To pick up from the last blog post I made some changes and had some success.
The first change was to reverse the polarity of the upper pointing electromagnet. I did this by loosening the nuts that hold the second foam board to the first and turning around the electromagnet rather than reverse the wires. Now when the electromagnets are powered on they don't oppose each other. Now they will make the top of the device be inclined to move towards one of Earth's magnetic poles and the bottom two towards the opposite pole.
The second change was to add permanent magnets at a right angle to the foam board near the top. Now when the electromagnets are off the permanent magnets are inclined to point north/south with Earth's magnetic poles.
The third change was to alter the Arduino sketch (program) so that the electromagnets only go on and off several times a second instead of reversing polarity.
Now it is set up as an earlier model that I somewhat understand. The device will oscillate as the electromagnets and the permanent magnets interact with Earth's magnetic field. Since the point where they oscillate is not the center of the mass of the whole device and neither at the center of the point of magnetic torque from the magnets acting like a compass and rotating, the device will go in a line.
As can be seen in the below (time lapse 6x) video this works.
The red cardboard 3 fold boards are reducing the effect of air draft and you can see the device accelerate a little over the several minutes it is on. Note that it is only turned on after some period of time to establish that it is standing still in the water.
Below is a second trial. (Also at 6x time lapse.)
This worked several times but then a problem with what I am almost certain is static electric charge occurred.
Here is what happens.
The 10 inch aluminum pan is used specifically as opposed to Styrofoam. In previous experiments Styrofoam caused this problem but now it happened anyway. I am not really sure exactly how it works but the static prevents the device from moving in the water, even when deliberately physically pushed. It will go for a short time then come to a perfect standstill. (My experience is that in any of the previous experiments a small push would make the model drift quite a bit.)
The thrust generated by this device is not powerful enough to overcome the resistance to motion from the electrostatic charge.
One thing that seemed to work in the past was to empty the container of water completely and refill it. I am now waiting for warmer weather to empty the water and try again. Also it may be that the choice of foam board for the structure of the model was not the best one.
20180125 Thursday - I did some math and realized that turning the magnets on and off at short intervals is not as effective as doing so at longer intervals. (Not necessarily with this exact model but it explains why this model is weak.)
To get a centrifugal force from a pendulum the speed of the pendulum's motion is important. The faster it moves the more centrifugal force. To show that it will move faster in one longer motion than in many shorter motions we only need to show that any velocity will be greater in one continuous acceleration from 0 than many small accelerations each from 0 and to do that we need to do some Calculus. But is it pretty easy Calculus.
Velocity = Acceleration * Time
So for our example we will compare a 10 second continuous acceleration from 0 verses 10 1 second accelerations from 0 and show that the 10 second continuous acceleration amounts to much greater velocity.
Since Acceleration will be considered a constant we will remove it from the equation or just call it 1.
For 10 separate accelerations of 1 second each from 0 the definite integral is:
10 Sum Time from 0 to 1
equals:
10 Time^2/2 where Time is 1 or 10 * 1/2 = 5 in lets say inches per second.
For one continuous acceleration for 10 seconds it is:
1 Sum Time from 0 to 10
equals:
1 Time^2/2 where Time is 10 or 1 *100/2 = 50 inches per second.
The difference between 5 inches per second maximum speed over 10 seconds and 50 inches per second maximum speed over 10 seconds is enormous.
Now when we consider earlier experiments where the device rotates completely in a circle it makes sense that it is a much more efficient use of energy to produce linear motion since the acceleration takes place continuously for at least several rotations until a terminal speed is reached due to factors such as friction.
