough for any airplane I
fly.

Maybe if you replaced the connectors or soldered everything well....
But, I'm reluctant to suggest USB to my users for anything more
permenant than a mouse.

There is already an adequate standard buss for interconnecting avionics
devices. I'm sure you can purchase a PC card to interface to that.

But you're on the right track. Non-stop, fault-tolerant computing is
the issue.

The Space Station uses IBM 760xd laptops for their glass cockpit.

http://spaceflight.nasa.gov/gallery/...l/sts105-304-0
25.html

http://spaceflight.nasa.gov/gallery/.../iss002e5478.h
tml

http://spaceflight.nasa.gov/gallery/.../iss003e5552.h
tml

http://www.spaceref.com/news/viewnews.html?id=213

http://www.cisco.com/en/US/tech/tk65...study09186a008
00b53b6.shtml

wrote:
What about attitude and heading?


You know how electrical engineers are. They drool over the thought of
making a gadget that is compatible with an interface like USB. If it
can be done, someone is probably doing it.

altitude
velocity (translational and rotational)
acceleration (translational and rotational)
humidity
windspeed
attitude
temperature
pressure
GPS coordinates
illumination
etc....

All they do is take the existing devices, throw on USB interface, and
and it's good to go (and cheaper).

-Chaud Lapin-