Early Brainstorming
3D Model Tolerance Test
For tight fit, give a tolerance of 0.2mm.
Design 1: Pure Magnetic Suspension Design
Concept Art & Block Diagram
Week 7:
3D Models
Print Result 1
Forgot 0.2mm is for TIGHT fit but we need it to slide. Decided to test 0.4mm and 0.6mm.
Both fit and slides down but not smoothly. WD40 has limited success. Would be nice to smoothen the walls but sanding of 3D printed parts not allowed due to prevent microplastic production. Will try 0.6mm for now for proof of concept.
Turns out the magnets are separated too far apart, so the magnets on top fall into the gap between the magnets at the bottom.
Diagram above:
Ideal configuration (top): magnets on top (red) stays repelled above the magnets below (white)
Actual (below): magnets on top (red) falls into gap betwen magnets below (white)
Other observations:
Fooling around with other things but getting close???
Week 8:
The magnets are rearranged so that they repel without the need to spin. The new arrangement simulates a ring magnet more closely.
From the previous week, we saw that the distance between the floating platform and the base is not very large, but by adding a stronger layer of magnets in between to repel both, we can increase this distance. However, whether the layers of magnets in between or the strength of the magnets in between is more significant would require more testing.
Another problem more relevant to the development process rather than the product itself is making the holder for the stronger magnet is difficult since it is difficult to put the magnets in and take them out if needed (see above video, the middle layer of magnets is difficult to put into its holder (black) because it is a tight fit)
A second holder that has thinner walls is made so that it can bend to make space for the magnets to be put in easily. Its flexible property also squeezes the magnet in place to keep it from falling out. It is also easier to push the magnets out since the holder can bend open wider.
Next we investigated the effect of a moving magnet in a solenoid. We wrapped an insulated copper wire around a tube and dropped a magnet through it.
The falling magnet generates an e.m.f. in the solenoid according to Lenz’s law, but this value is very small and has no visible effect in opposing the magnet’s fall.
Week 9:
We realise that according to our previous physics education, if the magnet is fully inside the solenoid, there will be no current in the solenoid since the current generated on either side of the solenoid would cancel each other out resulting in no current. This is also why the current goes from positive to negative (when entering and leaving the solenoid) when we drop the magnet through the solenoid (2nd closest video above) and a positive AC reading when shaking it in and out of the solenoid (1st closest video above).
We decided to cut the solenoid in 2 and separate them a distance equal to the length of the magnet so that the magnet will always have at least one coil opposing its motion.
Update: no, it did not make a difference.
Week 10-12:
The increase in distance due to the middle layer of magnets is due to the greater strength of the middle layer of magnets and the increase in number of layers. If the magnets are stronger, we need less layers. For our case and practicality sake, we can simply replace the magnets in the base and the floating platform with stronger magnets to reduce number of parts and simplify everything while achieving a sastifactory height.
We also tried exploring using electromagnets instead of permanent magnets. We tried tying our own coils, but they weren’t very strong, even when we applied a large voltage across it.
In the vidoe above, around 0.16A was applied to the coils, and we can see that the neodymium magnets were only moved slightly. While we can technically apply a larger current, we do not foresee being able to do that with a conventional power source.
We built a small rolling platform to test our suspension system on. We plan to attach two motion sensors, one on the suspended platform and one on the base (unsuspended).
The motion sensor will measure the acceleration (reflects how much force is experienced) and the distance moved (to calculate transmissibility).
Sample result:
Sem 1
Decided to add a gimbal-like design so that the suspension system is always upright, so that the force experienced by the suspended body is purely linear and not rotational which will complicate our calculations.
Rough sketch:
Gimbal-like system requires an IMU, Arduino, and servo motors.
Testing our code:
Test 1: IMG_2137
Test 2: IMG_2241
Test 3 (ran into some issue with this one): IMG_2242
The servo motors came with accessories like these to attach to it:
We 3D printed a small attachment to attach the motor accessory to the suspended body.
Before assembling:
After assembling:
Update: the small 3D printed attachment is not strong enough. May need tiny screws instead.
Update 2: the attachment is not needed.
Afterwards, we realised that the weight of the suspended body is too large, which made the gimbal not function as well as it cannot hold the weight of the body. The induced magnetic force is also negligible, so we removed the magnets at the side and the aluminium can so reduce the weight.
At this point, there is no longer dampening of the suspended body’s movement, so we just focused on how much did the gimbal improve the original design.