Solutions
I started this development process by drawing quite a detailed block diagram. This helped me understand the circuit and allowed me to think of possible solutions.
555 timer
The first solution that I through of was how I would build the ultra sounding unit. I decided that using a 555 timer was really the only clock mechanism that would work easily and effectively, I have also used them before which is an advantage so it is probably best for the job.
Amplification
Next I thought about the amplification system I would use, this had to give enough voltage to power the buzzer at the end of the circuit. I thought of the transistor amplifier and the operational amplifier. The difference being that the first would allow me the exact voltage for appliance, and the second would actually amplify the signal received. I felt it would be safer to use the operation amplifier, because it would more sensitive to the surrounds outside and I have experience with the op-amp.
The op-amp is designed to give a large gain from the input voltage, which is exactly what I need.
Switch for the alarm
The final subsystem I needed to think about was the method of switching the buzzer on and off. Logically I found that I needed a latch that would stay on once triggered and then could be reset. I considered a relay circuit of some sort using a solenoid, but instead I decided to use the D-type flip flop gate, because it uses digital information and is easy to configure. The chip needs to be capable of withstanding vibrations and motion experienced by cars as they travel.
Latching device
The gate needs the signal from the receiver to trigger its switch. I will use a Schmitt trigger to square the wave entering the gate. From here the output of the gate will enter and open the transistor amplifier, allowing the buzzer to sound.
Development
Building the ultra sound unit.
My main aim was to find a circuit to power the ultra sound transmitter and act as my investigation into Ultra sound so I could make an efficient circuit for the device. To start with the pulses for the transmitter need to be regulated at 40 kHz, because this is the optimal frequency for the sensitivity and the sound pressure level in the receiver. A commonly used chip for this job is the 555 timer, which I have used in the past. The circuit to the left will give me a square wave of 40 kHz.
The equation used to find the values of the resistors in the circuits is as follows;
1.44
f =
(R1+2R2).C
The Amplification System
Using this equation I achieved my frequency goal, that will power the ultra sound transmitter, in turn this will be received by the ultra sound receiver. The receiver leads into the op-amp circuit. From a decided distance of 40 cm the amplitude received read 1.8 mV, I needed the voltage to be amplified so it could switch the trigger. The switching voltage of the Schmitt trigger is of the order of about 2 V. This showed me that I needed a gain of about a 1000, which is a lot so problems would occur.
The theory proved successful and the gain was achieved; however the op-amp had a lot of regular spikes on it, exceeding the required voltage, also some clipping occurred. If this persisted then it would fault the Schmitt trigger and so the circuit would not work. To stop this I needed the source of the problem, so firstly I tried a smaller gain, which made the spikes smaller, so did not really help. I tried changing the resistor input to the op-amp; this did alter the spikes and clipping but not to a significant enough effect. I eventually found that by using two power supplies, one for the 555 timer and transmitter and the other used with the receiver and the rest of the circuit. The problem was that the 555 timer pulses were affecting the positive and ground rails. The double power supply system gave me a smooth sin curve from the receiver.
Still though the receiver was not sensitive enough so I had to make the amplification larger. I did this firstly by changing the resistor values but then I found that the op-amp seemed to have a limit to its maximum amplification and would not amplify over about 1.7 V, which meant the circuit was still defective. I decided to amplify the first op-amp with a second op-amp. This was about 3 times as sensitive, because smaller resistance values were needed on each op-amp. The circuit was now triggering the Schmitt trigger, but still the wave on the oscilloscope was still uneven and not always triggering, so I had to think about how to change the Schmitt trigger to give a more reliable circuit.
Building an effective latch in the Proximity Detector
I decided a better switch was needed instead of the Schmitt trigger. A clear square wave would allow the D-type gate to latch the data and work more effectively.
To solve this I found a circuit in the circuit book, ‘A Practical Approach to Systems Electronics’ called the positive feed back latch using an op-amp. This is the same as the Schmitt trigger, but because I am going to build it I can control the voltage out limitations. With this wave on the oscilloscope, it showed an even square wave which meant could proceed to the building the rest of the sensor.
The square wave was then lead in to the D-type Flip flop gate, which switches data from the positive rail to the output of the gate when the amplitude of the wave was big enough. In more detail the D type flip flop then on the first clock pulse larger is than 2 V the data from the positive rail (e.g. 5 V) is moved to the Q output. When making this I found that the voltage out from the Q pin could not power a buzzer so I knew I had to use a current-operated switch to directly power the buzzer, I used a npn-transistor due to them being simple to use. This would trigger on about 0.5 mV and then allow a larger current across the emitter-collector pins, this would power my buzzer.
Additional Modifications
With the system working satisfactorily and triggering the alarm at the necessary position I decided to make one more modification to the circuit. I added a variable resistor to the second amplifying op-amp so make the triggering distance ideal. I used a 1.3 MΩ variable resistor. This made the circuit as sensitive as I desired.
The Sensor project in action
In order not to move the prototype circuit the object will move to the sensor. The object starts a metre away and the alarm is quiet, when it moves through 40 cm, this is the distance the alarm is triggered and the sounds shows that the car is close enough to the object. At this distance the variable resistor gave a reading of 65 kΩ.
How it works
The 555 timer is causing the transmitter to pulse at 40 kHz, the receiver is placed next to the transmitter and receives a small signal. The received wave is amplified once in the first op-amp then again in the second op-amp. The wave is now large enough to trigger the positive feed back op-amp. This means that as the sinusoidal wave passes across the later circuit, it is latched into a square wave. The wave is then passed into the clock input of the D-type Flip Flop gate and when the first clock waves amplitude This then triggers a npn-transistor to open the gate between the positive and ground rails to power the Buzzer. A reset button is in place to reset the gate, however this will only re-trigger if parked so the power supply or (ignition) is our on/off supply.
Testing
In the future the sensor needs to be tested as the next stage. The system appears to work because it sounds the alarm when 40 cm away from the object. However because it is designed for the car I should experiment to make sure that the circuit is able to run off a car battery. Then I could consider the average reaction time of a human to react to the alarm so a standard triggering position can be set.
Assessing
The whole circuit I am pleased to say works well and reliably, however the ultra sound has to be placed in the exactly the right direction, because the waves are easily reflected by any surface. This means the unit will have to be mounted in a suitable position so that the waves will be reflected off the bumper line of vehicles.
Limitations and modifications
Most of the development took place in the construction and I have detailed these in the sub-systems report. In addition to these I would like to mention some ideas that would improve the sensor. The first is that the sensor only gives one piece of information feedback, if I were to improve this then the more information would make the system safer and more reliable. For example addition latches could be added to trigger the D-type at different distances from the object. This would allow the driver to be more alert when parking. The Final modification I would consider would be to allow the driver more control over the sensor by adding a specific on/off switch and also a timer for the alarm after it is triggered so it does not become annoying for the driver and passengers.
Summative Evaluation