Michael Perrone
Overview:
We purchased two neodymium ring magnets and forced them together with a 3D printed casing that was secured with a bolt.
This assembly was tied to fishing wire and suspended from a pulley system, on the other side of the pulley we have a plastic bottle filled with water that acts as a counterweight. We repeated the test at least 50 times in each arrangement while recording with a camera.
A python script analyzes the video and detects the position of the assembly in each frame which is then fit to a parabola versus time for that drop to get the acceleration. Code found here.
The external assembly is made entirely of non conductive or magnetic materials and the experiments were done in a parking lot where a metal detector was used to verify that no metal objects were nearby. The bolt securing the magnets is conductive and may be necessary for the effect, it does not move relative to the magnets and therefore makes no effect on the acceleration of the assembly according to existing understanding.
Results:
We found that the attracting magnet assembly had an acceleration 10.5% greater than the repelling magnet assembly. A T-test indicates a 16-sigma statistical difference between the two datasets. We are currently investigating if this could be due to systematic error in our measurement process or a fundamental change in gravitational attraction. Data and analysis procedures found on github.
Possible Sources of Error:
There are many possible sources of error that could have an effect on our results:
- Mechanical
- The camera moved between attracting and opposing mode trials and a bit throughout the trials.
- The stand was adjusted between attracting and opposing mode and a bit throughout the trials.
- The bottle placement at the start of the run has a variety of horizontal and vertical offset between trials.
- The fishing line was released by hand, there may be some slippage before full release of the line.
- Kinetic friction of bearing can drift with rope angle, temperature, lubrication, and load.
- Wobbling by the magnet, bottle or adjustment of the stand can affect this.
- The fishing wire may slightly loosen over time changing the force on the bearings or initial height.
- Environmental
- Wind can apply forces on the assembly affecting results.
- Nearby conductive or magnetic items around can cause errors.
- The earth's magnetic field has an effect on the assembly.
- Other
- The code does not always properly identify where the center of the object is.
- Could be affected by camera angle, magnet rotation or shadows
- Small relative motion of the magnets relative to the bolt can cause eddy currents to form slowing down the fall.
- The Bolt is not homogenous material and magnetizes in some directions easier than others. This may cause instability and wobbling.
- The code does not always properly identify where the center of the object is.
Next Steps:
To improve on this experiment we would redo the trials with the following modifications:
- Repeat the trials with the camera angle and position of the stand constant throughout.
- Swap between repelling and attracting mode multiple times to account for any systematic errors that change over time.
- Use a clean background to improve object tracking.
- Track the bottle as well as the magnet to add redundancy and improved accuracy.
- Use a control object with similar weight but no magnets.
- Repeat measurements with a nylon bolt instead of metal.
Comparison With Other Experiments:
The Action Lab: Found no difference in the acceleration of objects in his replication, he did not have a metal bolt traveling with the magnets suggesting that may be critical. Note his test was conducted under vacuum with weaker magnets and a short drop distance.
Robert Francis: Also found that attracting mode magnets fell faster. He used data from an accelerometer that was very close to the magnets being dropped and would likely be attracted to them, this could have affected his results.
Trial Videos:
Read more »
Voja Antonic
lumbric
Mike Rigsby
Lars Friedrich