Jim
The torque forces from the tiller are pretty low in engineering terms. You
don't have to match the strength of the shaft, you only have to match the
torque the tiller and the person on the other end can develop. And you can
only pull so much before you fall over ... You can develop much more
torque from a wheel because of pulley ratios, also because you can pull a
spoke up wit ha lot more force than you can pull a tiller sideways. Don't
know how long your tiller is going to be, or it's dimensions, but I'd bet
any coupling that looked like it was big enough would be plenty oversized
when looked at in engineering terms. Whist I am with Steve in that screws
have their issues, it's possible to design them out. There is a thing
called a Ringfedder ( or something similar) coupling which requires no
keyway in either shaft, and will take all the torque and end loads you could
possibly throw at it with a person on the end of a couple of feet of wood.
They are not cheap, but neither is one-off machining of a split coupling if
you have to pay for it. I had one made recently to replace my engine to
output shaft coupling, the "real"price was around $600 Aust. It cost me
about half that, but there was no material or profit margin in that figure.
I'd try to work out the loads and shaft sizes, and then start talking to
someone in industrial transmission supplies. If you can figure out how to
do it with a commercially made item, you will have some guarantee that it's
appropriately sized.

I'm not a mechanical engineer so I won't do the calcs for you, but I know
the principles and can point you to some relevant suppliers in Melbourne.
David Flew


"jimthom" wrote in message
om...
Thanks again for your ideas.

The problem is really what type of coupling to use to resist the
torque forces of the tiller. If I could use flanges, the total
x-section area of the bolts that hold the flanges together should be
the equivalent to the x-section area of the rudder post, so there is
the potential for a lot of torque. As you said, splines may losen and
wear, set screws may wonder, and even the bolting method that you
describe may not be stong enough. Once a little movement occurs there
is accelerated wear and eventual failure.

I did a search for mechanical couplings and found a wealth of
resources; now it is a matter of sifting through them to see what I
can come up with. There is even have one device that resists torque
forces by heat-shrinking the coupling to the shaft!

Cheers, Jim

On Wed, 14 Apr 2004 10:06:21 -0700, "Evan Gatehouse"
wrote:


"jimthom" wrote in message
. com...
Thanks again for your ideas.

The problem is really what type of coupling to use to resist the
torque forces of the tiller. If I could use flanges, the total
x-section area of the bolts that hold the flanges together should be
the equivalent to the x-section area of the rudder post, so there is
the potential for a lot of torque. As you said, splines may losen and

Uh I don't think so. The polar moment of inertia should be the same
shouldn't it rather than cross sectional area?

So a few little bolts on a very large diameter flange would have the same J
as a rudder stock that is very close to centerline.

Evan Gatehouse

Evan's point should not be hard to grasp in the context of a
tiller/rudder post. A determined push from Jimthom's hand on a three
foot tiller could wring off a one inch rudder post entirely.
Think of it as a 3 foot wrench on a big bolt. Ah, the power of the
lever.
But in this thread, the traditional fixing for the bicycle crank keeps
coming to mind. The old-fashioned pedal crank fixing seems to have no
trouble handling a repetitious, cyclic load of up to 240 pounds at an
arm of 10 inches, 60 rpm for 20 years plus on a 3/4 inch shaft.
All done with smooth round holes, and a single flat worked on the
shaft.....
The cotter bolt is what works the magic: its tapered flat of less than
1/12 slope will hold the shaft tight, even without its fixing nut and
lock washer....

Brian Whatcott Altus OK

Thanks for your offer, David. I'm confident that yours and other
contributors to this thread have put the problem in perspective for me
and I can go forward from here. Cheers, jimthom