CoasterBot Time!
Tortellini the Turtle Bot in the making!
Your Subtitle text
Tortellini the turtle bot 
Hi! My name is Jasmine. I'm 11, and I'm learning about robotics. My dad has taught me some of the basics, and how to solder. I learned a lot during the process of building our robot for Make's CoasterBot contest

My dad is a mechanical engineer, and he sells tiny electric motors, and our robots are usually driven by them. We have built a few line followers, and a robot-like car. The microcontrollers for them have been BASIC Stamp, PIC, and Athena. We have been collecting many parts for this occasion, and we are really making progress on our turtle bot.

Tortellini the Turtle Bot is programmed to walk on four legs, yet he is balanced on a mini-CD. The point of the contest was that the robot would be CD-ROM based, and Tortellini has a CD shell molded into the right shape. With creativity and serious contemplation, we've finished Tortellini the Turtle Bot.

I posted some pictures of our robot making and I also posted the finished pictures and a special video of Tortellini. Watch it in the Meet Tortellini section of our website.

Here are some of the concepts we wanted to incorporate into our robot.

1) Modularity.  -- The device should be able to be assembled from modules which can be tested separately.

2) Multi-purposing - By using components like the gear motors as structural members a more compact structure is formed.

3) Experimentation - Being our first walker and by breadboarding each module we felt our way through this.

Technical Summary

Using a mini-cd as a base, 4 gear motors actuate the legs.  The motors only are operated in the forward directly allowing the use of a single quad h-Bridge.  The walking motion is realized by a offset rotation on top and a sliding pivot point in the middle of the leg. Rubber grommets were used as feet to increase traction and decrease noise.  Due to the lack of a 3.3V h-bridge, 2 Li-ion batteries were used and regulated to 5 V with a 7805.  4 photo reflector sensors are used to note top dead center for the legs to control the gait. Walking is achieved by moving diagonal pairs 180 degrees out of phase. In addition, a servo motor controls the direction of the head which has a higher power IR led and photo transistor pair.

A PIC 16F690 is used for contol. 4 Digital Outputs control individual PWM for the motors through a 500 usec interrupt routine.
This is also used to create the pulse for servo motor which is the 5th output.  To save IO's all sensors are powered continually from the 5 V regulated supply. This, along with the decision to only run the motors forward, saved 10 outputs and a driver chip.  A six pin header was added to be able to program the micro by removing the top board with a pickit 2.

Lessons Learned

1) Buy the thinnest gage solder wire possible. The first 2 weeks with 0.125 inch solder wire and a 40 watt chisel tip solder iron resulted in using 5 feet of de-solder braid. After switching to a smaller solder iron and the thinnest wire the shack had to offer, the only solder braid I used was as a ground jumper.

2) Stacking two through hole boards works best if you flip the top one. It allows the male and female connectors to come 
together and make  good connection.  Also, soldering chips on the wrong side of the board did not work out as poorly as I imagined. With the thin wire it was a breeze.

3) Higher resistor values on the output side of the photo diodes greatly increases the range. I chose 4.7 Kohm for the analog high power and get some signal as much 12 inches from the sensor. For the leg sensors I used 51 Kohm. 

Unsolved riddles

After many hours of soldering a board, the h-bridge would work with out inputs. Even when I removed the pick from the socket the driver chip ran the motors at full speed.  I was able to ground out 3 of 4 pins but one continued, albeit more slowly even after I grounded the signal pin. I ended going back to the breadboard, and making a 4 board which was exactly like the breadboard. It finally worked but I wish I could have gotten the single board solution to work. The profile of the turtle would have been much better.

Points of pride

1) Use of a mini-cd. 5 motors on a tiny platform.

2) Us of the motors as structural elements. The 2 motors on each side and the wire harness form the perfect battery 
compartment that keeps the center of gravity low. Also I found a way to use the mounting holes of the motor to screw down
the first pcb. The corners of the PCB also line up perfectly with the leg tops allowing the sensors to go to the corners of the 
board and be pinned against the motor.

3) Use of connectors to make assembly and disassembly quick and easy. The head scratching that went into the position has been recouped many times over in convenience.

Points of shame.

1) Lack of capacitors. A reference design would show at least a dozen capacitors on a design like this. It has worked but I have a nagging concern that something may go wrong.

2) Use of CA adhesive. Spending my life around micro motors, I know very well the havoc that CA can wreak on the commutator of a brush motor.  In our factory we use UV adhesive but since I do not have a UV oven... CA has been my friend and the motors have held up better than should be expected.

Web Hosting Companies