The circuitry of the car is probably the most complicated portion of the project. First and foremost, there is the amplifier circuit. As the ultrasonic sensors only emit a very shallow signal in response to the presense of the beacon the signal must be greatly amplified in order for the ADC on the 555 to distinguish between the precense/lack of the beacon in a given direction. The amplifier circuit consists of 2 gain stages, and a final gain/level shifting stage. The first two stages amplify the AC component of the signal by 50x each, and the final stage applies another 10x amplification while shifting the signal down into the 0-5V range of the ADC. Some of the difficulty in designing the circuit stemmed from the lack of a negative supply voltage for the opamps. As such, a virtual ground at VCC/2 had to be created to center the voltage of the first two gain stages, and a 3*VCC/4 voltage level was used to center the final gain stage, prior to being rectified and passed to the ADC.

In order to avoid building the fairly large amplification circuit four times over, an anaolg multiplexer (TI cd4067b) was used to route all four sensor signals to the same amplification circuitry. Difficulties arose with this, as the control voltage requires to switch the mux was greater than the 55 could ouput. To deal with this, we used a BJT as a voltage amplifier to allow the control singals to switch between VCC and ground.

Our steering servo circuitry was minimal, requiring only a 5V regulator to drive the power line of the servo, with the PWM coming directly from the 555. The motor control circuity, on the other hand, proved to be quite a headache. Our first problem was that the motor could not be directly controlled by the 555 PWM output; rather, we had to use a FET as a current buffer. The second and larger problem arose from the interference between the sensors and motor. In order for the FET to switch in response to the 555 PWM output, it needed to share a ground line with the chip (and thus the sensors). However, doing so caused enough interference on the sensor lines that the direction control no longer functioned. We experimented with using a phototransistor to switch the FET, while keeping the motor and sensor power/ground lines completely seperate, but ultimately we ended up passing the PWM signal through a BJT, and then to the FET. The ground lines for the motor and sensors were common, but required a large amount of capacitance between VCC and ground, as well as software control modifications in order to prevent the motor from disabling the sensors.