pulling eight arms

Testing with eight arms. It works, however the PSU I have don't have enough amperage to pull all eight at once. Currently only seven motors are able to pull the arms.

Connecting the arms

I am struggling with connecting all the arms in the centre. I don't trust the objects to be sturdy enough to just glue them to the tubes. I think this could work if i glued the first layer of each arm to the tube and glued the tubes to the centre object. I guess the object could just be resting inside the frame then without having to tie it up. However i feel like it doesn't really work that well. Maybe i should remove the tubes all together?

Here they are just loosely threaded on.

David suggested I solder the brass tubes together instead as it will be much stronger.

workflow

Cutting all the layers for the arms. Then glueing the spacers on to separate the layers from each other. Without the spacer the angle of the thread will be too steep, so I think there is a chance that when it is stretched the layers will group together instead of spreading out evenly.

Then cutting the brass tube into 3.6 - 4mm ish length and gluing it onto the layer. When this is done the layers can be grouped together ranging from 0 - 20.

The tube cutter adds a bevel to the edge so each layer also needs to be drilled open again. In hindsight i realized that the sanding machine would've made this much easier and quicker.

Process for the arms.

experiment bigger brass tube

Tried with bigger brass tubes. Outer tube has some extra space around it to allow for some error. Inner tube is also slightly bigger, hopefully it won't bend as easily as the previous one. It works much better. However with the extra space around, the layers don't stay as straight as the other one. But it stretches pretty nicely. I think I should try to make the distance between each layer smaller as currently it strethes almost too much. If the layers are more compact the difference between the original object and the stretched one will be more clear.

Brass tube experiment

Tried to drill distortion from the pipe cutter away. It works pretty well. However when threading all the layers it does add some resistance. I am not sure if this is because the inner tube might be distorted or because it is not pulled at the right angle. The layers stay straight, but the distance between the layers is still a problem. Will try with more flexible glue.

knots between layers

I can't cut the brass tube correctly. There's always some distortion that either makes it impossible to slide it on to the smaller tube or it adds way too much friction for the motor to slide it easily.

This is with knots between each layer. The knots work, they just take a lot of time and they also remove a lot of the elasticity of the thread.

acrylic and latex

Tried with both acetone and chloroform to degrease the surface / make it rougher but neither worked, the latex doesn't stick properly.

One problem could be that the diameter of the thread decreases when you stretch it. There is a lot of extra room for the thread when it is stretched out fully.

object alternatives

Different alternatives for the object.

experiment elastic thread + glue

Tried to superglue the layers to the thread but I couldn't make it work properly unfortunately. The thread lost all elasticity and got stuck in a semi stretched state. I think i need to find another way to keep the layers straight.

If i use a tube i could glue them together, but then i can't stretch it anymore.

experiment elastic thread

Tried making one arm with elastic thread, based on the model from yesteday. 2mm ish acrylic with spacers, 20 layers. The market was closed by the time i got there, so will try to go again today to see if they have any other material.

It is pretty flexible and there's not too much resistance, however i think maybe rubber bands of the same size could be better? Also I think it could be possible to limit it to three threads instead of four as well to reduce resistance even more.

communication and power

Currently it just chooses a random servo and a random angle between 0 and 180. It goes back and forth like a a conversation, but I think it is not that clear.

I think I should make it so that it always chooses an angle that is far away from its current position (fixed this, just not shown in the video). Alternatively maybe do a "Simon says" version instead, where object B just copies A everytime? Or a version where they both copy each other but add on to the loop. E.g. first A sends [180], then B sends [180, 45], A sends [180, 45, 160] etc.

Or if it should just be an off - on situation?

Only did four servos this time, should do a version where i use more servos as well.

Testing the power source with eight servos running simultaneously.

Different alternatives for the spool. Tried to match the rest of the object a little more with these.

ideas for new object

A few variations for the object. Still have to make a spiky one.

If i place the object in a 45 degree angle inside the tank i can control each end of the obcect as well as the middle part. The middle part can be pretty much fully controlled as it will have 4 motors pulling in different directions. But for the ends to work, the motors need to be paired, e.g. the combined value of the motors can't exceed 180 degrees -> if one motor is turned 120 degrees, the other needs to be turned 60 or less (for the object to be able to move up and down without the thread becoming loose enough to jump off the bearing. However, there may be a lot of water friction if I use complete discs?

Instead of pull each arm outwards. It will pull the arm in the opposite corner of the motor to turn / twist the arm. Similar to the kinetic snake arms. But i think this one may be impossible as the arm closest to the motor will collide with the thread.

Laser cut the base shape and connect it together with gussets. Then connect the "spikes" to the base shape with elastic band. The layers need to be pretty tightly knit together for the shape to bounce back to its original shape.

This one is a combination of an acrylic cage and silicon. With a cage like this the silicon can be attached on the inside. Would also only need a 2D form.

Another similar to the first one, just in a circle instead. I think this one has less options in terms of possible distortion. As its a circle the distortion will in way be mirrored. Also if i use rubber bands / flexible bands to keep it together, there is no way it will retain its shape. Will end up in a square.

The rubber bands used here are way too strong. It is very stretchy but requires too much strength for the servos to be able to pull it. The model itself is also way too big.

lighting 01

LED chip test. This is a single LED chip placed 9.5 ish cm away from the tank. This is more or less the same as the LED strips, just a different pattern / distance. Would of course use just the LED chip and not the case around it. Found this laying around in 504.

The light is much warmer than the LED strip. It also has a pretty distinct hotspot, but i think this may be becuase of the case and not the light itself. The box that keeps the tank up is also casting a shadow which obstructs the light.

power use

If I keep the distance between each strip to around 3cm the light is pretty diffused on the acrylic when they're placed 2cm away (thickness of the steel frame). Each LED needs 60mA, which totals to 6.24A with 8 strips on the bottom plate. I think if the distance from the tank is more than 2cm (around 5-6) i could decrase the amount LEDs by alot.

The motors need a running current of 900mA and a start current at 2.5A. So i need at least 7.2A / 20A to power 8 servos.

The micro control unit needs at maximum 800mA.

So in total i will need at least 27.24A to power all the lights, servos and the MCU which is a lot of current (this can't be right can it?). In this case I think i should seperate the ligts from the servos / MCU so that maybe i could use one 5V 15A for the servos and one 5V 10A for the lights? Alternatively power the lights with something like this instead?

A standard computer power cord is usually rated for 15A so i could maybe use one for each PSU? Then find the correct AWG size cable to wire from the PSU up to the servos. I feel like my calculations here are way wayyy off. Can a wall socket even supply this much since i need double to power both objects?