"Allan Bennett" wrote in message
...
In article , riverman
wrote:

blade at all times. If all you are interested in is the resultant force,
put a potentiometer on the bow and brace it against a wall.

Unfortunately, like all the suggestions about tethering the boat, this
idea
misses the point by a mile: Kieran wants to measure the forces during a
paddling stroke - paddling against a resistance is just not the same.

--riverman (I love trying to sound like I know what I'm talking about)

...keep trying...


Allan Bennett
Not a fan of immovable objects

--

Well, okay but I was hoping that you'd come to this last point on your own.

You've got too damn many variables, Allan! You cannot run an experiment when
the variables include feather, fetch, grab, speed of stroke, blade depth,
variations of applied power, assault angle, retrieve distance and time,
stroke time, etc etc etc. Add to that a human doing the motions, and even
an isolated variable will have abberations. There is absolutely no way to
determine cause and effect if you cannot identify the role of a single
variable.

You need to isolate variables. Build a jig that will hold a blade and rotate
it in a circular motion. Place a paddle in the jig, and adjust the feather
angle, then let it wind out a few dozen times while you measure the pulling
effect on a rope tethering the boat. Change the feather angle, and go at it
again, until you have a 'feather angle vs forward force' graph. Change the
length of the paddle shaft until you have a 'blade depth vs. forward force'
graph. Change the rotational velocity until you have a 'stroke speed vs.
forward force' graph. Etc.

Then build a jig that will hold a paddle and move it horizontally with the
shaft vertical, lift it out and replace it a few feet forward. Maybe
something on a caterpillar tread. Place a paddle in this jig with no
feather, and run this several times and vary the feather variable until you
have results. Then change the speed, the length of the stroke, the angle of
the paddle, etc.

The build another jig that will do something else, and run a host of tests
on that.

When you are done, you need to solve each of these equations for the
representative curve, the K factor, and then meld them together into a
joint/inverse relationship equation that takes all the variables into
account with a single K.

And good luck!! IIWY, I'd identify 3 or 4 variables and call it a day. Just
think of all the minor adjustments a paddler makes within a single
stroke...and you want to quantify THAT?

--riverman