Robots

Vex Robot

This project is a simple drive base for 5.8 GHz FPV around my yard. One goal was to spend as little money as I could for components- in the spirit of frugality, I substituted preferred materials or components for those that I could make work. The chassis, for example, is 1/2” square walnut stock that I found in my garage. It was a little warped, but it worked fine. I figured I could tweak the axle blocks to align the wheels rather than depend on the accuracy of the chassis. A delta mix in my 2.4 GHz transmitter allows me  to control the tank-drive steering. I currently have no further plans for the robot, but it might serve some purpose other than exploring with FPV in the future.

As usual, the chassis began with CAD. I knew the vex geared motors that I used didn’t have the torque to move a robot over 3-4 lbs with any reasonable speed, so I made sure I kept everything as light and compact as possible.

I thought that anything thicker than 1/2” square stock would be unreasonably bulky for the frame, so that’s what I used. To put the frame together, I simply drilled holes and screwed in wood screws at the corners. I knew it wouldn’t be very strong at corners for impacts, but I wasn’t building this robot to be smacked around very much.

When the frame was together, I super-glued the motors to the frame, 6” from the end of the frame to the front of the motor casing. They were later braced on both side with more wood, also super-glued to the frame. I planned to brace the shafts on the other sides of the frame members with Vex bearing blocks, but the geometry was off just enough to cause additional friction on the shaft. The motors were secured to the frame well enough that I decided not bracing them from the other side of the frame would work- on a larger scale, the forces on the shaft would likely be high enough for that to be a problem, so I noted it for future robots.

I then bolted the bearing blocks to the frame with the 3rd holes in the lowest position for the greatest possible ground clearance. The wheels are from an ancient Lego Mindstorms kit. They are 4” in diameter and roughly 3/8” wide. They are meant to be mounted on Lego shafts, which have a + cross-section, but Vex shafts are the same dimensions along the side, so I could still attach the wheels to the shafts without any modifications.

To keep the shafts in place, I used 1/8” grub-screw shaft collars that came from an old erector set I had laying around. I only had enough collars to keep the shafts in place (not for the wheels as well) so I used small zip ties instead.

The Vex sprockets had such high friction with the axles that I didn’t need to keep them in place with shaft collars (or zip ties, for that matter). The sprockets on the motors have 24 teeth and the sprockets on each axle have 8 teeth. The free speed of the vex motors is 160 rpm, which when  geared 3:1 translates to roughly 12 feet/sec at the wheel. This is quite fast for such a small robot- for different terrains/applications I can simply replace the sprockets with others that came with the kit for more torque.

The receiver is hot-glued to a masonite sheet (also found in my garage) which is screwed into the underside of the frame. The plate also serves as a battery tray.

The battery is a modified NiMH pack from a cordless drill that broke years ago- I simply removed 6 of the 1.2v cells and soldered back the last two for a 7.2v output. The RC receiver can only handle 9v, which is why I didn’t want to go over 6 cells. The motor controllers are powered directly from the receiver- the motors don’t ever draw more than 3 amps.

To plug in the battery to the receiver, I soldered a pwm connector to the output + and – wires. I love pwm connectors for one reason- power is in the middle, which means I can never short out my receiver, regardless of the positions of the ground or signal pins.

The two vex speed controllers connect to the receiver’s aileron and elevator channels. The left goes to the aileron channel, and the right goes to the elevator. To secure the controllers down I simply bunched together the wire until it had roughly 1/2” of slack and zip tied the controllers and wire to the frame.

Everything I need for FPV is simply zip tied to the frame. I currently have all my FPV gear hooked up to the tricopter, which is why the picture just shows the completed drive base.

 

The project was an experiment in 4WD tank steering, which is likely what I will use to control YardBot, a large all-terrain robot that I am designing for general outdoor use. That project will have an article as soon as I finish the design/build process and document it.

---

Balsa Robot

I’m not exactly sure what I had in mind when I built this thing several years ago, but I’m sure it was extremely important at the time. I know that I built it from balsa because I had nothing else laying around. It’s not finished, and won’t ever be- I don’t think it does what it was supposed to do either, but the drive base and actuated arm work, so I might as well document it.

From an engineering standpoint, this thing is pretty stupid. I built it mainly for the chassis-I was experimenting with servos, and I needed an application. I modified some standard 180 degree rotation servos to rotate continuously. Here’s how I did it

Servos use potentiometers to locate their stopping points at either end of rotation. When the output gear rotates, the potentiometer turns with it. To allow the servo to rotate continuously, the potentiometer needs to stay in a neutral position- I removed it and put back the gears into the case. I powered the servo on a receiver using the aileron channel and allowed it to rotate without any input. then I just used a pair of pliers to gently turn the shaft to the neutral position. It took quite a bit of fiddling to get it just right- if you here the motor buzz even a little, the potentiometer needs adjustment. When it stopped completely, I super-glued the potentiometer into place. The output gear has a little plastic disc that has the semicircle shape as the shaft of the potentiometer. I removed that too, so that the output did not engage the potentiometer at all.

All that’s left to do is to cut off the little plastic nub that stops the output gear from rotating past the compound gear underneath it, and the servo will rotate continuously.

I had some large plastic gears that I took out of a toy car a long time ago, so I put them to use. I initially intended to gear the servos to rotate the wheels, but I thought reverse-gearing the servos to turn wheels quickly was a stupid idea- the servos are slow, but have enough torque to rotate big wheels, so I just geared them in a belt drive 1 to 1. I probably could have just attached the wheels to the servo output head, but I believe I was inspired by my FIRST team (it was my first year) to do a belt drive of some kind.

  I didn’t have a belt of any kind, so I just used some rubber bands that I had laying around-this was probably was the worst decision of the entire build. Belts stretch for tension, but rubber bands stretch for stretchiness- applying tension doesn’t do anything, so they kept slipping. At very slow speeds, however, they work just fine.

The arm servo is also continuous rotation ( no idea why I did that) but unlike the chassis servos, it is geared down, which gives the arm unnecessarily high torque. Here’s where I made a huge mistake; I made the servo continuous rotation with my radio’s trim tab all the way to the left. In other words, the arm would only stay put when it was held left as much as the transmitter would let it. I never built a clamping hand for this project, so it didn’t matter too much, but it meant that the arm couldn’t be adjusted at all.

The arm needed to fold to fit within the small frame. I considered a slew of methods, including adding another servo, but I decided against it because I didn’t want to complicate the middle joint and lengthen the amount of time required to extend it. Instead, I just hinged it to fold one way, so that it could fold into the frame sides but still hold load from the top; it works the same way your elbow does. If I had built a hand for it, I might have been able to test how much load it could take. For now, I’ll say it works in theory.

Although I never finished this project, it was a fun one to build. In the future, I might super-size it and properly design and build a robot for a specific purpose. Probably the most important part of the project was that it did give me a ton of experience with digital servos-maybe one day I’ll build a camera pan/tilt mechanism.