Overall Electrical System
Our WARRIOR was controlled by 3 Teensy LCs: Main Controller Teensy, Sonar Teensy 1 (front/rear), and Sonar Teensy 2 (left/right). The Main Controller Teensy controlled all the movement (driving motors), and ball launching (rotating plate stepper motors, and flywheel motors). Sonar Teensy 1 controlled the signals from the HC-SR04s on the front and the back of the WARRIOR, and Sonar Teensy 2 controlled the signals from the HC-SR04s on the left and the right of the WARRIOR. These two Teensy’s sent high or low signals to the Main Controller Teensy where the signal was interpreted.
Overall Electrical Diagram
Power
The robot was powered by the 2 - 7.2V rechargeable NiMH battery packs that were provided for the project. We connected the batteries series to have a total of 14.4V+ for our motors. When fully charged, the batteries actually produced a voltage greater than 16. A switch was used as the main on/off circuit breaker. In order to achieve the 5V required for the Teensy’s, we used 2 LM7805 voltage regulators to step down the higher 14.4V.
The ball launcher voltage was controlled by a separate Buck Converter (DC-DC). We chose to do this so we could have precise control of the ball launcher wheel speeds and have better control ball launcher.
Several large capacitors were placed in the teensy and ball launcher circuits to help minimize power fluctuations. During testing we noticed that sometimes the robot would unexpectedly start the program over. We concluded that the large current draws from the drive motors and/or ball launching motors would rob the Teensy’s of their supply voltage and casue them to reboot. Adding the capacitors helped correct this issue.
The ball launcher voltage was controlled by a separate Buck Converter (DC-DC). We chose to do this so we could have precise control of the ball launcher wheel speeds and have better control ball launcher.
Several large capacitors were placed in the teensy and ball launcher circuits to help minimize power fluctuations. During testing we noticed that sometimes the robot would unexpectedly start the program over. We concluded that the large current draws from the drive motors and/or ball launching motors would rob the Teensy’s of their supply voltage and casue them to reboot. Adding the capacitors helped correct this issue.
Motor Drivers
A total of three L298N Motor Controller were used on our WARRIOR. Each set of drive wheels (front/rear, left/right) was controlled by a L298N Motor Control Module, and the stepper motor for the ball launcher was controlled using another. The drive motors were controlled using a PWM signal generated by the Teensy’s built in analogWrite function. We were able to control the PWM signal for each wheel separately and make adjustments as need. There were some challenges when using this motor controller for the stepper motor. When programming the desired behavior, the stepper motor speed and acceleration didn’t behave exactly as desired. While it accomplished the desired task, we would have prefered better control. Ultimately, we weren’t able to determine it it was an issue with the motor controller or the stepper motor.
Calculations for the motor performance are below:
Calculations for the motor performance are below:
|
|
Sonar Sensing
We used 4 HC-SR04 ultrasonic distance sensors - 1 for each side of the WARRIOR - to aid in the navigation and wall detection. They were arranged as follows:
We programmed the front and rear sonars such that they would output a high signal to inputs 0 and 1, respectively, on the main controller teensy and light up the LED on the sonar teensy 1 if there was an object within 1 inch. We programmed the left and right sonars such that they would output a high signal to inputs 3 and 2, respectively, on the main controller teensy and light up the LED on the sonar teensy 2 if there was an object within 18 inches.
Challenges:
We initially ran into issues with getting consistent readings from the sonars regarding the time it took for the echo to ping back. This inconsistency led to speculations about the current sourcing from the pins and we realized that we needed to use the 20mA pins on the teensy and not the 5mA pins to support the HC-SR04’s 15mA working current identified in its spec.
We then ran into the issue that the sonar sensors would not work when the teensies were unplugged from the computer. We started probing the output pins with the oscilloscope to see what was going on and found that it would work then the probe was there. We then came to the conclusion that the probe was providing very high resistance to those pins, on the order of 1 megaohm, which then allowed the signal to be set appropriately. To resolve this, we put 10 megaohm resistors across all the output pins and ground.
We then ran into the issue that sonars 1 and 3 and sonars 2 and 4 were connected and would trigger the same events unintentionally. This turned out to be a code error in which we initially were outputting signals to pins 13 and 14 for sonars 1 and 3 respectively, but then also outputting a high signal to pin 13, because it is the LED pin, for both sonars. Once this code error was identified and resolved, the sonars proved to work as intended.
Positives:
The HC-SR04 sensors were accurate for distances ranging 2cm - 4m which was more than sufficient for our intended purposes.
We programmed the front and rear sonars such that they would output a high signal to inputs 0 and 1, respectively, on the main controller teensy and light up the LED on the sonar teensy 1 if there was an object within 1 inch. We programmed the left and right sonars such that they would output a high signal to inputs 3 and 2, respectively, on the main controller teensy and light up the LED on the sonar teensy 2 if there was an object within 18 inches.
Challenges:
We initially ran into issues with getting consistent readings from the sonars regarding the time it took for the echo to ping back. This inconsistency led to speculations about the current sourcing from the pins and we realized that we needed to use the 20mA pins on the teensy and not the 5mA pins to support the HC-SR04’s 15mA working current identified in its spec.
We then ran into the issue that the sonar sensors would not work when the teensies were unplugged from the computer. We started probing the output pins with the oscilloscope to see what was going on and found that it would work then the probe was there. We then came to the conclusion that the probe was providing very high resistance to those pins, on the order of 1 megaohm, which then allowed the signal to be set appropriately. To resolve this, we put 10 megaohm resistors across all the output pins and ground.
We then ran into the issue that sonars 1 and 3 and sonars 2 and 4 were connected and would trigger the same events unintentionally. This turned out to be a code error in which we initially were outputting signals to pins 13 and 14 for sonars 1 and 3 respectively, but then also outputting a high signal to pin 13, because it is the LED pin, for both sonars. Once this code error was identified and resolved, the sonars proved to work as intended.
Positives:
The HC-SR04 sensors were accurate for distances ranging 2cm - 4m which was more than sufficient for our intended purposes.
IR Sensing
Although the functionality was not included in the final design, the IR sensor circuit was constructed and is shown below. In testing, we determined that there needed to be a much higher gain resistor across the transresistive amplifier to get a meaningful voltage swing. The comparator's hysteresis points were reset to match the new performance, which had a much reduced amplitude of the signal, even with the higher resistance.
To narrow the field of view of the phototransistor, we added a short tube of heat shrink. This reduced the effective viewing angle of the photodiode, and gave a meaningful train of high pulses from the output of the comparator when the sensor was against the front edge of the playing space and within one inch of the lateral position of the other sensor. This accuracy was sufficient to be useful had we needed to target with IR.
The new design gave a very small variation in voltage - the largest signal was only 1V in amplitude, from 1 V to 2 V. This required a redesign of the comparator hysteresis as well.
To narrow the field of view of the phototransistor, we added a short tube of heat shrink. This reduced the effective viewing angle of the photodiode, and gave a meaningful train of high pulses from the output of the comparator when the sensor was against the front edge of the playing space and within one inch of the lateral position of the other sensor. This accuracy was sufficient to be useful had we needed to target with IR.
The new design gave a very small variation in voltage - the largest signal was only 1V in amplitude, from 1 V to 2 V. This required a redesign of the comparator hysteresis as well.
Final IR Sensor Circuit Schematic
Final Comparator Hysteresis Calculations
Electrical Parts List
