|
The system we are designing will test an
individual's blood alcohol level by analyzing the breath
exhaled into our system, in which the sytem will decide
whether to let the individual start the car. We will be
using an "Ensure" made by "Alcohol Countermeasures" to
measure the blood alcohol level. We will also be
including a numeric touch pad (state machine) that will
replace the existing key ignition system. The driver of
the car will have to key in their pin number instead of
using a key. The the driver will then be signaled to
breathe into our modified breath analyzer. If the driver
passes the test he/she may start the car by pressing a
start button that will control the starter for the
vehicle.

The objective of this design is to help
keep drunk drivers off the road. The system will be
designed so that it will be able to interface with all
current ignition systems. Our system will also eliminate
the use of keys to start the vehicle, therefore
eliminating the need for keys. The user of the system
will be given audio feedback at each stage (ie. whether
they passed the keypad test and then the blood alcohol
test). This will be done by integrating standard voice
chips into the system. These could be replaced by LED's
to control the cost for future designs.
BACKGROUND
The breath alcohol tester is an
electronic device used to test the blood alcohol content
(BAC) in a person's blood stream. Several other methods
of alcohol detection exist, such as chemical reactions,
optical- interference, infrared absorption, and heat
generation from a catalytic reaction on a hot platinum
wire. These methods, however, have the disadvantages of
being difficult to handle and maintain. They are also
very expensive and have a very limited lifetime.
This device utilizes a semiconductor
sensor to analyze a person's exhaled breath. This method
is both simple and affordable. The semiconductor device
offers the an added advantage in that it is robust; it
is highly immune to shock and vibration, and therefore
more reliable.
The sensor consists of a semiconductor
material mounted on a ceramic tube; it is protected by a
double layer of 100-mesh-per-inch stainless steel
screening. When the alcohol tester is turned on, the
sensor is heated to normal operating temperature. During
this process, oxygen is absorbed into the semiconductor
surface. This causes a depletion region to form between
the surface and the body of the sensor. In turn, this
increases the resistance of the sensor.
When an intoxicated person blows on the
sensor, the alcohol is absorbed into the semiconductor
surface and reacts with the oxygen that is already
there. This decreases the sensor's overall resistance in
proportion to the alcohol concentration.
When the alcohol tester is turned on,
the heater coil in the sensor is energized with 5 volts
from a 7805 voltage regulator (IC5). The circuit cycles
through a self-testing and warm-up period. The circuit
resets itself to 0-alcohol reading and sets off the
ready light.
When an intoxicated person blows into
the device and the sensor detects alcohol, its
resistance decreases; this changes the input voltage to
the detector circuit. The detector circuit is made up of
a quad op-amp. The detector circuit is calibrated by 2
resistors (R3 and R4), and the inputs to each section
are controlled by a voltage-dividers (R21-R23). As each
section is activated, the outputs go low and the
sample-and-hold circuitry will latch onto the highest
input value and trigger the appropriate LED. Different
colored LEDs represent different alcohol levels.
If the detected alcohol level is above
the legal limit (0.1%), another op-amp will trigger a
buzzer and the LED corresponding to an "over the limit"
level of alcohol. This indicates that the person is too
drunk to drive.

Time Chart for the CKT
The alcohol tester must be calibrated
using the potentiometer; this must be done before
initial use and should be repeated after 10 uses.
After a test, the sensor takes a few
seconds to ready itself for another test. During this
time, the circuit resets itself and the ready light will
come on again.
This design includes a low battery light
indicator. When this light comes on, the batteries must
be changed to insure accurate test results.
To get accurate test results, do not
drink or smoke 15 minutes prior to taking the test.
MATERIALS
R1 ............. 25k ohms trimmer
potentiometer
R2 ............. 2k ohms trimmer
potentiometer
R3,R4 .......... 5k ohms trimmer
potentiometer
R5-R7,R11-R13 .. 220 ohms
R8-R10 ......... 120 ohms
R14,R18 ........ 10K ohms
R15,R23 ........ 1K ohms
R16 ............ 880 ohms
R17 ............ 4.7K ohms
R19 ............ 1.3K ohms
R20 ............ 390 ohms
R21,R22 ........ 430 ohms
C1 ............. 10uF, 16 volt,
electrolytic
IC1,IC2 ........ LM324 quad op-amp
IC3,IC4 ........ 74LS00 quad 2-input
NAND gate
IC5 ............ 7805 5 volt regulator
D1 ............. 1N4001 diode
LED2,LED4,LED10 Green (LEDs)
LED3,LED5,LED6 Red (LEDs)
LED7 ........... Amber (LEDs)
LED8,LED9 ...... Yellow (LEDs)
SEN1 ........... Nemoto resitive
semiconductor alcohol sensor
BZ1 ............ Piezo buzzer
S1 ............ SPST (on/off) switch
Block Diagram
To implement the combination number
phase of our ignition system, we used a standard ten
number key pad out of a push button phone with two PLDs
(programmable logic device). The first PLD was used to
convert the 7 bit number produced by the key pad (first
3 bits represent the 3 columns and the second 4 bits
represent the 4 rows) into a 4 bit number (N0-N3 on PLD
#1) which represents the numbers 1-9 in binary. The
first PLD also outputs a 1 bit number called the
keypress bit, which will be high if any of the buttons
are pressed. The logic handled by the first PLD is shown
in the following truth table. The 7 bit numbers which
are not shown below do not represent a number on the
keypad and will be handled by the PLD as "don't cares".
Since these numbers can't be triggered by the keypad
they will be ignored.

PLD truth table
The second PLD will handle the state
machine needed to check and see if the three number
combination from the keypad actually matches our
security code - 332. Since state machine logic gets
rather complicated we had to limit our security code
numbers to two binary bits (numbers 0-3 on the keypad)
so that the logic would fit onto one 20 pin PLD. We are
using the state machine shown below to handle this task.
For the state machine to move from state 0 to state 1 it
needs a high input from the keypress. For the state
machine to move from state 1 to state 2 it needs the two
bit input to equal the first number in our security code
- 3. If it recieves a two bit input not equal to 3 it
will go to state 15 where it will stay until the reset
button is pushed triggering a high at the reset pin. The
state machine will then wait for the keypress to go low
indicating that the user has let off of the 3 button on
the keypad. When this happens the state machine will go
to state 3. This process is repeated until the user
enters the last number in the security code (2). When
this happens the code_ok pin goes high, turning on the
blood alcohol tester segment of the ignition system. The
state machine stays put in this state until the reset
button is pressed. Note that only the two least
significant bits are inputed into the second PLD. Thus
the numbers 776 would work just as our code of 332 does.
This decision was made for simplicity, and a more
advanced and usefull combination device would need 4
bits to the input of the second PLD and more than 3
numbers in the combination.

Flow Graph
With the completion of our system we
relized that this idea could be very effective against
drunk drivers. We found a couple of commercial models
which range from $500-$700 dollars. The final cost of
our sytem was $200. Note that our system included a
keyless ignition control system and a voice chip to
signal the driver that they could start the car.
The most difficult part of our design
was bring together all the different systems and
intergrating them into a robust and reliable product.
When testing the design we noticed that it would be
aggravating for the driver to wait 35 seconds for the
alcohol sensor to warm up. In future designs of this
system the designers should look at ways of speeding up
this process.
The sensor, which claims to be a blood
alcohol tester, actually measures the amount of alcohol
that is in your breath. We noticed that if you drank on
sip of beer and quickly breathed on the sensor that you
would be regestared as legally drunk when in fact you
weren't. Overall, the system work better than we
expected and hopefully the state will look at using such
system to keep drunk drivers off the road. |