automatic rain operated wiper reportsfinal

8/12/2019 Automatic Rain Operated Wiper Reportsfinal

1/39

CONTENTS

1. INTRODUCTION

2. DESIGN

3. TECHNICAL DETAILS

4. TEST DATA

5. CONCLUSION

6. REFERENCES

ABST

RACT


2/39


3/39


4/39

To reduce man power

To increase the efficiency of the vehicle

To reduce the wor$ load

To reduce the vehicle accident

To reduce the fatigue of wor$ers

To high responsibility

%ess &aintenance cost

The major components of the Automatic rain operated wiper' are follows

(onductive #ensor

(lass frame and #upporting #tructure

)attery

Wiper &otor and its arrangement

*elay

1.2. BASIC #RINCI#LE

Many attempts have been made at constructing an effective, reliable, and

cheap rain detection and wiper control system for vehicles. perfect system could

subtract one more tas! from the driver"s wor!load, and allow them to better !eep

their eyes on the road and hands on the wheel during foul weather. #espite this,

automatic rain$sensing wiper systems are relatively uncommon in modern vehicles

for a number of reasons. They are often too e%pensive, too unsightly, or too


5/39

unreliable to be desired in new automobiles. &hile a number of different design

approaches have been made to improve upon these issues, none have been

successful enough for the technology to become widely adapted in new vehicles.

'y far the most common rain detection method is the use of an optical sensor.

These optical sensors function by transmitting an infrared beam at an angle

through the windshield and measuring the reflection to determine the presence of

water. This is a relatively difficult tas!, reuiring comple% circuitry and precision

manufacturing. ptical sensors are thus somewhat e%pensive and can produce false

readings when dirt or other particles on the windshield cause a reflection

mimic!ing that of rain. 'ecause it relies on an infrared beam for detection, the

optical sensor also suffers from a very small sensing area on the windshield,

limiting its effectiveness in rapidly responding to light rain. *n addition, the sensor

housing is physically bul!y, reducing its appeal in lu%ury vehicles.

These issues can largely be mitigated by using a capacitive sensor rather than an

optical one. *nstead of sending an infrared beam through the windshield glass, a

capacitive sensor wor!s by emitting an electric field which can pass through the

glass to interact with ob+ects resting on it. 'ecause water and other ob+ects such as

dirt or roc!s interfere with the electric field in very different ways, the sensor will

be less li!ely to be fooled if designed correctly. Unli!e a standard capacitor, which

confines the electric field lines between two conductors in a tight pac!age, a

capacitive sensor allows the field lines

to spread out, and is designed to ma%imie the fringing of the electric field lines

away from the conductors. These electric field lines are !nown as fringe fields-,


6/39

and are vital to the operation of a capacitive sensor. 'ecause they e%tend away

from the conductors, which are typically +ust copper traces laid out flat on a printed

circuit board (C'), the fringe fields can be interacted with by other ob+ects. &hen

conductive or dielectric ob+ects interfere with these fields, it changes the

capacitance of the capacitive sensor, as seen in /igures 0 and 1. This change in

capacitance can then be detected via circuitry and used to modulate an output

signal. Capacitive sensors can detect the presence, position, and type of conductive

or dielectric material interfering with their fringe fields. &hen multiple capacitive

sensors are connected in an array, they can also be used to detect movement of a

conductive or dielectric ob+ect.


7/39

The electric field is created by applying an alternating current (C) voltage to one

of the conductors forming the sensor traces. typical button sensor reuires only

two conductors, which never physically connect but are separated by a small

distance and patterned into shapes. #epending on the application of the sensor, the

sensor traces can ta!e on a variety of different sies and patterns. The layout of the

traces is often designed to ma%imie the fringing fields over a given area. The

traces, along with the materials surrounding them, also form the base capacitance

of the system, typically along the order of 1 2 13 pico$/arads (p/) in magnitude.

'ase capacitance should be minimied when possible, as the change in capacitance

resulting from fringe field interference is often less than 3.4 p/, and detection is

easiest when the changing capacitance value is close to the base value.


8/39

The idea to use capacitive$sensing to detect rain on a windshield is not

entirely new, as seen in United 5tates atent U56378790, among others. :owever,

technical limitations have largely prevented such designs from being commercially

viable. &ith advances in modern integrated circuits over the past decade, however,

this problem can now be avoided under the proper design. :TC* had previously

been contracted with ;nterprise ;lectronics to design a capacitive sensor for this

application, but development was halted.


9/39

2. DESIGN

WORKING OPERATION:

VEHICLE

MIRROR

SENSOR

CONTROL

UNIT

RELAY

WIPER

MOTOR

BATTERY


10/39

The battery supplies the power to the sensor as well as rain operated

motor. Wiper motor is automatically +, during the time of rainfall. The senor is

fi-ed in the vehicle glass. The conductive Touch/ sensor is used in this project. It

senses the rainfall and giving control signal to the control unit. The control unitactivates the wiper motor automatically. This operation is called Automat! "a#

o$%"at%& '$%"(

2.1. DESIGN S#ECIFICATIONS


11/39

*n designing the capacitive rain sensor, the following design specifications

were provided by the sponsor>

A' FUNCTIONALIT(

#etect and report the presence of one drop of water placed on top of a 6mm

thic! glass windshield above the sensor trace area

B' ACCURAC(

Must not falsely trigger the wipers when a hand is placed in pro%imity of the

sensor trace area

rovide at least two different output signal levels depending on the amount

of rain present on the windshield

'e shielded from the vehicle interior to avoid interference? only water on the

windshield should activate the wipers, not ob+ects or circuits inside the

vehicle

Maintain all performance characteristics across the temperature range from

@@ 2 013 degrees /ahrenheit

C' CO%#ATIBILIT(

#evice fits in e%isting :yundai optical rain sensor housing area (0143 mm

1)

#evice mounts to interior of windshield via adhesive

#evice can operate on vehicleAs 01 B power supply


12/39


13/39

Unli!e some of the more open$ended pro+ects, :TC* provided #esign a

specific solution to the problem of detecting rain on the windshield through the use

of capacitive$sensing. The most critical component in the design was a circuit to

monitor the capacitance of the sensor traces and modulate an output signal

correspondingly. The conceptual design descriptions thus represent a number of

different variations on this critical component.

2.3.1. DESIGN A CA#ACITI*E+SENSING CIRCUIT

The first proposed design was to build a capacitive sensing circuit from

basic components such as op$amps, comparators, and passive components. #ue to

e%perience in analog circuitry, the team realied that the capacitive sensor traces

form a variable capacitor that changes as ob+ects interfere with the fringe fields.

Many circuits e%ist that utilie the time$constant principle of an


14/39

*t was determined that this circuit would not provide the high level of accuracy

needed to determine the presence of minute amounts of rain through a 6$9 mm

thic! glass windshield. Considering that the base capacitance (steady state) of the

sensor traces was going to be around 4 2 04 p/, and that the changing capacitance

from rain was e%pected to be between 3.0 2 3.4 p/, this would result in a change in

a very small change in output freuency. This would be difficult to differentiate by

a microcontroller and would also be highly prone to errors from noise. *n addition,

designing a capacitive$sensing circuit, when highly accurate dedicated circuits

were available on the mar!et, was a ris! that would not only lower the accuracy of

our product but ta!e precious development time.

2.3.2. %ICROCONTROLLER CA#ACITI*E+SENSING %ODULES


15/39

Many recently released microcontrollers include specific hardware modules

for capacitive sensing. /or e%ample, Cypress 5emiconductor has a popular

Cap5;5;- module, and Microchip has the appropriately named Capacitive

5ensing Module-. These modules were investigated as potential solutions to the

capacitance sensing circuit. This method would simplify the system, as the

microcontroller could perform two tas!s 2 monitoring the capacitive sensor traces,

and processing the change in capacitance to determine wiper action. 5ince a

microcontroller would still be needed for processing if a separate circuit were used

to monitor the sensor traces, it would be convenient to have the microcontroller

perform both tas!s. Unfortunately, these hardware modules are primarily designed

for human touch applications, and it was determined that they would not possess

the e%treme accuracy needed for the product, and offered by other, stand$alone

circuits such as the nalog #evices #DD84. :uman touch applications are

relatively easy wor! when compared to a rain sensing application, as the covering

of the sensor traces is often only 0 2 1 mm thic! instead of the 6 2 9 mm of glass

covering a standard windshield. /urthermore, the change in capacitance to a sensor

from a human finger is much larger than a change in capacitance from a few

raindrops. Thus, while the capacitive$sensing modules would be a very convenient

solution to any human interface application, they don"t provide the accuracy

needed for reliably detecting rain through a windshield.

2.3.3. ANALOG DE*ICES CA#ACITANCE+TO+DIGITAL CON*ERTER

5tand$alone integrated circuits often off better performance than integrated

modules. nalog #evices" offers a series of highly regarded capacitance$to$digital


16/39

converters (C#Cs). These chips offer industry leading accuracy in a variety of

different configurations for applications reuiring only one sensor to ones

reuiring up to 08 sensors. table of all nalog #evices C#Cs is illustrated in

Table 0. 5ince measuring only one sensor on the windshield, which can be thought

of as a button sensor-, only one channel of conversion was reuired. This

narrowed our search down to either the #D040E#D04@ low$power, 01$bit, one$

channel C#Cs or the #DD84E#DD8D 18$bit, one$channel C#Cs. ower

consumption was of little importance to the design, as it would be low regardless

and the device would be running off of the vehicle"s power system. The

#D040E4@ 01$bit C#Cs only cost appro%imately F0.D4 per, while the #DD84E8D

cost closer to F8.43 per. :owever, the #DD84E8D offer 18$bits of accuracy on

capacitance readings from the sensor, while the cheaper, low$power #D040E4@

offer only 01$bits. s performance was the most critical criteria, the decision was

made to focus on the #DD84E8D C#Cs from nalog #evices.

These circuits are designed for one channel of conversion, enabling one

single$ended capacitive button sensor- or two differentially operated capacitivebutton sensors- to be monitored. The term button sensor- simply indicates that

the capacitive sensor is ta!ing only one series of measurements over the single

capacitor formed by the sensor traces. *t does not indicate that the sensor is to be

usedas a human$interface button, although it potentially could be. *t is useful to use

the term button sensor- to differentiate a single point calculation as opposed to a

slider- or array of sensors, which are integrated together to perform analysis of

moving ob+ects. 'oth the #DD84 and #DD8D operate on either @.D B or 4 B #C,

and have a built in e%citation source generator, which is a @1 !: suare wave

with pea!$to$pea! amplitude eual to the operating voltage (Bdd). This e%citation

source is connected to one conductor of the capacitive sensor traces, and the other


17/39

conductor is tied to the Cin- pin. The primary difference between the #DD84

and the #DD8D is that the #DD84 is designed for floating capacitive sensor

traces, while the #DD8D is designed for sensors in which one trace is grounded.

'ecause the #DD8D"s sensor capacitance is between the e%citation conductor and

a grounded conductor, any e%traneous capacitance between the e%citation pin and

ground will accumulate as a parasitic capacitance, ma!ing the base capacitance of

the sensor appear larger than it should be. 5ince the capacitive sensor base

capacitance is only between 4 2 04 p/, any additional parasitics can easily

dominate the base capacitance of the system, leading to errors. lternatively, the

#DD84 is designed for floating capacitive sensors, in which the Cin- conductor

is not grounded but is instead floating. nly capacitance formed between the Cin-

trace and the e%citation trace add to the base capacitance of the sensor? any

capacitance to ground does not increase the effective capacitance of the sensor.

arasitics to ground can form very easily through shielded cables or on C'

layouts, so this ma!es the #DD84 design more robust. The decision was made

early on to focus on implementing the design with the #DD84, with the #DD8D

as an alternative if problems arose.

2.4. FEASIBILIT( %ATRI,

The /easibility Matri% is a useful development tool allowing for uic!

comparison between a numbers of different design schemes based on weighted

design factors. #esign Team 6 concluded that accuracy was the most important

design factor, as the capacitive rain sensor would be useless if it could not

accurately detect a change in capacitance caused by rain through a 6 2 9 mm glass


18/39

windshield. 'eyond this, cost was evaluated as the second most critical factor, as

one of the primary pro+ect goals was to develop a sensor with an estimated

production cost less than the current optical sensor. Through the /easibility Matri%,

#esign Team 6 compared the three proposed designs and determined that the use

of a stand$alone capacitance$to$digital converter from nalog #evices, the

#DD84, would provide the best solution for a capacitance monitoring circuit.

2.5. #RO#OSED DESIGN SOLUTION


19/39

has developed an accurate and ine%pensive capacitive rain$sensing system utiliing

the bloc! diagram architecture shown in /igure 8. This device has four primary


20/39

components> a capacitance monitoring circuit, a microcontroller, a voltage

regulator, and the sensor traces. These components are mounted on a stac! of two

two$layer C's which are neatly housed in a plastic enclosure and mounted to the

interior of the windshield. The lower C' contains the sensor traces, which adhere

directly to the windshield, on one side and a wire connector on the other side. The

upper C' mounts appro%imately 0 cm above the lower and contains a protective

ground shield on the bottom layer, and the surface$mount components and

connectors on the top layer. The device layout is illustrated in /igure . The

prototype to be displayed at #esign #ay contains the microcontroller in a separate

housing to allow it to interface with a laptop, which will display the wiper

operation and sensor data through a computer program. fully functioning wiper

system for display purposes was not realistic, however, an actual :yundai

windshield will be on display with the sensor mounted to it. roduction$level

prototypes will have the microcontroller on the windshield$mounted unit itself, and

these circuits will be on display at #esign #ay to give viewers a better image of

how the final product will loo!.

2.5.1. CA#ACITANCE %ONITORING CIRCUITR(& ANALOG DE*ICES

AD--45


21/39

s described in sections 1.@.@ and 1.8, the nalog #evices #DD84 was

chosen as the capacitance monitoring circuit. The #DD84 interfaces with both the

capacitive sensor traces and the *C microcontroller processor. *ts primary role is

to sample the changing capacitance of the sensor traces and output that data as a

digital signal to the microcontroller for processing. The #DD84 communicates

with the microcontroller via a two$wire *1C standardied communication system.

This is a MasterE5lave system with a Master$generated cloc! line and bidirectional

data line. The #DD84 is powered by the 4 B #C output from the #@@30 linear

voltage regulator. *t produces a @1 !:, 4 B suare wave e%citation signal to be

routed to one of the sensGor traces, and the other sensor trace connects to the Cin-

pin. The #DD84 comes standard in a 06$pin surface$mount (T55$06) pac!age.

2.5.2. %ICROCONTROLLER& %ICROCHI# #IC1F452/0#IC16F126

microcontroller is necessary in the design to control the #DD84 and

process the incoming capacitance data. The *C09/8413 was selected for use in

the prototype display unit due to its free availability in the M5U ;C; 893 lab. /or

production$level prototypes, the very similar but smaller *C06/0916 will be used,

as it contains only 09 pins as opposed to the 83 on the *C09/8413.

The *C09/E06/ is a popular and affordable 9$bit microcontroller which runs off a

4 B power supply and comes in a #* or surface$mount pac!age. *t can be

programmed using the C programming language to perform a wide variety of

tas!s, and has @.4 !' of program memory. *n the capacitive rain sensor, the *C


22/39

serves as the Master in the *1C communication system with the #DD84. *t is

responsible for configuring the #DD84 into the correct operating state, polling it

for capacitive and other data, and interpreting that data by comparing it to !nown

capacitance values gained through e%tensive testing of the device. *f the incoming

capacitive data falls into a certain range over a certain number of samples, the *C

will output a signal instructing the wipers to engage. /urthermore, the *C can

differentiate between varying levels of rain to ad+ust the speed of the wipers, and

prevent false positives by ignoring capacitance values outside the range of rain.

2.5.3. CA#ACITI*E SENSOR TRACES& CUSTO% DESIGN

The capacitive sensor trace layout is critical to the performance of the

capacitive sensor system. The shape and spacing of the two traces forming the

capacitive sensor are directly related to the electric field lines produced when the

e%citation voltage is applied. s the rain to be detected is present through 6 2 9 mm

of glass, the sensor traces should be designed as to ma%imie the fringing fields

away from the plane of the C'. =lass has a relatively high dielectric constant of

around 8.4, allowing easy transmission of electric fields through it. onetheless, 6

2 9 mm is a very large distance away from the sensor traces to have to measure, as

most capacitive touchscreens have an overlay thic!ness of only 0 2 1 mm. The

software CM5H was used to model a variety of different sensor layout designs,

where parameters such as trace patterns, conductor width, conductor spacing, and

total sensor sie could be ad+usted to find the perfect layout for the system. These

parameters had a large impact on the total system capacitance, which had to be lessthan 06 p/ due to the range of the #DD84 C#C, and the shape and strength

of the fringe fields. Using CM5H, an e%act sensor trace pattern was decided

upon, and empirical results mirrored that of the software"s predictions.


23/39

2.5.4. *OLTAGE REGULATOR& ANALOG DE*ICES AD#33/1+5

*n a vehicle, the typical battery voltage can range from 00 2 [email protected] B

depending on the strength of the battery and the operating state of the vehicle and

the alternator. 'oth the #DD84 and *C microcontroller reuire 4 B #C for

operation, so a reliable voltage regulator was reuired to scale the vehicle power

supply voltage to 4 B. The nalog #evices #@@30$4 is a linear voltage

regulator which can accept up to 08 B of input voltage, and outputs a preset 4 B

#C. *t can source up to 033 m of current, more than enough for the entire device.

*t offers high linearity, a wide operating temperature range, and is available in a

surface$mount pac!age. The #@@30$4 reuires a capacitor on the output pin of

at least 3.8D u/ in magnitude for proper operation.

3. TECHNICAL DETAILS

3.1. SENSOR TRACE DESIGN

The capacitive sensor traces are formed by two copper conductors, closely

spaced, laid out flat on a C'. This C' adheres directly to the interior of the

windshield with the use of @M 869M adhesive transfer tape. This tape is non$

conductive and has been recommended for similar applications (see


24/39

Capacitive measurements are ta!en over the sensor trace area on the windshield?

therefore, only rain hitting this area will be detected. This still offers a larger

detection area than the current optical system, however. 5ince the purpose of the

sensor is to detect between rain, no rain, and other ob+ects on the windshield, no

data is reuired about the movement of the ob+ects, +ust the presence of them. This

negates the need for a comple% trace layout such as a slider or a touch$pad, which

are used to trac! movement, typically a human finger. Therefore, a single button

sensor design forming one capacitor to be measured was used, reuiring only the

two traces mentioned previously. The layout of these traces can ta!e on a variety of

different shapes and sies. ;%amples of sensor trace layouts are illustrated

/or a button sensor with two conductors, the primary design variables to consider

are the pattern of the two conductors, the width of the conductors, the spacing

between the conductors, and the overall sie of the sensor layout. ll of these have

a substantial impact on the total capacitance of the system, as well as the

distribution of fringe field lines. Typical patterns include concentric circles,

parallel lines, or interweaving fingers-. The sie of the sensor layout is chosen to

match the system environment. *f the sensor is to detect a human finger touch, the

overall sie should be close to the sie of a fingertip. /or the capacitive rain sensor,


25/39

the sie was chosen to be as large as possible without e%tending beyond the sie

constraints provided by :TC*. #ue to the #DD84 C#C"s ability to only null out

the base capacitance up to 06 p/, the sie also had to be ad+usted so the sensor did

not e%ceed that value.

The spacing between the conductors is critical to the overall capacitance of

the sensor, as well as the distribution of fringe field lines. continuous spacing of

3.14 $ 0 mm between conductors is common, as this typically provides a good

combination of large fringing fields and small base capacitance. The thic!ness of

the windshield overlay presented a considerable design challenge, and because of

this the fringe fields too! primary concern. *f the fringe field lines did not e%tend

all the way through the glass, the change in capacitance from any ob+ect on the

windshield would be much smaller than if the lines did e%tend all the way.

:owever, as the sensor traces move closer together in a design, the capacitance

will increase, so a balance must be struc!. The ideal sensor trace layout for the

capacitive rain sensor is the one that produces the farthest e%tending fringe fields

and covers the largest area, while minimiing the sensor capacitance (ma%imum of06 p/ due to #DD84).

ssuming an effective sensor design, care must also be ta!en in the materials

surrounding the trace area. The dielectric constant of a material, is a measure of

the materialAs ability to transmit an electric field.

:igher values of indicate a better transmission of electric fields. The dielectric

constant of air is 0, standard /


26/39

present between the sensor trace area and the windshield, as any air gap would

wea!en the field above it.

3.1.1. CO%SOL %ULTI#H(SICS

CM5H Metaphysics is a power scientific tool that allows the use of

visual environments to model and implement engineering problems. The software

uses artial #ifferential ;uations (#;s) to solve for complicated models. *t is

basically a computer program that allows the modeling and simulation of a wide

variety of physical phenomena. Technical problems relating to the field of>

acoustics, electromagnetic, heat transfer, fluid dynamics, structural mechanics and

M;Ms (Micro ;lectro Mechanical 5ystems) can be modeled and studied using a

rich and interactive user environment.

;ven though the software allows modeling of comple% applications, it does not

reuire an in$depth !nowledge of numerical or mathematical analysis. *t is possible

to build models by simply defining the physical parameters li!e area, length, width,

flu%es, and constraints rather than defining the euations. nce the parameters are

defined and the sub$domain and boundary conditions are set, CM5H

automatically compiles a set of #;s to represent the entire model. #ue to the

simple user interface and easy modeling, CM5H was chosen to model the

capacitive sensor and observe the base capacitance before actually fabricating the

C' design. There was also a time constraint and the team did not have enough

time to e%plore other alternatives.


27/39

nother reason the team opted to use CM5H was because of the

availability of the software. rior to using this software, the team had designed a

capacitive sensor on ;agle C' #esign to verify if the design wor!s. The sensor

wor!ed surprisingly well for our first try but a more accurate design was needed as

reuired by our sponsor. That is why CM5H was used to design the capacitive

sensor model and then compare various models to see which one is more accurate.

couple of designs were laid out and the best one was chosen based on the

CM5H results. *n the following, only four designs will be discussed to give an

overview of how CM5H was used to optimie the design of the capacitive

sensor.

3.2. ANALOG DE*ICES AD--45 CDC

s described in section 1.4.0, the #DD84 is the mediate between the

capacitive sensor traces and the microcontroller. The core of the #DD84 is a 18$

bit 5igma$#elta architecture #C which is modified to convert capacitance

directly to a corresponding digital signal. simplified diagram of this capacitance$

to$digital converter in the #DD84 can be seen in /igure 09, and a more detailed

circuit schematic of the 5igma$#elta C#C is displayed in /igures . t a high level,

the 5igma$#elta C#C functions by balancing charge through two capacitors 2 the

variable sensor capacitor, Csensor, and an internal reference capacitor. The

capacitors are switched between a fi%ed input voltage to charge them, and then


28/39

discharge through an integrator. This can be thought of as a charge amplifier, as

illustrated in /igure , which produces a voltage proportional to the total charge. s

sensor increases, more charge is pumped into the integrator from that branch

because>

I J C K B

&ith increasing sensor, the output of the integrator will grow larger. This is fed to

a comparator to produce a series of eros and ones, which vary with the charge

needed to balance the feedbac! loop. The feedbac! loop connects only to the

reference capacitor, so as sensor increases, the output voltage increases, which is

fed bac! to the reference side and increases the charging voltage of that capacitor

to balance the two branches. The feedbac! signal is also fed through a third$order

digital filter to produce the digital result which can then be output to a

microcontroller for processing. The Bref(L) and Bref($) signals are reference

voltage signals supplied by an internal temperature sensor to The #DD84 can

measure up to LE$ 8.376 p/ changing capacitance, and outputs the result as a 18$bit

digital signal. *t can, however, accept up to appro%imately 06 p/ of unchanging

base capacitance from the sensor traces. This base capacitance can then be nulled

to appro%imately 3 p/ using the on$board C#C. The C#C can be thought

of as a programmable negative capacitance value which can be added to the Cin-

pin to null the base capacitance to around 3 p/. The #DD84 will then be able to

measure the full range of LE$ 8.376 p/ of changing capacitance from there. *f one

were not to use the C#C, and had a base sensor capacitance value of wellover 8.376 p/, the data output would be a constant reading of 8.376 p/- and the

sensor would be useless.


29/39

The #DD84 interfaces with the *C microcontroller using the *1C

communication system, in which the #DD84 is the 5lave and the *C the Master.

*1C stands for *nter$*ntegrated Circuit-, and is technically a multi$master serial

single$ended computer bus. The Master can control multiple 5lave devices,

although only one is used in this design. The *1C system contains only two wires, a

cloc! line !nown as 5CH- and a data line !nown as 5#-. The lines are open$

drain type and reuire pull$up resistors. The 5CH line is generated by the Master

and used to synchronie the two devices, while the 5# line transmits data bit by

bit bidirectional but is controlled by the Master. 5tandard operating freuency is

033 !:.

The #DD84 contains 07 eight$bit registers, many of which must be set to

configure the C#C into the correct operating mode for the rain sensor system. s

the *C microcontroller is the Master, it is responsible for writing the correct he%

codes into the registers. The *C is programmed to perform an initialiation

seuence upon start$up.

The #DD84 is first reset to clear any data or settings. The e%citation signal is then

setup to be full strength of Bdd. e%t, the C#C"s are set to null out the base

capacitance of the system close to 3. The capacitive channel is setup to put it into

single$ended mode at a sample update rate of 61 ms. fter this, the temperature

channel is setup to ta!e temperature measurements every 61 ms as well. /inally,

the #DD84 is placed into continuous conversion mode, where it will start

producing 18$bit capacitive data readings appro%imately every 61 ms. The data isstored in three registers, each of eight$bits, and must be read seuentially to ensure

that no data corruption occurs. fter the initialiation seuence is complete, the

*C polls the status register of the #DD84 to determine when a capacitive data

sample is available to be read. &hen the status register bit goes high, the *C reads


30/39

from the three capacitive data registers seuentially and stores the data for

processing. Temperature data can be read in the same way. *f the initial starting

temperature is below @1 degrees /ahrenheit, the system will shut off as the product

is not intended to wor! in sub$freeing conditions, as per the sponsor :TC*.

3.4. AD#33/1+5 *OLTAGE REGULATOR

The capacitive rain sensor is designed to operate on a vehicle"s power

supply, which can range from 00.4 2 [email protected] B #C depending on the strength of the

battery and the current state of the vehicle and alternator. The *C and #DD84

both operate on 4 B, so a voltage regulator was reuired to regulate the changing

vehicle voltage to a steady 4 B. *t also had to meet the current reuirements of the

two components. The #DD84 uses only 3.D m of current, while the *C06/0916

uses appro%imately 4 2 03 m depending on its operation state. The nalog

#evices #@@30$4 is a fi%ed 4 B #C output, up to 08 B #C input, linear voltage

regulator capable of supplying up to 033 m of current, more than enough for the

device. The #@@30$4 produces no high freuency switching noise to possibly

interfere with the sensor. *t comes in an 9$lead 5*C surface$mount pac!age which

is mounted close to the other components on the top layer of the upper C'.

bypass surface$mount capacitor of 3.8D u/ is present on the voltage input pin to

increase voltage stability at the input. The #@@30$4 reuires a capacitor of at

least 3.8D u/ on the output pin for proper operation, and an additional surface$

mount capacitor is mounted onto the C' for this purpose.

3.5. #CB LA(OUT DESIGN

The layout and geometry of the C's were critical to the functionality of the

sensor. *nitially, a four$layer C' was considered, as this would reduce the

comple%ity and cost of assembly of the device. *n this case, the bottom layer would


31/39

contain the sensor traces, the second layer would be empty, the third layer would

contain a ground plane (ground shield), and the top layer would contain all surface$

mount components and connectors. This design was not pursued because of the

worry that the four$layer C' would not be fle%ible enough to curve to fit the

geometry of the windshield, concerns that the ground shield would be too close to

the sensor trace area, and the difficulty of layout out four component traces on only

one layer.

*nstead of a four$layer C', a two$layer design was used with the C' cut into

half and mounted vertically on top of each other with a spacing of 0 cm in

between. The bottom C' contains the sensor traces adhered directly to the

interior of the windshield, and a connector on the top layer. The top C' contains

a ground shield on the bottom layer, and the surface$mount components and

connectors on the top layer. The top C' mounts above the bottom C' by fitting

into the small plastic enclosure which covers both devices. ;%pressC' offered

the most ine%pensive prototype C' production service. Using the Miniboard-

option, fi%ed sie @.9- % 1.4- C's can be produced at three boards. ;%pressC'

reuires that their software, C'Hayout, be used in designing the C's. This

software offers industry standard features and an intuitive user interface, so that

layout out the board designs was very straightforward. arts can be selected using

the built$in part finder tool, which places the pad geometry of the selected part on

the design. *f the part to be used is not in the catalog, a custom pad geometry can

be constructed. The design includes a total of three integrated circuits, three


32/39

terminal bloc! connectors, and four passive components, all of which are surface$

mount. The sensor trace design also had to be placed onto the C'. 'ecause the

board was intended to be cut in half, the bottom C' was placed on one side of the

design, and the upper C' on the other. 5ee ppendi% # for figures relating to the

C' layout.

4. TEST DATA

4.1. AD--46 E*ALUATION BOARD TESTING

*n the early design stages of the pro+ect, once the decision to utilie the

#DD84 was made, an evaluation board for the #DD86 was ordered to allow for

rapid prototyping of sensor trace designs. The #DD86 evaluation board contains

the #DD86, which is the e%act same as the #DD84 e%cept for it allows for two

channels (two sensors) to be measured instead of one. The evaluation board

contains built in circuitry to allow the board to connect directly to a laptop, and

includes a C# with software to run a program allowing all data from the #DD86

to be displayed visually on the laptop. Capacitive sensor traces can be connected to

the input pins of the #DD86 and performance can be +udged through use of the

software program. This allows for easy and rapid testing of different sensor trace

layouts to determine best performance.


33/39

#esign Team 6 utilied the M5U ;C; 5hop"s capabilities to construct simple two$

layer C's for performance comparison between sensor designs. Using the

CM5H software, sensor trace layouts were created and analyed. These were

then transferred to a C' layout using the software ;=H;. They were then

provided to the ;C; shop, which produced small sensor trace C's for use in

testing, as seen in /igure . These test C's were then connected to the evaluation

board, and adhered to a small test piece of windshield glass. b+ects could then be

placed on the test piece of glass and the performance results, such as the change in

capacitance, were displayed on the laptop. Using this process, the team was able to

analye real$world performance compared to predicted performance in CM5H.

The design predicted to wor! the best in CM5H also performed the best in

testing, and this sensor trace design was chosen as the pro+ect moved forward and

was used in the C' layouts from ;%pressC'.

4.2. #IC I2C INTERF ACE INITIALIATION TESTING

nce the sensor trace layout was decided upon, the ne%t step was to begin

testing with the #DD84 and the *C microcontroller without the evaluation board

present. This was done by using a standard protoboard with the *C and #DD84

mounted, and a serial and U5' connector hoo!ed up to the *C to allow for

programming and data transfer to a computer. 'ecause the #DD84 is a surface$

mount component, a T55$to$#* adapter was reuired to be able to be able to

mount it in the protoboard.

n image of the protoboard used for testing purposes is shown in /igure 19.

The #DD84 and *C were connected as would be in the final design, as seen in

the schematic diagram in ppendi% C, e%cept for the addition of the U5' and

5erial$port connectors to the protoboard, which interface with the *C


34/39

microcontroller. The prototype sensor trace C' was connected to the #DD84

using short cables, and adhered to the test piece of glass so that meaningful data

could be produced.

'efore the *C could communicate with the #DD84, the *1C

communication system had to be integrated into the programming of the *C. The

software used to program the *C was Microchip"s MHab *#;. The *C can be

programmed using C code, which can be entered into the computer running the

MHab *#; software, and then transferred to the *C using the *C#1

debuggerEprogrammer which connects to the *C using a U5' interface mounted

on the protoboard. The MHab *#; software library includes a .h- code file

containing all of the standard *1C communication commands, such as 5tart-,

5top-, c!-, otc!-, *dle-, etc. Using these commands, the initialiation

seuence described in section @.1 was built in C code and programmed into the

*C. The purpose of this initial testing seuence was to determine if the *1C

communication system was functioning correctly, and the data values were in fact

being written to the #DD84"s registers.

*nitially, the system did not wor!. This was determined by writing a he%

value to a register, and then reading the value of the register. *f they did not match,

then the communication failed. The problem was determined to be with the

55##- register value of the *C, which determines the freuency the *1C

communication line. The system is designed to operate at a standard freuency of

033 !:. This freuency is determined by the he% value written to the 55##register. This register counts down from the programmed he% value twice for every

cycle, and so is dependent on the cloc! freuency of the *C. The datasheet of the

*C had an incorrect formula for determining the value othe 55##. nce this

problem was corrected, the *1C communication system wor!ed flawlessly.


35/39

4.3. FINAL #ROTOT(#E TESTING

nce all the wor!ing code and Bisual 'asic program was complete, the C'

layout was ordered from ;%pressC'. Three identical C's were received,

although one had some of the component pads soldered together and had to be

scrapped. The other two were cut into their respective sies using a fine$tooth saw.

The surface$mount components were then soldered onto the boards, and wires fit

into the terminal bloc!s connecting the upper C' to the lower, and the upper C'

to the blac! bo%- containing the *C09/8413 and laptop$interface euipment. The

lower C' was adhered to the test windshield using the 869$M adhesive.

plastic enclosure from


36/39


37/39

APPLICATIONS

"our wheeler application

DISADVANTAGES

3. This system applied in the case of water falling on the class only.

4. Addition cost is required to install this system to four wheeler.


38/39

5. CONCLUSION

#esign and developed an accurate and cost$effective capacitive rain operated

utomatic wiper system, improving on nearly all of the faults of the current optical

sensor. Testing confirms that it can accurately detect varying levels of rain through

a 6 mm thic! vehicle windshield, and uses this data to turn the wipers to three

different settings depending on the amount of rain present on the windshield.

/urthermore the sensor can differentiate between rain and other ob+ects, such as

leaves and human hands, that a placed above the sensor, thus preventing false

positives and inappropriate wiper operation. The entiproduction$level unit is selfcontained in a compact plastic enclosure mounting near the rear view mirror. This

enclosure is smaller than the optical sensor unit, yet still provides a substantially

larger sensing area for better detection of sparse rain. The system contains only

three integrated circuits anfour passive components, thus allowing for efficient

assembly, low comple%ity, and easy repair. 5light refinements will be needed to

incorporate the capacitive rain sensor into an actual vehicle, as the control signals

to be sent to the vehicleAs 'CM are not implemented at this time, as this was not a

sponsor reuirement. :owever, these control signals can be easily generated by the

on$board *C microcontroller and output to the 'CM, and should reuire little

additional design effort.


39/39

6. REFERENCES

Hee, Mar!. Cypress 5emiconductor Corp. NThe rt of Capacitive Touch

5ensing

'rychta, Michael. NMeasure Capacitive 5ensors &ith 5igma$#elta

Modulator.

/airchild 5emiconductor. NMCD9OOEHMD9OOEMCD9OO @$Terminal 0

ositive Boltage

automatic rain operated wiper reportsfinal

Documents