The Tracked Robot Project 2010

Page 4

We will be getting to the electronic brain of this robot in a little bit... the part that gives it the ability to avoid things, turn motors on and off... along with many other importan tasks such as blinking LEDs to look kool. But for the moment... let's look at two important components before we decide on a brain. Let's wire up the sensors to detect wheel rotations, and let's get a motor driver up and running so the brain can actually have a robot to control.

parts
This is the basis for our wheel sensors... a very simple design.

Looking at the package, the IR LED (On the left) uses 1.2 volts at 100mA of current. I am going to run the robot from a 9v battery that I am then going to regulate into a steady 5v supply. If we do the math... 5.0 volts - 1.2 volts = 3.8 volts more than the part can withstand. It would be damaged but the higher voltage. We need to add a resistor... so we take the voltage (3.8 volts) and divide it by the current (100mA) which can also be looked at as 0.100 Amps.

Here is the formula:   3.8 volts / 0.100 amps = 38 ohms

I don't have a 38 ohm resistor as it is an odd value... but I do have some 39 ohm resistors... what would that do?  Let's find out...

Here is the formula: 3.8 volts / 39 ohms = 0.097 amps or... 97mA of current.

So... the LED will glow 97% of its maximum intensity... I can live with that... and the LED will last longer at a lower current too. Actually... this is a great deal of current for a battery operated device, so likely we will experiment with increasing the resistance to drop the light output and current further. A normal LED runs about 20mA so this is almost FIVE TIMES the current. For now I am connecting a 2N2222 NPN transistor to the LED voltage supply. This will allow me to turn on the LED, make a reading, and then turn off the LED. If I run the LED only 1/10 of the time that alone could cut the power needed 90%. If I can also lower the current, I get additional energy savings, and the batteries last longer. At full load they would last about 3 hours. By being energy conscious I may be able to get 12 hours from a charge.

circuit
Here is the schematic working drawing of how the sensor gets wired.

prototype
The circuit set up as a prototype on a breadboard.
The red traces are 5v, the blue are Ground.
The yellow wire at the bottom runs to an Atmel Mega 88P for testing.

setup
This is my test setup sitting on my messy desk.

To get this mounted into the tracked robot we need to mount the components on some type of circuit board. Rather than go through the expense of having a board etched or the time needed to etch our own board we can create one using some prototyping circuit material. This is a circuit board with plated holes... you just cut to the side you need and mount the components as needed. A rotary tool or scroll saw is used to cut the board... do this outside or with adiquate ventilation... you do not want to breathe in fiberglass or phenolic dust.

Sensor
(View from SOLDER SIDE of board)
Simple Circuit Layout for the Phototransistor and resistors needed.

Circuit board cut to size
   
Components installed, leads bent and cut to shape
 

Solder applied to leads, wires not yet attached.
 
Component side of board - wires not yet attached.
 
This is the Phototransistor board assembled and ready to be mounted into the tracked robot for testing. We need to attach three wires to it so that it will have 5vdc, Ground, and the signal back to the microprocessor to read the wheel counts. Be sure to properly align the phototransistor or you will need to reassemble the board which is no fun to do... and you can overheat and damage the components if not careful.
 
Finished Board
Finished board showing connections to the main circuit board.

Remember: You will need to make two of these boards... one for the right wheel and one for the left wheel.
 
Sensors With Wire
Sensors with wires installed, ready to mount.		 Installed
Sensors mounted in the chassis.
 
closeup
Mounted with a little bit of hot glue.
 
By counting both sides and doing a comparison we can tell if the robot is drifting to one side and make corrections in software. (We will figure out how to control the motors and count the steps later in this project.) For now, we need an IR light source for the sensor to detect. This is accomplished by installing an IR LED on the outside of the wheel so that it's light can shine through the hole and reach the sensor that we just constructed. Since the IR LED operates at only 1.2 volts it will need to have the resistor installed so that we do not damage or destroy the LED.
 
LED
View from the SOLDER SIDE of the board.

The original testing was done using a 39 ohm resistor on the LED board. This produced almost 100mA of current for each of the two LEDS, one per side. This meant that a total of almost 200mA was being drawn from the batteries when the LEDs were running. That's almost 1/4 amp... so the batteries would run out in about 3 hours on a full charge. By increasing the resistance we can lower the current... a typical LED runs about 20ma so these should be able to operate between 20ma and 50ma. Let’s calculate the resistors for each and experiment with the lower power setting.

Assuming 5 volts to the circuitry, we would use 1.2 volts for the LED and need to compensate for 3.8 volts.

3.8 Volts divided by 0.050 amps ( or 50mA ) =  76 ohms
3.8 Volts divided by 0.020 amps ( or 20mA ) = 190 ohms

If we were to use a common 150 ohm resistor, that would give us 3.8 volts divided by 150 ohms = 25mA (approx.)

I think 25mA is a good compromise... so that is the value we will use for testing.



Video showing IR light beam on sensor