Scott Speirer wrote:

With regard to Perry Mick's tension bolt failure, I highly doubt that coating the new
or original tension bolts during the rebuild is going to prevent a recurrence of a
failure.  You might as well cover them with bovine scatal matter before installation
- the amount of protection will be the same.  I strongly suspect that Perry will have
another tension bolt failure with his rebuilt engine around the same number of flight
hours as for the first occurrence.

If the bolts were fine with 130,000 miles while in the car, and one failed after a
proportionately much smaller number hours in the airplane, it must have to do with
the airplane environment, not the age of the engine.  After reading the posts
regarding this incident, I don't recall anyone discussing vibration very much.  I
suspect that the engine, direct-drive and propeller - as a powerplant system - has a
somewhat different set of vibration characteristics than the engine had while in the
car.  It is likely that in the vibration spectrum for the powerplant there is a
component which matches the natural resonant frequency(ies) of the tension bolts,
while they are under tension in the installed condition.

What I think would be a prudent course of action would be to have a spectral
vibration analysis done on the powerplant, once the engine is rebuilt and reinstalled
in the airplane.  This would involve the use of a multi-channel spectrum analyzer and
a set of vibration sensors.  The vibrations in a system like this can be understood
as being composed of set of sinusoidal oscillations of various frequencies.  The
amplititude of the oscillation at each frequency can be seen clearly with a spectrum
analyzer.  The engine would be run at a number of different rpm's, and especially
those ones most used during cruise, climb and descent.  The vibration sensors are
placed on the engine, including some of the tension bolts, and especially the tension
bolt located where the last one failed. A university engineering department may be
willing to take on the analysis at little or no cost.

If the results show that there are just one or two specific rpms where the tension
bolt vibration is high, you can just fly avoiding those rpms.  No modifications would
be necessary.  If the vibration is high throughout the operating ranges for cruise,
climb and descent, you've got a problem.  You'll have to figure out a way to
significantly increase or decrease the natural resonant frequency of the bolts,
and/or modify the composition of the propeller-to-engine drive train to remove those
frequencies with some kind of dampening device.

Scott Speirer

My thoughts exactly Scott and I think I have ARP talked into doing just that.
Mazda, for some strange reason, threaded the bolt much more than it needs
to be. The engineer at ARP was surprised and mystified when he saw it. 
When ARP make bolts like that there are no threads beyond the threaded nut or 
hole they are screwed into. The bolt fatigued at the root of the last thread
in the front cover.

Tension bolt failures are not unknown in car racing so the guys came up with
coating them with RTV. It seems to work. It restrains the bolt from large
amplitude vibrations all along its length reducing the stress.

Paul Lamar
 
The AirCraft Rotary Engine NewsLetter.              Powered by Linux.

http://home.earthlink.net/~rotaryeng/            http://www.linux.org