VEX Starstruck Single-motor dumper experiment

This was an experiment to see if you could build a Starstruck robot that will score stars, cubes, and perform high hanging all with the power of a single motor… Actually there are total of three motors: two powering non-symmetrical holonomic drive (it just turned out this way) and a third motor powering linear lift and dumper (using a differential):

From the power standpoint - the experiment was a success. One motor could easily lift both game objects to 25" height as well as performs high hanging. While nobody expects it to be fast, it is not very slow either.

There are, obviously, numerous problems. While robot could decently handle stars, it is difficult to keep cubes from falling off the forks. The frame is not well balanced and, with the total weight of just 4.5 lbs, back wheels have difficulty generating enough traction when carrying a cube. Aligning for hanging didn’t have to be this difficult, etc… In any case, the time would be better spent improving designs that have access to more power. :slight_smile:

The main highlight of this robot is differential:

IMG-20160908-00092-diff1.jpg

It splits the power (from the force geared motor) between linear lifting and flipping (dumping) motion. Total spread between two outputs of the differential is 9x. Lift is geared up 1:3 to be faster and the arm is geared down 3:1 for the higher torque that is required for the flipping.

Essentially, motor is mounted on the arm and has 12T pinion connected to its output shaft. With the pinion it tries to climb onto the 36T gear. If there is no load on the arm it will flip first, hit backstop, and then 36T gear will start rotating and, in turn, driving 12T/36T compound gear that interfaces rack gears.

If there is a game object on the arm then flipping would require more torque and it will be easier to start turning 36T gear first which will move the carriage up. Once it reaches the top and hits backstop then flipping will be the only available motion.

There are several types of differentials that could be used not only to split power from a single motor into two destinations, but also to combine the power from two independent sets of motor (driving separate motions) when high torque is required.

Finally, while this robot was never intended to go to a competition, here is a proof that it could fit into 18" cube, in case anybody would doubt it :slight_smile:

IMG-20160908-00093-folded.jpg

1 Like

i am in awe

1 Like

Is that front omni wheel cut in half?

Wow! That’s really cool, especially considering that with only 3 motors on the robot, you could make a tether bot and have essentially 4 robots instead of one.

1 Like

A true minimalist robot! Excellent!

1 Like

@vex2.0 and @Team80_Giraffes, thanks for the comments!

Yes, there was an extra omni and it was tempting to see how it would work. The conclusion is that you wouldn’t want to use it for anything except an idle wheel carrying very little load. It loses a lot of the firmness and could be easily bent.

If you have to cut a wheel the better choice would be a thicker 2.75" omni like on 185A NbN robot (not easy to see but they have a pair of those around the intake).

Yeah, you could, probably, fit two of those in 18" cube! However, the expectation is that a single robot with 12 motors will be much more effective.

if you could pick and score a cube in 20 sec and a star in 10 sec with just 1 lift and 2 drive motors, then 4 motor lift and 4 motor drive should be able to do that in 5 and 3 sec respectively and, hopefully, even faster.

1 Like

Holy crap that differential is ingenious. I’ve seen the concept happen numerous times when I first started building in vex, but always thought of it as a design flaw and a mounting of the motor in a bad spot. Way to go dude, freaking god.

1 Like

Please tell me all of the parts you used on this robot (c-channel, gears, screws, etc.) and create detailed instructions so I can make an exact clone of your robot.

Joking, just joking. Seriously though, nice robot. It reminds me of The Little Robot That Could from NBN. I don’t think this one will qualify for worlds, though. Even so, good job.

1 Like

I thought differential gears don’t combine torque; their torque is equivalent to that of the input with the least torque.

VERY interesting project. It’s amazing how elegant even the simplest designs can be.

1 Like

Yes, this is a good observation, and I should have used word “power” third time in that sentence, because that is what I was thinking about. But first, lets get a little background info on differentials.

In general, any mechanical system with three input/output channels is a differential if sum of the changes over all channels is guaranteed to be zero.

Let say you have a black box with three shafts x, y, and z coming out; then it is a differential if it is always true that:


 A dx + B dy + C dz = 0 

, where A,B,and C are some constants.

For example, classical automotive differential (with wheel axles connected to x and y, and motor power coming through z) has A=B=1/2 and C=-1:


(dx+dy)/2=dz

. Essentially, average speed of the wheels always equals to the output from the motor. This allows outer wheel to rotate faster than inner wheel when vehicle makes a turn:

Simple automotive differential works well when neither of the wheels loses traction with the ground. There are now all sorts of advanced designs that limit wheel slip, control torque distribution, etc…

Now, back to the question about combining the torques. Let say you have two input channels x and y and an output channel z. Also, let say you moved all up/down gearing outside of the differential, such that:


dz = dx+dy

. This case could be illustrated by this “great” example of forklift safety violation:

SAFETY2.JPG

Yes, you are correct that large load force could backdrive the weaker input. However, we still add output power of two sources.

If you pretend that weight of the higher forklift is negligible and both forklifts generate equal lifting force F then the maximum weight they could lift will correspond to F (or the weaker of two, if not equal). However, the load will move up faster because you add up lifting velocities. Therefore you effectively combine power of both forklifts: P=F*(v1+v2)

If your VEX robot has two stages: first set of motors attached to the base and the second set of motors riding the arm (that could have an independent motion) then you are also combining their power.

I’ve seen robots in previous years that have successfully used this principle and, if you are careful how you do it, you could cross link those stages (to get some power flowing) and get very interesting results.

With Startstruck game play calling for three distinct motions with slightly different force, direction, and power requirements (launch stars, dump stars or cubes, and lift the robot) it feels that some sort of differential could offer competitive advantage by using motors more effectively.

1 Like

A differential does work when the power from the inputs are always the same. However if the power of the inputs are always the same, why don’t just link the two power sources together?

It’s true that you have to balance forces between differential inputs, however the power flowing from one input could be greater than that coming from the other (see the forklift example above). Also consider non-symmetrical differentials that could be implemented with planetary gears…

For various reasons, I don’t think it is a good idea to share at this time a design that my team is trying to implement, but if you have seen it it would be much easier to understand what I am trying to explain. We will reveal it once it is built and tested (assuming that it actually works). However, we are not such a fast builders as 8000A, so it may take a while. In the mean time, I will try to give an example of a similar concept:

Let say you have built a multi-object dumper on one side of the robot and a catapult on the other. They are powered by 4 motors each. The way they are built there is no way to directly re-use either of them for hanging and, with 4 motor drive, you don’t have any motors left.

Now imagine that you somehow attached one end of the string to a special point on the catapult and another end to the dumper. Then there is a pulley with a hook that could easily move up and down as you pull or release the rope. When there is no load on the hook (z) both catapult (x) and dumper (y) could move freely and independently. Position of the hook (z) will somehow correspond to the sum of x and y: z=x+y

Assuming that you picked locations to attach the string such that both (x) and (y) will provide matching forces you will now have robot lifting capability powered by 8 motors with very minimal interference with original functionality of either dumper or catapult.

This is, obviously, not what we are trying to build but, I hope, you get an idea of how you could have two independent mechanisms (motions), each powered by its own set of motors, and then you will be able to combine the power of all motors to do something else.

1 Like