Aerospace Vehicle Fabrication
Satellite Launching Rockets
A major aerospace manufacturer was in the process of making the
prototypes of a new series of rockets. There rockets would be tasked to
launch larger and heavier satellites. The computer controlled press brake
used to bend the heavy, flat sheets of material into conical nose cones
malfunctioned. Instead of coming straight down, one side of the press
was going much faster. This action caused some heads of the large fasteners
holding the machine together to violently shoot more than 30' (9m)
across the room.
There were none, although two people from the plant that were working
near the machine were almost hit.
The equipment was heavily damaged, as enormous pressure was involved
the operation to bend the plates. The machine was repairable, but valuable
time was lost towards meeting the target launch dates for the rockets.
There was a 2-1/2" (65mm) diameter dowel pin used as a
between the main plates. It was cleanly sheared. The fasteners were 2"
(50mm) x 7-3/4" (197mm) specially made socket cap screws. The screws
appeared to have the smaller head diameter of the 1936 Series standards,
and the fillet radius was undercut into the bottom of the head. The screws
also had a reduced body diameter adjacent to the fillet. Tooling marks
made it obvious that these fasteners were machined from a bar, and that
the hex socket was made via the "drill & broach"
rather than forged. Additionally, the clearance holes for the screws seemed
Laboratory testing of the dowel pin showed the material to be 1117
that had not been heat treated. The screws did have the head diameter
of the small 1936 series, and we learned that the undercutting of the
fillet radius was intended to maximize the head contact area. Oddly, there
was ample room available for the current, larger 1960 series head.
of the reduced body under the head was smaller than the
diameter of the threads. Lastly, a normal clearance hole
size is 1/16"
(1.6mm) greater than the body diameter. However, in this
through hole was 1/8" (3.2mm) larger for the first
under the head, and then enlarged yet an additional 1/8"
for the next 3" (75mm).
This collection of linking faults could have seriously hurt someone.
the computer malfunctioned and skewed the load. Then a soft dowel pin
allowed a load change breaking the screws — made much easier by
having the cross-sectional area under the head less then the stress area
of the threads. This shifted the loads to the weak fillet. The oversized
holes then created a fulcrum point, and that facilitated the prying action
that broke the heads off.
First, get the computer fixed. Next, the dowel pin should have been
from a better material and through hardened. A standard 1960 series socket
screw design would have worked fine — and with a normal fillet radius.
A chamfer at the top of the hole would protect the fillet, and the 1/16"
(1.6mm) normal clearance hole could have been used.
As appeared in Fastener Technology