control failure

Don: Be sure to share what you find. I think that I'm only scratching the
surface of some stuff. One thing I found and had never seen before was an
equation relating the increase in stress due to a crack. It scares the hell
out of me. I'm afraid to use a metal fork in my salad. The equation
basically says that the max stress is 2 times the load divided by the area
times the square root of the crack length divided by the radius of curvature
of the end of the crack!!! If the radius of curvature was equal to the
crack length, the max stress is already twice what you would calculate using
the applied load and the element cross section. Now put a reasonably sharp
crack and see what happens....as the radius approaches 0.001 times the
length of the crack......????


"Don W" wrote in message
et...
Stuart Fields wrote:

Don: I've got numerous phone calls to anodizing firms who have confirmed
the decrease in fatigue life due to anodizing. I've got a photo of a
fatigue failed anodized control tube, none of the non anodized control
tubes in any of the other similar helicopters even those with more hours
have failed. Further if you consult the excellent text:
titled Fatigue Design of Aluminum Components & Structures, Sharp,
Nordmark and Menzemer, a chart, page 110, shows decrease in fatigue life
due to pre-cleaning as well as the affects of Alodine and a couple of
different thicknesses of anodic coatings.

Further:

In a report authored by Thart, WGJ and Nederveen, the following was
stated:

"Constant amplitude fatigue tests on anodized unnotched specimens reveal
that sulfuric acid and sealed chromic acid anodic layers cause the
largest decrease in fatigue strength. Phosphoric and unsealed chromic
acid anodic layers do not significantly affect fatigue life. Scanning
electron microscopy of fracture surfaces confirms that fatigue cracks
initiate at cracks in the anodic layer".



Mo

Shiozawa, Kazuaki; Kobayashi, Hirokazu; Terada, Masao; Matsui, Akira.
Japan Society of Mechanical Engineers, Transactions A. Vol. 66, no. 652,
pp. 74-79. Dec. 2000

"The anodized film is fractured at an early stage of the repeated tensile
fatigue process, because it is too brittle to accommodate the substrate
metal."

Mo A P.E associated with the anodizing community said it even
stronger. "Never anodize flight critical components"

Van's of Van's RV aircraft and the subject was anodizing spars, said that
anodizing has been known to reduce fatigue life as much as 50%.



Boeing Aircraft has a special process whereby the ameliorate the effects
of anodizing on some parts.

I had a 36' McGregor catamaran with an anodized mast that I sailed in the
open ocean in the South Pacific. Even with the cracks in the anodized
layer, the frequency of vibration in the mast was much lower than the
17hz associated with the helicopter. Looking back I would expect the
mast on the sail boat to have a much longer life than helicopter parts.

More data. The failed control tube was inspected by a laboratory in
Canada and they proved that there was no existing flaw prior to the
anodizing. The crack started after the anodizing and the control tube
with a small load applied, but subject to the vibrations produced by a
helicopter, failed in fatigue with very few hours.

Experience can be misleading. I've been in the amateur helicopte game
since 97 and I'm a retired engineer but I had never heard that the
fatigue life of anodized parts could be reduced as much as 50%.


Stuart,

Very interesting! A lot of things that you run into in engineering are
counter-intuitive, and this is apparently one of them. I had not heard of
this phenomenon before now.

I'll certainly look into this some more when I get some time.

Don W.

Stuart & Kathryn Fields wrote:
Don: Be sure to share what you find. I think that I'm only scratching the
surface of some stuff. One thing I found and had never seen before was an
equation relating the increase in stress due to a crack. It scares the hell
out of me. I'm afraid to use a metal fork in my salad. The equation
basically says that the max stress is 2 times the load divided by the area
times the square root of the crack length divided by the radius of curvature
of the end of the crack!!! If the radius of curvature was equal to the
crack length, the max stress is already twice what you would calculate using
the applied load and the element cross section. Now put a reasonably sharp
crack and see what happens....as the radius approaches 0.001 times the
length of the crack......????

That is because the load is not equally spread
across the part, but is concentrated at the end of
the crack.

Think about a piece of metal bar in tension with a
crack halfway across it. The cross sectional area
that has already seperated cannot bear any load at
all as it has already failed. The end of the
crack is taking a lot of load because the crack
pulls apart when it is under tension.

Solution: Don't use cracked parts!! (Well duh)

Don W.

Don: Reading in a text book on Fatigue design for aluminum structures,
there is at least one design technique that assumes that there are cracks
inherent in the material to start with that are not readily detectable by a
visual inspection. The design philosophy progresses from there. This whole
area of study is explaining why the time life specs on parts can be very
meaningful. It also can allude to the excessive safety factors that can be
used in determining time life specs.
It looks like I've got a lot more studying to do in this area.

--
Stuart Fields
Experimental Helo magazine
P. O. Box 1585
Inyokern, CA 93527
(760) 377-4478 ph
(760) 408-9747 publication cell
"Don W" wrote in message
...


Stuart & Kathryn Fields wrote:
Don: Be sure to share what you find. I think that I'm only scratching
the surface of some stuff. One thing I found and had never seen before
was an equation relating the increase in stress due to a crack. It
scares the hell out of me. I'm afraid to use a metal fork in my salad.
The equation basically says that the max stress is 2 times the load
divided by the area times the square root of the crack length divided by
the radius of curvature of the end of the crack!!! If the radius of
curvature was equal to the crack length, the max stress is already twice
what you would calculate using the applied load and the element cross
section. Now put a reasonably sharp crack and see what happens....as the
radius approaches 0.001 times the length of the crack......????

That is because the load is not equally spread across the part, but is
concentrated at the end of the crack.

Think about a piece of metal bar in tension with a crack halfway across
it. The cross sectional area that has already seperated cannot bear any
load at all as it has already failed. The end of the crack is taking a
lot of load because the crack pulls apart when it is under tension.

Solution: Don't use cracked parts!! (Well duh)

Don W.

Stuart & Kathryn Fields wrote:

Don: Reading in a text book on Fatigue design for aluminum structures,
there is at least one design technique that assumes that there are cracks
inherent in the material to start with that are not readily detectable by a
visual inspection. The design philosophy progresses from there. This whole
area of study is explaining why the time life specs on parts can be very
meaningful. It also can allude to the excessive safety factors that can be
used in determining time life specs.
It looks like I've got a lot more studying to do in this area.

Hi Stuart,

If you've ever looked at aluminum under a high
power microscope you have seen that it is not a
smooth material at all, but consists of "grains"
of material more or less mashed together. If it
is alloyed with copper, you can see the copper
grains around the edges of the aluminum grains.

Fatigue of material is an interesting subject. Do
you remember the problems introduced at the drive
shaft on the Rotorway 162F when people started
changing out the chain drives for belt drives?

The chain drives needed an oil bath which required
oil seals, and a housing all the way around the
chain. Since the only way to change the bottom
seal when it started leaking is to remove the
housing, and that requires removing the chain and
drive sprockets it seemed that replacing the chain
sprockets and chain with a gilmer belt and
tensioner was a good idea--at least until the
drive shaft from the motor started failing right
on the other side of the support bearing.

The problem was/is that the drive shaft is
supported by only one bearing below the drive
pulley, and the gilmer belt must be kept in much
more tension than a drive belt. Consequently the
end of the drive shaft is bent ever so slightly by
the belt tension. This would be okay except that
when it rotates, the direction of the bend keeps
changing. The effect is like bending a piece of
metal back and forth 50 times per second. Even
the fairly thick drive shaft failed in a hundred
hours or so given that treatment. The failures
puzzled everyone at first, and they called it
"fretting fatigue"--I guess because the designers
were fretting over it ;-)

Don W.