The analysis goes through each section testing the circuits specified in the investigation. Once each circuit was working fully they are then constructed on strip board. The circuits can then be connected together adding the required batteries and then placed on the mouse. The first stage was to build and test the sensors.

6.1 Analysis of the sensors

Having built the damped tuned circuit the sensor could be tested to see the voltages produced by placing the inductor near the track. The sensor was placed 10mm vertically away from the ground and then systematically moved horizontal distances away from the track. Fig.14 shows the voltages produced by the sensors as the distance varies. These voltages are vital to know for the later circuits and the positioning of the sensors. Fig.14 Table showing generated voltage of inductor at different distances from wire track

Distance From Wire(cm)		Generated Voltage(mV p-p)		Distance From Wire(cm)		Generated Voltage(mV p-p)
0.5		109		3.0		24
1.0		81		3.5		18
1.5		50		4.0		14
2.0		31		4.5		12
2.5		28		5.0		10

The results in fig.14 clearly show that the inductor generates less voltage as it moves further away from the track. The position of the sensor must generate a sufficient amount of voltage to control motor speed. The closer the sensor gets the faster the motor will go and the opposite when the sensor moves further away. The distance of the sensors from the centre of the mouse at the front must be sufficient to allow a voltage to power the motors but not a full speed. At 1.5 cm from the wire half the maximum induced voltage is produced. This would be an adequate position that produces enough voltage for motor control. The closer to the wire more voltage is induced increasing motor speed, whilst the other would be opposite, slowing down, correcting the path of the mouse. The horizontal position of the sensors may have to be altered to produce better angle control as some variables in the circuits can change the results. Through testing once the mouse is complete the ideal position can be found. A further test for the sensors was to change the frequency through the wire to ensure that it responds to a slight frequency change. The frequency was changed from 19 KHz to 21 KHz and showed the sensors to be working within a �5% range.

The maximum voltage produced was 109mV peak to peak. This needs to be considered as the voltage is going to be used in the sensor circuit. By using the maximum voltage to do the circuit calculations ensures that no component of the circuits (especially the motor) receives too much voltage. The maximum voltage will more than likely be reached to when setting the op-amp gains this value must be used. The sensor circuit will need to amplify this signal to uses to control the mouse. By using a gain of 100 will produce a voltage of 10V peak to peak. This is more than enough voltage to power use within the rest of the circuit. Once the signal has been rectified the voltage will be only five volts (as no negative voltage) and voltage will be lost on the rectifier. This will all need to be considered in the analysis of the sensor circuit.

6.2 Analysis of the sensor circuit

The first sensor circuit was chosen as a group decision. The idea was chosen, as it was a seemed a different idea and in this case the full wave rectifier was slightly simpler set-up. The full wave rectification should produce a less rippled affect for the wanted steady voltage. The only disadvantage was the loss of voltage across the diodes but was corrected in the circuit with the second op-amp. The project was based on the decision to be as light as possible to be faster around the track, so battery supplies need to be limited and number of components as they all add weight. The reduced number of components also reduces the amount of noise created by the system. Both designs try to use as little components as possible and the difference in numbers between them is minimal. After many build attempts and testing by each group member the final circuit was constructed.

The circuit was constructed on bread board first to allow any modifications to be easily implemented. By using the signal generator to produce as 100 mV peak to peak sine wave the sensor circuit was tested. Looking at the first stage was the amplification of the voltage. Using the determined gain of 100 the voltage provided was more than adequate. The oscilloscope provided visual verification of the voltage wave shape and magnitude produced. The measured results matched the expected results of 10v peak to peak. The voltage then went through the full wave rectifier where a lot of voltage was lost. As the negative voltage was rectified the voltage became 0-5V. However the voltage reduced even further as the diodes need 0.7 V each to activate. So at each rectification 1.4 V were lost. However more voltage was lost due to resistances in the system and the output was only one volt. The voltage value needed to be greater; as this was the maximum inputted voltage the wider voltage range means the wider control of speed of the vehicle. So using an op-amp that doubles the voltage solved this problem. The smoothing capacitor was placed before the op-amp smoothing the waveform. However the ripple waveform was amplified too much distorting the steady output. Therefore testing revealed the smoothing capacitor should be placed after the amplification.

Using the oscilloscope and testing the circuit on bread board allowed modifications to the circuit that would not have been allowable on strip board. This allowed the movement of sections of the circuits and freedom to change resistors to provide ideal gains. Once the final design had been established the results on the oscilloscope were clear showing a ripple voltage at about 2.5 V. The next stage was to test the circuit using the sensor on the actual track to ensure that the output was still produced. The voltage produced at the output of the circuit was slightly noisier and distorted but generated very similar voltages. Moving the sensor further away from the track lowered the output voltage as the distance increased. This demonstrated that the circuit worked and would provide adequate guidance control for the mouse. The next stage was to construct the circuit on vary board.

When constructing the circuit on vary board the last op-amp�s feedback resistor was changed to a variable resistor. This allowed the circuit output voltage to be modified at any time to produce the wanted voltage. This also allowed both sensor circuits to be matched as voltage losses and resistances can be different when soldered onto the board. The circuits could be matched so one sensor does not provide more voltage to one motor than the other at a given input voltage. Once again the circuit needed to be tested to ensure everything was working properly. Using the oscilloscope the output of the sensor circuit was identified. The output had become even noisier and the variable resistor had to be adjusted to increase the gain and produce the desired output. The results were not as good as the previous test and the voltage produced was a lot more distorted but should be adequate to control the motors.

Having created two working sensor circuits the best way to test if they will be adequate to control the motors of the mouse they needed to be connected to the mouse motor drive circuit. The sensor circuits must be connected to motor drive circuits of opposite sides therefore controlling the speed of each wheel. So when the mouse deviates away from the track then the voltage reduces to the opposite wheel slowing it down on one side. The faster wheel on the opposite will drive it back on course so the sensor picks up equal voltage again to drive straight. The next stage is to look at the analysis of the control implementation.

6.3 Analysis of control implementation

The characteristics of the motor are very difficult to measure. The rotor resistance in the D.C. motors were not constant depending on the coil in contact. The constant Kw and Kt are difficult to measure and the inertia J, was very difficult to measure as the actual it changed depending on the angular velocity of the vehicle. The characteristics of both motor are so similar that only one motor was calculated. They are so similar this would be able to demonstrate the characteristics of both motors almost identically. The following shows the motor characteristic equation followed by the angle guidance equations.

For Motor 1:

Gf(s) = 1004566/s2 + 3367s + 59369 G�(s) = �(s)/V�(s) = 1004566 . 0.286 / (s + 3348.5)(s + 17.73) .s G�(s) = 287000 / (s+3348.5)(s+17.73) s = 86/(s+17.73)s as (s+3348.5) ignored as pole so far out. Therefore with unity feedback then T(s) = 86Kc /s2 + 17.73s + 86Kc choose ?=0.5

Then sing the ideal characteristic equation s2 + 2 ?Wns + Wn2 given that equals the denominator of the transfer function above. Then Wn=17.73 rad/s and Kc=314/86 = 3.66. This reveals characteristics of the angle control of the system. Theses values can be used for analysis of the systems to predict responses of the mouse before it is built. The mouse built in this project has variable resistors that will allow for compensation in different characteristics of the motors and sensor circuits. These will have to be adjusted before using the mouse. These variations can be accounted for by using control simulations but the differences between motors are so small that it is easier to adjust resistances in the circuit to compensate. The simulation of the forward motion was run showing the motor response was very quick. This demonstrates that using the angle control and motor control should be easily implemented within the circuit. Now that control of the mouse has been implemented and simulated the next stage is to build and analyse the motor drive circuit.

6.4 Analysis of the motor drive circuit

As a group both designs for the motor drive circuit were considered and as a team the first design was chosen looking at all the considerations. The two motor drive circuits consider would easily meet the specifications for the mouse. The bidirectional motor has the advantage of being able to slow the motors down the hill. This would be very useful as the gravity make take over the control of the motors down the hill so the direction control would not work. However the friction of the hill should slow it enough and the track is straight for the mouse to allow it to maintain control. The first design provided enough current to run the motors at sufficient speed and therefore is a perfectly adequate design. This design was chosen as a negative voltage supply was not required reducing the number of batteries for the motor drives by half. This will therefore reduce the weight of the mouse, which will allow it to go faster.

A as a group the motor drive circuit was designed and tested. Each member built the circuit each trying slightly modified designs. These were then combined to produce the final design. One design included the use of a Darlington pair in the circuit to see any improvements to the circuit. There was very little difference between the designs but the Darlington pair was a much more temperamental circuit and also the components would cost more. After many builds of the circuit the final design was achieved. Using the oscilloscope the voltages were measured throughout the circuit and the motor voltage was constantly measure ensuring that no more than three volts reached them. The easiest way to check the circuit was working was connecting the motor to the circuit. This provided clear visual result for the circuit but any problems in the circuit the oscilloscope was used. To represent the input of the sensor circuit there was a dc voltage of 2.5 volts used in testing. The next stage was to test it by connecting it directly to the sensor circuit. Using the output of the sensors it was connected to the drive circuit. The results were immediately clear as too much relief the motors worked and varied when the sensor moved closer to the track.

Once the circuit was established to be working fully the circuit had to build on vary board. Having built the board the whole circuit had to be tested. The sensor circuit was connected directly into the drive circuit. Adding the battery holders to the circuit it was ready for testing. A switch was added to the connections between the sensor circuit and the motor drive circuit so that the mouse could be stopped at any time. Once the switch was turned on the mouse could be seen to be working. The motors responded to the voltage produced by the inductors. As the sensors got close to the track the opposite wheel increased in speed therefore correcting the course of the mouse. The variable resistors had to be adjusted to ensure the wheels were turning at the same speed when the sensors were equal distances apart. The next test was to try the mouse working on the circuit connecting up all the batteries.

The batteries required were two 9V batteries, six 1.2 rechargeable batteries. Each motor drive circuit used two batteries in series to produce almost three volt supply. A further battery was added in parallel to add to the current of the supply when the hill control pendulum was connected. The two nine volt supplies were connected taking a reference of nine volts allowing the batteries to be connected as �9V. Having connected the batteries in appropriate holders the mouse was ready to test on the track. The mouse was switched on and began to follow the track. The mouse managed to go down the hill and begin to follow the corners. The variable resistors had to be adjusted to slow the speed of the mouse to allow it respond to the corners better. The mouse was also tested on the slope. The mouse without using the hill control made it up the hill very slowly, which would be cause problems for the race. Using the hill control the parallel batteries were added giving the extra current up the hill. This increased the power to the motors and increased its speed up the hill very slightly. The hill control therefore tested well and provided adequate hill control.

Having completed full testing on the circuits and devise on the mouse it was established the mouse with adjustments to the variable resistors would be able to make it around the course. However due to the use of mechanical hill control and not circuit speed control the mouse reduced a lot of speed on the hill. The sensor circuits were not as sensitive as would have been liked so the speed of the mouse had to be reduced. The mouse is therefore complete and ready for the race.