RC Car Modifications

Update 2014-02-13: added information about servo and motor controller

This entire tutorial/post can be skipped with the use of a hobby-grade RC Car. They can easily be identified by either the price tag or high quality interchangeable parts in the car itself. The mods outlined here are to upgrade cheaper toy-grade RC cars to resemble the functionality of their hobby-grade counterparts.

Choosing a car

In my experience, it is far easier to use a smaller or regular sized RC car rather than the enormous (1:6 scale+) toys. The larger toys are both more expensive (especially when it inevitably breaks), less convenient to tear apart and harder to modify. In terms of modifications, my main concern was hacking together the motor controller. In smaller cars (powered by AA’s) the voltages are much smaller and the controllers are made from transistors and smaller voltage components (read: cheaper and easier). Larger toys have relays which, in addition to being slightly more difficult to control, switch slower than transistors. This makes speed control far more inaccurate and difficult since it needs a rapid on/off cycle to mix speeds (aka PWM - pulse width modulated signal).

Tearing Apart the Car

Pretty self explanatory. Take off the fluff and outer shells. Keep the electronics in place (for now). Get to the core (“chassis”) of the toy. Locate the major components (and separate if desired).

Modifications

There are several modifications necessary to emulate the driving functionality of a real car. First, as mentioned above, is the motor controller. This way the toy has more controls than just 100% forwards, 100% backwards and neutral. Similarly, a servo-based steering control will replace the binary left-right steering on the toy and allows for (gasp) varying turning angles.

Motor Controller

Rather than trying to reverse engineer the electronics inside the the car, I choose to start over. I got a TI SN754410 dual H-bridge from Sparkfun.

NYU has a really nice tutorial on how to use this chip with an Arduino.

I wired up the H-bridge and tested it with my DIY lab bench power supply and everything worked perfectly. The problem is that the inputs must be 5v and up. Since the Beaglebone has a 3v3 logic level, I will have to use transistors to amplify the 3v3 from the BBB to 5v for the H Bridge. From there I can input whatever voltage I need (probably ~6v aka 4 AA batteries) and drive the motors.

Servo Steering

Toy RC cars are extremely limited in their steering options, typically only turning hard left or right from center. When the car is going forwards (not turning left or right), the front wheels are pulled to the center with springs. By using a servo that constantly maintains control over steering angle, the car has a much stronger and finer grip on steering.

To install the servo, I tore out the existing motor and used a thin piece of aluminum as a flat baseplate. Small strips of brass connected the servo horn to the steering rack on the car.

Next the servo needed to be connected to the Beaglebone. An external 5v power supply is used to power the servo and common ground is shared between the Beaglebone and the power supply. The PWM signal is connected to pin 8_13 via an optional 1k ohm resistor.

A quick python script tested the functionality of the servo.

# Code from Adafruit's Tutorials on Servo Usage with BBB
# Updated to work with the new PWM duty funtionality in the
# Adafruit_BBIO library

import Adafruit_BBIO.PWM as PWM

servo_pin = "P8_13"
duty_min = 3
duty_max = 14.5
duty_span = duty_max - duty_min

PWM.start(servo_pin, (duty_max), 60.0)

while True:
    angle = raw_input("Angle (0 to 180 x to exit):")
    if angle == 'x':
        PWM.stop(servo_pin)
        PWM.cleanup()
        break
    angle_f = float(angle)
    duty = ((angle_f / 180) * duty_span + duty_min)
    PWM.set_duty_cycle(servo_pin, duty)

The servo can now turn the wheels back and forth.

I’ve tried to mount the servo horn so straight forward approximately 90 degrees. Superb accuracy is not paramount (and might not even be possible since the servo’s range of motion is actually a little greater than 180 degrees). PID control will correct for any minor mounting inaccuracies. Given the imperfect connection between the servo horn and the steering rack, relying just on a good mount will actually be detrimental in the long run as the joints loosen. Intelligent and adaptive algorithms are needed to ensure reliable performance.

Now I will need to impose some kind of hard limits on the PWM signal so the servos don’t oversteer and trash itself. Maybe a limit switch of some kind…