"Catastrophic" failure of mountain bike fork lowers (Manitou Skareb Comp)

The magnesium lowers of my 2003 Manitou Skareb Comp had begun to crack last
year, about 6 months before I stopped riding it. I've kept it stashed away
for a while until today, when I was cleaning my room and found the lowers.
I gave the legs a light (maybe 2.5-4lb) squeeze, and they snapped. Although
I didn't mean to snap them, they lent some interesting insight into their
failure.

Necessary background:
Manitou hailed their reverse arch technology in 2003 as being revolutionary
because it allowed lighter forks, stiffer blah blah blah. The Skareb, Six,
and Axel were the forks I can remember off the top of my head as having
identical lowers. These bikes for 2003 were assembled and shipped with the
fork/stem turned backwards, which, traditionally, puts the arch towards the
back of the bike. But with the reverse arch, the arch faces forwards, and a
bike's weight is all put directly on the arch during shipping. This
stresses the fork in a manner that it was never meant to be stressed
(perpendicular to the axis of the stresses normally encountered on a MTB
fork). As a result, as these bikes were ridden, the lowers began their
downward spiral into separation.

My experience:
From the first day I bought the bike, the front wheel had always rubbed the
v-brake pads when turning. I thought this was just because it was a light
bike (23lb Giant XTC2). But as I continued to ride, my expectations turned
to frustration, as the larger tire that I had put on was actually beginning
to rub the arch when braking hard. This wasn't really a problem until I
began riding more aggressively. It was at that point that noticed that the
wheel was beginning to get cocked to the side with the mere application of
weight. I knew this was out of the norm and searched for the problem, which
I found in the form of a crack in the fork leg that contained the coil
spring. New updated lowers with a beefed-up arch/leg join were sent, and I
installed them without a hitch. After another 2 months of use, I broke the
damping assembly inside the fork. Rather than fix it or pay to have it
fixed (I wasn't working at an LBS at the time) I left it as it was,
spring-only. After another 3 months, I acquired my current bike (Giant
STP2) and a Fox Vanilla. The Skareb with the updated, uncracked, flexy
lowers was put aside.

Today:
I broke the lowers manually and was surprised to see what had transpired.
The crack was much more extensive than I had thought, making an interesting
path parallel to the casting wall.

Photos: (I'm a much better photographer than I am a rider)

Here it is a year ago when I first identified the crack:
http://plaza.ufl.edu/phillee/rbt/p1.jpg
http://plaza.ufl.edu/phillee/rbt/p2.jpg

Here is the upper half, the arch portion:
http://plaza.ufl.edu/phillee/rbt/u1.jpg
http://plaza.ufl.edu/phillee/rbt/u2.jpg

Here is the lower half, the leg portion:
http://plaza.ufl.edu/phillee/rbt/l1.jpg

Here are the component parts, for reference:
http://plaza.ufl.edu/phillee/rbt/o1.jpg

(note: shiny marks on the pointy parts of the cracks are due to my
matching-up of the two broken pieces, which, I assume, smoothed over the
apexes of broken edges.)

Analysis:
Dark corrosion shows fatigue cracks weaving its way through the magnesium
(if it really is magnesium) up to a certain point about 2mm away from the
surface. From that point, the cracks then take a 60-degree turn upward
towards the arch. This suggests that at least some of the stress
experienced by the lowers were shear stresses (45 degrees if fully shear).
The cracks actually occupy a rather large portion of the total cross
section. Failure would have been imminent had I continued to ride it. The
final fracture indicates brittle material, with a smooth sandpaper-like
surface. Beach/clamshell marks are not very prominent, but they are there.

Corrosion is also visible at the very inside edge (the portion with the
smaller radius) of the casting. This is particularly worrisome, as it
appears that the fork had begun to fail not just at the thick portion, but
at the thin portion too. If this were the case, failure would occur even
faster since the forces acting along the thin part of the cross-section
would guarantee high bending moments, and thus high tensile stresses. If
they were acting along the long part of the cross-section, the forces at the
ends of the cross-section would not be as high.

The location of the inside edge corrosion also suggests that the casting was
improperly radiused/chamfered, at least in the original design. These sharp
corners should be avoided.

Final word: I'm lucky these didn't totally fail on me while I was riding it.

--
Phil, Squid-in-Training