• We're building a go-kart. That can strafe.
    12 replies, posted
Hey guys, So I've been posting a bit about this in WAYWO but I figured I should make a thread to chronicle my progress. [b]What am I doing?[/b] I, along with my friend Brandon, are creating an electrical go-kart. We are using four 2.5'' CIM motors that run off 12V DC. If power allows we may use EIGHT of these bad boys. [thumb]http://puu.sh/ggFgw/82074b367a.jpg[/thumb] A CIM Motor We will be using Mecanum wheels, which are specially designed wheels that have a net vector in the diagonal. They are placed in a diagonal configuration -- that is, the top left and bottom right wheels have the same direction, and the same pairing for the top right and bottom left. When all four wheels go forward or backward, the sideways vectors cancel and allow for pure "vertical" (forward/backward) motion. However, when the diagonal pairings are run differently, the vertical vectors cancel and it allows for [b]pure side-to-side motion.[/b] You can use a mix of this to, for instance, move diagonally. [thumb]http://puu.sh/ggFsf/47f72ff46f.jpg[/thumb] A Mecanum Wheel. Finally, we're using a Raspberry Pi and an Xbox 360 controller to actually drive the motors. The motors can be driven by connecting them to a motor controller, which takes a signal from the Pi and allows a variable amount of current through. For this specific instance, we're going to be using Talon SRX Motor Controllers, which allow for a large variety of input signals. We're going to use PWM (Pulse-Width Modulation) signals to control motor speed. [b]Mark I Gokart[/b] Brandon and I have already created a Gokart using Mecanum wheels in the past week -- however, this was done by ripping apart a FIRST Robot Chassis and crudely sticking a chair on top. This had a number of flaws -- it was uncomfortable, and it required thousands of dollars in proprietary computer equipment. When we make our new gokart, it will have a custom design, with a non-proprietary Raspberry Pi. The Pi costs ~$35, while the technology on the previous kart (a cRIO, Laptop Computer, and Digital Sidecar) are combined MUCH larger than the pi, require much more energy, and cost [b]at least $1,000 more[/b] than the pi. A video of the mark 1 (Brandon is driving): [vid]http://puu.sh/gcVZF/1c36e8931e.mp4[/vid] [b]Current Progress[/b] I'll be updating this with the current progress and keep a history of all progress posts. Currently, I have linked a raspberry pi to a talon to drive a motor. It uses an Xbox 360 as input to drive the motor forwards or backwards a variable speed. Unfortunately, the PWM signal is by definition VERY timing-specific. The Raspberry Pi does not send PWM signals using hardware, only software; which means that any PWM signal will be very unstable, as shown in the video below: [vid]http://a.pomf.se/juewjv.mp4[/vid] As you can see, the motor jitters consistently unless being driven fully forward or fully backward. This is completely unacceptable. Luckily, there is a solution. I'm purchasing an addon board that sits atop the Pi and provides hardware PWM outputs; this allows me to send a completely stable PWM signal out for complete motor control. That's all for now -- if you have any questions or comments, I'd love to field them.
What's the potential ceiling for speed on something like this? I mean, if you hypothetically had an incredibly powerful motor, would you be able to stabilize the whole thing enough to prevent it from flipping when you move from a pure vertical to going sideways?
That's actually really cool. Looking forward to seeing where this goes.
[QUOTE=Mr. Bleak;47232130]What's the potential ceiling for speed on something like this? I mean, if you hypothetically had an incredibly powerful motor, would you be able to stabilize the whole thing enough to prevent it from flipping when you move from a pure vertical to going sideways?[/QUOTE] Going sideways is going to be slower than going forward, but the speed ceiling is whatever the motors can output and the battery can supply. I estimate for our new design a max of, eh, 20 mph if we can make it light as possible and tune the gearboxes correctly. It essentially will be impossible to flip when you do what you describe. In fact, it's impossible to flip on any flat terrain no matter what you do. It just doesn't have the speed/decelaration necessary to flip its VERY low center of gravity. I'm actually a software engineer with a small amount of electrical engineering experience (which is growing as I do what I'm doing here). Brandon is the mechanical engineer so he'd know better than I would. [QUOTE=woolio1;47232134]That's actually really cool. Looking forward to seeing where this goes.[/QUOTE] As am I! Unfortunately I don't think we'll be able to make it street-legal but I do look forward to racing against the Mark I with it.
[QUOTE=Ehmmett;47232185]That looks like it's a blast to drive[/QUOTE] It's much better on a hard surface than on carpet, but unfortunately there's too much snow (aka melty water) outside to take it for a spin. On carpet when you try to strafe it kind of pivots because of different levels of friction on each wheel -- on a hard surface it's pretty much a 1:1 movement and it is, indeed, a [b]blast[/b]. I kind of want to save this until I have kids and then build more so they can race each other (putting speed dampers on them of course). That'd be awesome.
Call it the GeelMobeel. Spelling is on purpose.
[QUOTE=NightmareX91;47232311]Call it the GeelMobeel. Spelling is on purpose.[/QUOTE] It'll be the reel geel deel.
Why did you choose brushed motors?
[QUOTE=OogalaBoogal;47233207]Why did you choose brushed motors?[/QUOTE] It's piss easy to design a solid & cheap H bridge driver for a brushed DC motor. Meanwhile it's going to be either a headache or fucking expensive with a brushless 3 phase AC motor. That said a 3 phase AC motor is both more lightweight, smaller and packs a harder punch.
I pretty much do all my shit with brushless. The headaches are very real, but usually only while starting out. Once you get it going it's very much worth it.
[QUOTE=OogalaBoogal;47233344]I pretty much do all my shit with brushless. The headaches are very real, but usually only while starting out. Once you get it going it's very much worth it.[/QUOTE] Problem with AC brushless motors is that you need a feedback loop, which usually consist of a simple hardware, but intricate programming solution, or complex hardware solution and simple programming. Most common solution is current sensing and encoder wheel + optical sensor, for the 'better safe than sorry' factor. Only brushless I voluntarily deal with are steppers, and that's because I can "just" use the 'bigger is better' methodology when in any kind of doubt.
[QUOTE=OogalaBoogal;47233207]Why did you choose brushed motors?[/QUOTE] It's what I'm used to and what I have.
Sorry, you need to Log In to post a reply to this thread.