I think Mr. Ridgeway's question is ill-formed.  The question 'are composite parts safe?' is too general.  The question should be 'is this composite part safe?'  He makes it sound like this material was discovered last week and put on an aircraft.  It is possible that this specific rudder was underdesigned.  It is also a fact that no design can witstand all possible flight condidtions.  If the plane was flying to fast for the turbulent conditions it was in, then _no_ design (which would exclude those conditions by design) would survive.  It's time to allow the crash investigation to determine what happened.

-Kirby

Kirby Wilkerson

Remember, first define the problem, then solve it.

Carbon fiber is strong, and stronger than aluminum, but here's my take on it's weakness that is my guess on the cause of the crash.

Carbon fiber conducts poorly (3 ohm per square) compared to aluminum and needs metallization over the fibers. Carbon Fibers tend to move when stressed which moves the metallization which is painted on typically. This will create microridges in the metallization surface (or possibly small gaps) which creates an inductive surface. The impedance of an inductive surface on a lightning strike is high due to the fast rise times of the lightning pulse. This high impedance surface makes lightning go to a lower impedance path through the carbon fiber and into solid metal under the carbon fiber. Lightning jumping through a surface explodes the surface violently, it doesn't just make a little peep hole.
 
Hence, I'm voting lightning strike hit that went through the carbon fiber surface. Hence the bright flass seen by two other pilots who were flying behind the ill fated aircraft.  

I think eventually, they'll add solid metal strips onto some surfaces of carbon fiber (front of wings, front of fuselage).

If a tail broke off due to high speeds, maybe you'd get some arcing of electrical wires, but fuel would have to ignite for a bright flash to occur. That's still possible of course, but I'll go with lightning first (mainly because I've working in the lightning area and the carbon fiber area while at Raytheon and I know the headaches Beech Aircraft had in metallizing their first carbon fiber aircraft for protection against lightning strikes).

kch

Higgler,

I though that carbon was fairly conductive.  Although not as conductive as aluminum wouldn't it do a fairly good job of dispersing the charge over a larger area?  or is the resin so non-conductive that even if the charge can be carried by the carbon tows, it can not easily spread from tow to tow.  It sounds like you are the right person to ask from your experience at Ratheon.  Don't some fabric weaves include a metal filament with the fiber?

Thanks for sharing details about this.

-Kirby

Kirby Wilkerson

Remember, first define the problem, then solve it.

Vertical stabilizers have been breaking off airplans for a long time.  The Navy Tacamo program lost a vstab off of a 707 during flight testing, a B-52 came home without the vstab, there has been combat damage that left airplanes coming home without it as well.  Also crop dusters regularly slice off the vstab flying under wires they shouldn't be flying under.  The metal ones break off sometimes, why would we think a composite part isn't subject to the same failures.
..
And I'll throw in my 2 cents on pitot tubes.  I cant think of any pilot in cruise flight who would change power settings due to an erratic airspeed indication, particularly if it wasn't associated with groundspeed deviations on the FMS.  You just leave the power set and evaluate the problem.  If 1.80 EPR got you there, it keeps you there.  And in a thunderstorm, erratic altitude and rate of climb info isn't too unusual, but if it correlates to the direction the airplane is tossed around, it is ok, and pilots are trained to hold pitch and bank.  And the aiutopilot will do the same with altitude hold off.  And as long as the airplane has the structure to handle it, they pop out the other side off altitude and heading, but in one piece.
..
And if the tail comes off for any reason in a thunderstorm, there would be no way to keep the pointy end in front, and any of the big airplanes would begin to turn side on to the relative wind and begin breaking up.
..
But then noone knows if the tail came off, of if it did, why.

KirbyWan, I found some info on the internet.

for an aluminum layer that is 200 microns thick (or 0.02cm thick), since the bulk resistivity of aluminum is 0.0000027 ohms-cm 0.0000027 / 0.02cm =0.000135 ohms-per-square or 1.35 x 10^-4. This assumes that the aluminum film is pure, since the bulk resistivity value was taken from the periodic table of the elements.


Which makes carbon fiber 3/0.000135 or 22,000 times more resistive than aluminum.

I got the carbon fiber 3 ohms per square value from a former Raytheon CoWorker who designed reflector antennas for space applications at Loral Space System in San Fran. That's the value they put in the computer for calculating antenna performance.
Carbon fiber being a group of fibers makes it inherently less conductive compared to a sheet of conducting metal. Fibers are also much more inductive than a sheet based on their inherent shape. You make conductors out of thin long wires.
The thing I recall about the Raytheon Beech metallization of their Starship was that lightning was a huge concern that they had to fix the problem by adding a better conductive coat to the carbon fibers. I'm guessing they did lightning arcing tests on carbon fibers by themselves and saw some bad results (purely a guess on my part). The way it was worded to me from a mechanical Engineer in Wichita Beech was that lightning was a major concern on the Beech engineering team.

another article from the web on carbon reinforced concrete;

One property of carbon fiber is its piezoresistivity, and
it is observed that concrete reinforced with carbon fiber has a self-sensing ability. Therefore,
damage can be monitored by measuring D.C. electrical resistivity. Damage increases the
electrical resistivity of the specimen. In the case of major damage, this increase is irreversible
and fiber fracture is found to be the controlling factor.

If you can measure resistivity in carbon inside concrete, it's not very conductive.

Excessive tension causes the resistivity of the fiber to increase, due to damage. Thus, carbon fiber in epoxy is a sensitive piezoresistive strain sensor.

Feels to me that in violent winds, add a lightning strike on just the wrong location, and bad things happen.

All pure guesswork of course. I'm no expert.


kch

'ohms per square' is not a unit of resistivity. The typical fibers used in airframe (HS and IM) have longitudinal reistivities of 14–16 micro-ohm.m, or ~0.0015 ohm.cm in Higgler's preferred unit. They are about 500 times more resistive than pure aluminum (7000-series alloys are about 0.033 micro-ohm.m, so not hugely different).

Most external carbon airframe is protected from lightning strike by an outermost layer of Coppermesh or similar. Protecting the material by metallizing the fibres with nickel or similar is sometimes done, but at the moment a highly conductive outermost layer is preferred. This has been shown to give good protection in artificial strikes mimicking the worst known naturally occurring ones.

L/S is unlikely to be the cause, though of course it remains a possibility.

The Airbus and Boeing 777 carbon fins and tailplanes have had very good service histories, with the loss of the fin in the AA587 Belle Harbour crash being due to an estimated load of about 2x limit according to the NTSB (finding 6 in AAE0404 below).

Accident report:
http://www.ntsb.gov/publictn/2004/AAR0404.pdf

Has an interesting material analysis article:
http://www.ntsb.gov/publictn/2006/JRN0601.pdf
 

Ohms per square is indeed a measure of resistivity, and may be more appropriate for sheet materials than bulk resistivity: http://www.four-point-probes.com/ohms-cm.html

Listing of recorded aviation incidents involving lightning: http://www.aviation-safety.net/database/dblist.php?Event=WXL

TTFN

FAQ731-376: Eng-Tips.com Forum Policies

Apologies for my ignorance (I thought it might have been a weird typo). Thanks for the correction IRstuff.
 

No problems.  

It's an extremely common term in analog IC design; you determine the length of a resistor in the layout by counting "squares" of length, multiply by the sheet resistivity and get the total resistance.

TTFN

FAQ731-376: Eng-Tips.com Forum Policies

My old boss said that you could make anything [any design] work. The trick is making it work analytically and in practice. Learn from testing and discovered mistakes in design.