Department of Electrical Engineering

IIT Delhi

FADING CHANNEL SIMULATOR

(B.Tech Project)

Submitted by

Gaurav Sharma (98185) Mehakdeep Singh (98197)

Under the guidance of

Prof. H.M. Gupta

Certificate

We, Mehakdeep Singh and Gaurav Sharma, are submitting this report detailing work

done during our B.Tech Project in the year 2001-2002. The work done and the results

contained are our own and are genuine and all material taken from other sources (books,

manuals, internet, other thesis, etc.) has been fully acknowledged.

Mehakdeep Singh Gaurav Sharma (98197) (98185)

Date: April 26, 2002

Mr. Gaurav Sharma and Mr. Mehekdeep Singh have worked under my direct supervision

during this semester. I have read this report. It meets my expectations and it accurately

reflects the bonafide work done by the students.

Prof. H.M. Gupta (Supervisor)

Date: April 26, 2002

2

Acknowledgement

We hereby express our gratitude to our supervisor, Prof. H.M. Gupta of the Dept. of

Electrical Engineering, for his guidance in the execution of the project. It was indeed a

great honor to work under such unerring teacher, perfectionist and a person of an ever-

helping nature. We are especially grateful for all the help he provided and the resources

that he made available to us without which the project would not have reached its current

stage.

We would also like to thank our parents whose good wishes and silent blessings always

remained with us throughout the course of the project.

Mehakdeep Singh Gaurav Sharma

3

Table of Contents

Certificate 2

Acknowledgement 3

List of Figures 6

1. Introduction

1.1 Motivation 8

1.2 Problem Definition 9

1.3 Relevant Background 9

2. Theoretical Background

2.1 Multipath Propagation 11

2.2 Multipath Fading 12

2.3 Types of Fading 13

2.3.1 Rayleigh Fading 13

2.3.1.1 Effect of Motion 14

2.3.1.2 Phasor Representation 14

2.3.1.3 Probability Distribution Function 15

2.3.1.4 Power Spectrum 16

2.3.2 Rician Fading 16

2.3.3 Nakagami Fading 18

3. Design of Rayleigh Fading Simulator

3.1 Noise Generation Circuit 20

3.2 Shaping Filter Design 20

3.2.1 Peaking Amplifier Design 21

3.2.2 Low Pass Filter Design 22

3.3 Balanced Modulator Design 23

4

3.3.1 Using LM1496 Chip 23

3.3.2 Using AD630 Chip 24

3.4 RC Phase Shifter 25

3.5 LC Band Pass Filter 26

3.6 Adder using Op-amp RCA3142 27

3.7 PCB Design 28

3.8 Results 31

4. Verification Of The Output

4.1 Demodulation of Output 33

4.1.1 Squarer 33

4.1.2 Low Pass Filter 34

4.2 Sampling using 80196 Processor 34

4.3 Histogram Plotting 34

4.4 Results 35

5. Demonstration of Flooring

5.1 Flooring 36

5.2 PSK Modulation and Demodulation 37

5.3 Determination of BER 43

5.4 Results 43

6. Problems Faced 44

7. Future work 45

Appendix 1 46

Appendix 2 48

References 49

5

LIST OF FIGURES

Fig 1: Typical Multipath Propagation Environment

Fig 2: Multipath Fading, Shadowing and Path Loss phenomena

Fig 3: Phasor diagram of a set of scattered waves (in blue), resulting a Rayleigh-fading

envelope (in black)

Fig 4: Power density spectrum of a sine wave suffering from a Doppler spread

Fig 5: Block Diagram of Rayleigh Fading Simulator

Fig 6: Experimental Setup for noise generation

Fig 7: Experimental Setup for Shaping Filter

Fig 8: Peaking Amplifier Circuit

Fig 9: Low Pass Filter Circuit

Fig 10(a): LM1496 Chip and its Internal Circuit

Fig 10(b): LM1496 Chip and its Internal Circuit

Fig 11: AD630 Chip and its Internal Circuit

Fig 12: RC Phase Shifter

Fig 13: LC Band pass Filter

Fig 14: Op-amp RCA3142 Internal Structure

Fig 15: Complete Circuit Diagram of Rayleigh Simulator

Fig 16: PCB Layout

Fig.17 (a): Shaping filter frequency response

Fig 17(b): Band pass filter frequency response

Fig 18: AD630JN Chip Circuit

6

Fig 19: Low Pass Filter Circuit

Fig 20: Histogram of Output Samples

Fig 21: Comparison of Output with Rayleigh Distribution

Fig 22(a): (C/N) vs. Pe for different severity of Fades to demonstrate Fading

Fig 22(b): Set up to demonstrate Flooring

Fig 23: Set up for PSK Modulation

Fig 24: Set up for PSK Demodulation

7

1. INTRODUCTION

1.1 Motivation

Mobile radio channel simulators are essential for repeatable system tests in the

development, design, or test laboratory. Field tests in a mobile environment are

considerably more expensive and may require permission from regulatory authorities.

Because of the random, uncontrollable nature of the mobile propagation path, it is

difficult to generate repeatable field test results. The Rayleigh Fading Simulator can be

used to test the performance of radios in a mobile environment in the lab, without the

need to perform measurements whilst actually mobile. The mobile fading simulation can

also if required be replicated, and the effects can be varied according to the ‘velocity’ of

the mobile receiver. This allows the comparison of the performance of different receivers

under standardized conditions that would not normally be possible in actual mobile

testing situations.

Network Fading Simulator can also be used for the performance analysis of different

modulation-demodulation schemes. Random binary Sequences are generated and

modulated using the desired modulation scheme (generally QPSK or PSK). This binary

sequence is then detected in the presence of additive white Gaussian noise and

multiplicative noise. This multiplicative noise is often Rayleigh Distributed and is

generated by the network-fading simulator. Thus a plot between the C/N ratio and the Bit

Error Ratio (BER) can be obtained. This plot can be used to demonstrate a well-known

effect called flooring.

8

1.2 Problem Definition

The objectives of our project can be listed as follows:

1. Hardware design of a Rayleigh Fading Channel Simulator

2. Using the Simulator for performance analysis of PSK modulation scheme

3. To demonstrate the effect of Flooring using the PSK modulation and the

Simulator

1.3 Relevant Background

The following courses offered by the Department of Electrical Engineering have helped

us to understand the concepts involved in our project:

1. EE206N: This basic course on Communication Engg. familiarized us with the

various modulation schemes, signal and system representation schemes and the

concept of coherent and non-coherent detection. Moreover, the course provided

an overview on sampling, quantization and A/D conversion. 2. EE330N: This course furthered the concepts taught in EE206N and helped us to

gain insight into the digital modulation schemes like PSK, QPSK, FSK, ASK etc.

An overview of the performance analysis of the various digital schemes and their

detection methods in presence of noise was also given. 3. EE432N: This course on Satellite communications provided a general overview

and technical characteristics of various communication systems, multiple access

techniques, synchronization techniques and reception systems.

9

2. THEORETICAL BACKGROUND

2.1 Multipath Propagation

A typical model of a land mobile radio, including PCS and digital cellular transmission

link, consists of an elevated base-station antenna (or multiple antennas) and a relatively

short distance line-of-sight (LOS) propagation path, followed by many NLOS reflected

propagation paths and a mobile antenna or antennas mounted on the vehicle or more

generally on the transmitter/receiver (T/R) or transceiver of the mobile or portable unit.

In most applications, no complete, direct LOS propagation exists between the base-

station antenna, also known as the access point, and the mobile antenna because of the

natural and constructed obstacles. In such environment the radio transmission path, or

radio link, may be modeled as a randomly varying propagation path. In many instances

there may exist more than one propagation path, and this situation is referred to as

multipath propagation. The propagation path changes with the movement of the mobile

unit, the base unit, and/or the movement of the surroundings and environment.

Even the smallest, slowest movement causes time variable multipath, thus random time

variable signal reception. For example, assume that the cellular user is sitting in an

automobile in a parking lot, near a busy freeway. Although the user is relatively

stationary, part of the environment is moving at 100km/hr. The automobiles on the

freeway become “reflectors” of the radio signals. If during transmission or reception the

user is also moving, (for example, driving at 100km/hr), the randomly reflected signals

vary at a faster rate. The rate of variations of the signal is frequently described as Doppler

Spread.

10

Fig 1: Typical Multipath Propagation Environment

Three partially separable effects known as multipath fading, shadowing, and path loss

characterize radio propagation in such environments.

• Multipath propagation leads to rapid fluctuations of the phase and amplitude of the

signal if the vehicle moves over a distance in the order of a wavelength or more.

Multipath fading thus has a 'small-scale' effect.

• Shadowing is a 'medium-scale' effect: field strength variations occur if the antenna is

displaced over distances larger than a few tens or hundreds of meters.

• The 'large-scale' effects cause the received power to vary gradually due to signal

attenuation determined by the geometry of the path profile in its entirety. This is in

contrast to the local propagation mechanisms, which are determined by terrain

features in the immediate vicinity of the antennas.

• Path loss models describe the signal attenuation between transmit and receive

antenna as a function of the propagation distance and other parameters. Some models

can include many details of the terrain profile to estimate the signal attenuation.

11

The large-scale effects determine a power level averaged over an area of tens or hundreds

of meters and therefore called the 'area-mean' power. Shadowing introduces additional

fluctuations, so the received local-mean power varies around the area-mean. The term

'local-mean' is used to denote the signal level averaged over a few tens of wavelengths,

typically 40 wavelengths. This ensures that the rapid fluctuations of the instantaneous

received power due to multipath effects are largely removed.

2.2 Multipath Fading

Multipath fading is characterized by envelope fading (non-frequency-selective amplitude

distribution), Doppler spread (time selective or time variable random phase noise), and

time-delay spread (variable propagation distance of reflected signals causes time

variations in the reflected signals). These signals cause frequency-selective fades. These

phenomena are summarized in Fig.2.

Doppler Spread is defined as the spectral width of a received carrier when a single

sinusoidal carrier is transmitted through the multipath channel. If a carrier wave (an

unmodulated sinusoidal tone) having a radio frequency fc is transmitted, then because of

Doppler Spread fd we receive a smeared signal spectrum with spectral components

between fc – fd and fc + fd. This effect may be interpreted as a temporal decorrelation

effect of the random multipath-faded channel and is known as time-selective fading. The

effect of Time-delay Spread can be interpreted as a frequency-selective fading effect.

This effect may cause severe waveform distortions in the demodulated signal and may

impose a limit on the bit-error-ratio (BER) performance of high-speed radio systems.

12

Fig 2: Multipath Fading, Shadowing and Path Loss phenomena

2.3 Types of Fading

Many models for the probability distribution function of the signal amplitude exposed to

mobile fading have been given. Out of these models Rayleigh fading, Rician Fading and

Nakagami fading models are most widely used. We will now discuss these three models

in brief and in the following chapters Rayleigh fading will be discussed in detail.

2.3.1 Rayleigh Fading

Rayleigh fading is caused by multipath reception. The mobile antenna receives a large

number, say N, reflected and scattered waves. Because of wave cancellation effects, the

instantaneous received power seen by a moving antenna becomes a random variable,

13

dependent on the location of the antenna. Let us discuss the basic mechanisms of mobile

reception. In case of an unmodulated carrier, the transmitted signal has the form

.

.

2.3.1.1 Effect of Motion

Let the n-th reflected wave with amplitude c_n and phase arrive from an angle

relative to the direction of the motion of the antenna.

The Doppler shift of this wave is

,

where v is the speed of the antenna.

2.3.1.2 Phasor representation

Fig 3: Phasor diagram of a set of scattered waves (in blue),

resulting a Rayleigh-fading envelope (in black)

14

The received unmodulated signal r(t) can be expressed as

An in phase-quadrature representation of the form

can be found with in-phase component

and quadrature phase component

.

Provide that N is sufficiently large and that all the cn are equal, then by the central limit

theorem, I(t) and Q(t) can be shown to be zero mean stationary baseband Gaussian

processes.

2.3.1.3 Probability Distribution Function

If we express r (t) as:

r(t) = R(t) cos2л fct + θ(t)

then, the probability density functions of R and θ are given respectively by

p(R) = R/σ2 exp(-R2/2σ2) R≥0

p(θ) = 1/2л -л≤θ≤л

The preceding equations indicate that the received faded carrier e(t) has a Rayleigh

distributed envelope r(t) and a uniformly distributed phase θ(t).

15

2.3.1.4 Doppler Power Spectrum

Assuming that many waves arrive each with its own random angle of arrival (thus with

its own Doppler shift), which is uniformly distributed within [0, 2 pi], independently of

other waves. This allows us to compute a probability density function of the frequency of

incoming waves. Moreover, we can obtain the Doppler spectrum of the received signal.

This leads to the U-shaped power spectrum for isotropic scattering,

Fig 4: Power density spectrum of a sine wave suffering from a Doppler spread

2.3.2 Rician Fading

The model behind Rician fading is similar to that for Rayleigh fading, except that in

Rician fading a strong dominant component is present. This dominant component can for

instance be the line-of-sight wave. Refined Rician models also consider

• that the dominant wave can be a phasor sum of two or more dominant signals, e.g.

the line-of-sight, plus a ground reflection. This combined signal is then mostly

treated as a deterministic (fully predictable) process, and

• that the dominant wave can also be subject to shadow attenuation. This is a

popular assumption in the modeling of satellite channels.

16

Besides the dominant component, the mobile antenna receives a large number of

reflected and scattered waves.

The derivation is similar to the derivation for Rayleigh fading. In order to obtain the

probability density of the signal amplitude we observe the random processes I(t) and

Q(t) at one particular instant t_0. If the number of scattered waves is sufficiently large,

and is i.i.d., the central limit theorem says that I(t_0) and Q(t_0) are Gaussian, but, due to

the deterministic dominant term, no longer zero mean. Transformation of variables shows

that the amplitude and the phase have the joint pdf

Here, is the local-mean scattered power and is the power of the dominant

component. The pdf of the amplitude is found from the integral

,

where is the modified Bessel function of the first kind and zero order, defined as

17

2.3.3 Nakagami Fading

Nakagami fading occurs for instance for multipath scattering with relatively large delay-

time spreads, with different clusters of reflected waves. Within any one cluster, the

phases of individual reflected waves are random, but the delay times are approximately

equal for all waves. As a result the envelope of each cumulated cluster signal is Rayleigh

distributed. The average time delay is assumed to differ significantly between clusters. If

the delay times also significantly exceed the bit time of a digital link, the different

clusters produce serious intersymbol interference, so the multipath self-interference then

approximates the case of co-channel interference by multiple incoherent Rayleigh-fading

signals. Following are some important facts related to Nakagami Fading.

• If the envelope is Nakagami distributed, the corresponding instantaneous power is

gamma distributed.

• The parameter m is called the 'shape factor' of the Nakagami or the gamma

distribution.

• In the special case m = 1, Rayleigh fading is recovered, with an exponentially

distributed instantaneous power

• For m > 1, the fluctuations of the signal strength reduce compared to Rayleigh

fading.

18

3. DESIGN OF RAYLEIGH FADING SIMULATOR

A detailed implementation concept of the Rayleigh fade simulator is illustrated in Fig 5.

The theoretical foundation and justification for the implementation of such simulators has

been explained already in previous sections. From our derivations and results, we note

that the in-phase i(t) and quadrature phase q(t) baseband control signals must be band-

limited Gaussian sources with a power spectral density W(f). This density is proportional

to:

W(f) = C. 1/( fD2- f2)1/2 |f|≤ fD

= 0 |f|≥ fD

where fD is the maximum Doppler Frequency and C is proportionality constant.

The Rayleigh envelope statistics are obtained by adding two noise sources in quadrature.

The theoretical spectrum of the received signal is approximated by shaping the spectrum

of the noise sources with filters.

19

Fig 5: Block Diagram of Rayleigh Fading Simulator

3.1 Noise Generation Circuit

Experimental setup for the generation of noise sources is given in Fig6. The zener diode

used was Z122. It had breakdown voltage of 8.7V. A voltage of 12V was applied to the

zener diode through a resistance of 20KΩ. The noise so obtained was amplified using an

operational amplifier circuit. This noise was then fed to the shaping filter to appropriately

shape its spectrum.

Fig 6: Experimental Setup for noise generation

20

3.2 Shaping Filter Design

Fig 7: Experimental Setup for Shaping Filter

Shaping filter circuit shown in Fig 7 can be divided into two parts:

• Low pass filter

• Peaking amplifier

3.2.1 Peaking Amplifier Design

Fig 8: Peaking Amplifier Circuit

The peaking amplifier can be represented as a Twin-T network as shown in Fig 8 and its

transfer function is given by:

21

=iV

V0

1)()(1)()(

331212

213232313131213132213

321321

2132

213213

321321

++++++++++++++

sCRCRCRsCCRRCCRRCCRRCCRRCCRRsCCCRRRsCCRsCCRRRsCCCRRR

In our case,

R1=R2=R3/2=R and C1=C2=2C3=C

=iV

V0

R3C3s3+R2C2s2+RCs+1 R3C3s3+5R2C2s2+7/2RCs+1

The desired peaking frequency was 100KHz.

The peak frequency is given by the formula, fp = 1/(2лRC)

We chose the value of C = 1nF, for this value of C the value of R obtained from the

above formula is 1.59KΩ.

From these values of R and C we obtained the values of R1, R2, R3, C1, C2 and C3.

3.2.2 Low Pass Filter Design

Experimental setup for the low pass filter is given in Fig 9. The low pass filter chosen

was Butter worth low pass filter of second order. The reason why we chose Butter worth

filters is that they offer the flattest pass band and provide a fast initial falloff and

reasonable overshoot. The design procedure is given below:

1. The corner frequency (fc) was chosen to be 100KHz.

2. The value of C was chosen to be 1nF.

3. R2 was calculated as, R2 = 1/(2πfcC) = 1.6 KΩ.

22

4. R1 was calculated as R1 = XR2, where X is dependent on the gain that you desire. The

gain was chosen to be 2 and for this value of gain X was obtained to be 0.5. This

gives you the value of R1 = 800 Ω.

Fig 9: Low Pass Filter Circuit

3.3 Balanced Modulator

The Balanced Modulation has been tried using two different circuits:

• Using LM1496 chip

• Using AD630 chip

3.3.1 Balanced Modulator Design using LM1496

The chip circuit is shown in Fig 10(a) and Fig 10(b). The LM1496 is a double balanced

modulator, which produce an output voltage proportional to the product of an input

(signal) voltage and a switching (carrier) signal. The LM1496 has adjustable gain and

signal handling, fully balanced inputs and outputs, low offset and drift and a wide

frequency response up to 100 MHz. The input to Pins 8 and 10 is the Carrier Input and to

Pins 1 and 4 is the Modulator Input. Pins 2 and 3 are for gain adjustment and Pin 14 is

connected to –V. The output is obtained on Pins 6 and 12. The internal structure of the

pin is also shown in Fig 10.

23

Fig 10(a): LM1496 Chip and its Internal Circuit

Fig 10(b): LM1496 Chip and its Internal Circuit

3.3.2 Balanced Modulator Design using AD630JN

The chip circuit is shown in Fig 11. The AD630JN is a high precision balanced

modulator that combines a flexible commutating architecture with the accuracy and

temperature stability afforded by laser wafer trimmed thin-film resistors. AD630JN is a

24

20-pin chip. The Modulation input is given to Pin 16 and the Carrier Input to Pin 9. Pins

2 and 3 and Pins 5 and 6 are for gain adjustment. The output is obtained on Pin 13. The

internal structure of the pin is also shown in Fig 11.

The configuration of AD630JN makes it ideal for signal processing applications such as

balanced modulation and demodulation. The 100dB dynamic range of AD630JN exceeds

that of any IC Balanced Modulator and is comparable to that of costly signal processing

instruments. Moreover, the op amp format of AD630JN ensures easy implementation of

high gain applications with no additional parts.

Fig 11: AD630 Chip and its Internal Circuit

25

3.4 RC Phase Shifter

To obtain the required Rayleigh Fading spectrum, the two noise sources have to be added

in quadrature. This shift of 90 degrees between two RF sources is obtained using the

circuit shown in Fig 12. Here the idea is to make the impedance of capacitor and

resistance same at the desired frequency of operation.

f = 1/ (2 ΠRC) = 1 MHz

C = 1nF => R = 160 Ω

Fig 12: RC Phase Shifter

3.5 LC Band Pass Filter

The basic circuit of LC bandpass filter is shown in Fig 13. It consists of two transistor

stages. The basic idea is that LC circuit shows very high impedance at resonance and

small resistance otherwise. Since the ac gain of the common collector configuration is

proportional to the resistance put at the collector, so by putting a LC circuit there we can

26

get a peaking at the desired frequency. The capacitors used in the two stages were both

variable capacitors. By properly adjusting the value of the capacitors you can make the

two circuits to peak at slightly different frequencies. This makes the transfer function

almost constant in the in between region and falls sharply after that. Hence a bandpass

filter close to ideal can be obtained. The reason why we did not use an active RC

bandpass filter is that in this case the resonant frequency is quite a complicated function

of all the resistances and capacitances. And also it is very sensitive to changes in any of

the component values. Slight changes in the component value can alter the passband and

center frequency by a large value.

Fig 13: LC Band pass Filter

3.6 Adder using Op-amp RCA3140

The two noise sources produced in quadrature cannot be added using the normal

operational amplifier HA17741 due to its low bandwidth. HA17741 adds a lot of noise

and gives distorted output at high frequencies. To avoid this problem, RCA3142 chip is

used. Its pin structure is shown in Fig 14. RCA3142 is a wideband monolithic amplifier

with improved speed and stability and low power consumption. It has a bandwidth of

10MHz and thus is ideally suited for our application. RCA3142 has the same number of

27

pins as HA17741 (8pins). Pins 7 and 4 are connected to +12V and –12V respectively.

Pins 2 and 3 are the input pins and Pin7 is the output pin.

Fig 14: Op-amp RCA3142 Internal Structure

3.7 PCB Design

Initially all the components of the Rayleigh fading channel simulator (as described

above) were made on the bred board. The complete circuit diagram is shown in Fig 15.

After verifying that all the components were working as desired we made a PCB of the

complete circuit. The PCB layout is given in Fig 16. Following were the additional

features of the PCB:

• The noise produced by the two zener diode is generally not equal in amplitude.

This noise is then passed through the amplifier, low pass filter and peaking

amplifier. The response of these can also be slightly different in the two cases.

But before combining the two noise sources we must make sure that their

amplitudes are equal. To achieve this we added an additional variable gain stage

in each of the two circuits. By varying the gain of this stage we can make the

amplitudes of the two noise sources equal.

28

• Finally the summer, which sums the outputs of two AD630JN chips, was also

provided with variable gain. By adjusting the gain of the adder, we could change

the overall output of the Rayleigh Fading Simulator.

• Some additional capacitors were added to reduce the effect of interference. And

the results were quite good, with all these capacitors and some other design

features interference was almost removed at the output.

29

30

Fig 16: PCB Layout

31

3.8 Results

The following were the results obtained:

• The power spectrum of the noise generated by the zener diode was found to be

flat over the entire frequency range of 10Hz to 1MHz. After passing the noise

through the amplifier circuit it was found that the higher frequency components

were slightly attenuated. This was expected because the operational amplifiers

used had a finite bandwidth.

• The shaping filters frequency response obtained is shown in Fig 17(a). Here the

value on the y-axis is the normalized gain obtained as:

Gn(f) = G(f) / G(0)

where Gn(f) is the normalized gain at frequency f Hz, G(f) is the actual gain at

frequency f Hz and G(0) is the gain at 0 Hz.

Normalized Gain vs Frequency

-25

-20

-15

-10

-5

0

5

10

15

0 50 100 150 200 250

Frequency (in KHz)

Nor

mal

ized

Gai

n (in

dB

)

Fig.17 (a): Shaping filter frequency response

The response above is the combined response of the low pass filter as well as the

peaking amplifier.

32

• The desired phase shift of 90 degrees was obtained between the two outputs

taken from the phase shifter.

• The response of the Band pass filter was found to be flat for frequency range

900KHz to 1.1MHz and response was quite small for frequencies below

500KHz and above 1.5MHz that is what was required. This is shown in Fig

17(b).

00.5

11.5

22.5

33.5

4

0 500 1000 1500 2000 2500

FREQUENCY (IN KHz)

GA

IN

Fig 17(b): Band pass filter frequency response

The above results were also obtained on the PCB and were found to be the same.

However, by the visual inspection of the output one could not verify whether the

amplitude was Rayleigh distributed or not. So for that we demodulated the output and

plotted the histogram of the output samples. This is illustrated in the next section.

33

4. VERIFICATION OF THE OUTPUT

4.1 Demodulation of Output

Since the output of the Rayleigh Fading channel simulator does not have any carrier and

has random phase component this makes the use of envelope detector and coherent

demodulation schemes impossible. So the only way it can be demodulated is using a

squarer followed by a low pass filter. The output of the squarer will be a term

proportional to the square of the output amplitude and some high frequency component

centered on twice the carrier frequency. The low pass filter removes the high frequency

component and gives the output proportional to the square of the received signal.

4.1.1 Squarer

The squarer is implemented using AD630JN chip. The output of the Fading Channel

Simulator is applied to the carrier as well as the modulating input. The output obtained is

fed to the low pass filter.

Fig 18: AD630JN Chip Circuit

34

4.1.2 Low Pass Filter

The low pass filter used was the same as the one used in the shaping filter. The cutoff

frequency was 100 KHz.

Fig 19: Low Pass Filter Circuit

4.2 Sampling using 80196 Processor

The demodulated output was fed to the input of 80196 microprocessor. This analog input

was then sampled and the corresponding 8-bit samples were stored in memory. The

program for Sampling followed by A/D conversion is given in Appendix 1.

4.3 Histogram Plotting

The above-mentioned program stores the histogram of the output samples in memory

locations. These values were then manually read and written into a file called input.txt.

The histogram of the output samples was then plotted and compared with Rayleigh

distributions with varying σ. The Matlab program for this is given in Appendix 2.

35

4.4 Results

The histogram of the output samples is shown in the Fig 20. The histogram was than

compared with Rayleigh distribution for different values of σ as shown in Fig 21. It was

found that the output was closest to Rayleigh distribution with σ equal to 0.8.

-2000

0

2000

4000

6000

8000

10000

12000

14000

16000

0 0.5 1 1.5 2 2.

Demodulated Output (in V)

No.

of s

ampl

es

5

Fig 20: Histogram of Output Samples

0 0.5 1 1.5 2 2.50

2000

4000

6000

8000

10000

12000

14000

16000

18000

VOLTAGE

NO

. OF

SA

MP

LES

Rayleigh (sigma = 0.7) Rayleigh (sigma = 0.8) Rayleigh (sigma = 0.9) Actual Output

Fig 21: Comparison of Output with Rayleigh Distribution

36

5. DEMONSTRATION OF FLOORING

5.1 Flooring

After verification of the output, our next objective is to demonstrate flooring. For that we

will be we will be generating a pseudo random binary sequence and modulating and

demodulating it in the presence of AWGN and multiplicative noise generated through our

fading channel simulator. For a given severity of the multiplicative noise as you keep on

decreasing the AWGN what you observe is that bit error rate does not fall after some time

as shown in Fig. 22(a). This phenomenon is called flooring. Different curves are obtained

for different severities of the fade. The Block Diagram of the circuit to be used to

demonstrate the Flooring effect is shown in Fig 22(b).

Fig 22(a): (C/N) vs. Pe for different severity of Fades to demonstrate Fading

Fig 22(b): Set up to demonstrate Flooring

37

The modulation scheme used to demonstrate Flooring is PSK and its explanation is given

below.

5.2 PSK Modulation and Demodulation

Phase-shift keying (PSK) means altering the phase of a (usually sinusoidal) carrier signal

with respect to a reference phase, in accordance with the value of the base-band signal.

PSK, to an extent often greater than with frequency-shift keying, has the advantage that

amplitude variations can be suppressed by limiting. It is also beneficial in reducing noise.

However it is more susceptible to sudden changes in transmission delay of a channel,

such as can happen with radio links, and the modulation and demodulation processes tend

to be more complex. The amount of phase-shift can for a binary (two-valued) baseband

signal can be ±90 or it can be less.

In this experiment modulation will be achieved by adding, to a constant carrier by phasor

C, a quadrature component represented by Q that may be reversed in phase to give

resultant signal of phase ±ø.

The method of demodulation to be used must depend on reconstructing an equivalent of

C. as a reference phase, with which the phase of the received signal can be compared.

This is easily done if the data format is such as to give equal periods of 0 and 1 signal in a

short interval, such as bi-phase code. It is then only necessary to use a phase lock loop

whose action is slow enough to hold the oscillator at the mean of the two signal phases.

38

For the PSK modulation and demodulation, the kit available in the lab was used and it

had the following components:

1 DCS297A Data Source

1 DCS297B Data Format

2 DCS297C Double Balanced Modulators

1 DCS297D Carrier Phase-shifter

1 DCS297E Voltage-Controlled Oscillator

1 DCS297F Data Clock Regeneration

1 DCS297G Data Recovery

1 DCS297H Data Receiver

1 DCS297K Audio Module

1 DCS297L Tuned Circuit

1 DCS297M Power Supply

1 DCS297N set of connecting leads

Dual-beam oscilloscope

Modulation

Connect the equipment as shown in Fig 23. Note that the bi-phase data waveform from

the Data Format module DCS297B always has the same dc component, averaged over

any one bit-time. This can be verified by measuring its dc component with a voltmeter.

Because of this, the capacitor-feed to the modulator can reject the dc and produce equal

positive and negative inputs to the modulator proper. Check this with the oscilloscope at

input b. Use the oscilloscope to verify that the Carrier Phase-shifter DCS297 D produces

39

carrier waveforms in mutual quadrature at links 10 and 12. The associated gain control

may be turned fully clockwise and the phase control should be adjusted initially to give

approximately equal output voltages. These two voltages are fed respectively to two

modulators. The lower modulator has a constant bias as its second input, so that it’s

output corresponds to the C phasor. The bias value controls the magnitude (and sign) of

this output. It should be set to a positive value by turning the control clockwise. Link 10

supplies the quadrature carrier, which is reversed in phase as the modulating signal (link

9) changes sign at the 'b' terminal. The output currents of the two modulators are

combined in the common load, producing the phase-modulated signal. The phase-

modulated signal can be examined as follows:

• Increase the time base speed to say 2 microseconds per division.

• Synchronize (using the external sync/trigger connection for later convenience) to

the 1.28MHz carrier, link 6, to which also the Y1 channel should be connected.

• Finally display the output, link 14, on Y2.

This will show the two output phases superimposed. The amount of modulation can be

varied by adjusting the 'phase' control. It should be set less then ±90.

Demodulation

Without disturbing the equipment already set up connect up the remainder as shown in

Fig 24. The links 14 and 15 are the communication links and are therefore the same as the

links 14 and 15 shown in Fig 23. With a signal on link 14, transfer the Y1 lead (but not

the sync/trigger lead) to the VCO module, link 16. This should show the carrier recovered

from the signal by the phase-lock loop. It will vary in phase somewhat, but not much,

40

because, although the loop tries to lock it to the changing phase of the incoming signal, it

is slow acting. Note that its phase is in quadrature with the mean phase of the received

signal. The latter's quadrature components (Q in fig 3.8) are therefore at 0° or 180° to it.

A sinusoid multiplied by another of the same phase produces a dc component in the

output:

(2sin2ω= 1 - cos2 ω t)

When one of them is shifted 180° this component changes sign. The modulator output on

links 20, 21 therefore contains a component that represents the original bi-phase data.

The shunt capacitor passes most of the component of frequency 2ω t, keeping the ripple

voltage small. Check this with the oscilloscope, Y1 showing the bi-phase data link 9, and

Y2 the recovered data link 20. The settings should be restored to the usual:

• Y1 and Y2; dc-coupled

• 5V/division time base

• 10µs/division, externally triggered by +ve going edge

41

Fig 23: Set up for PSK Modulation

42

Fig 24: Set up for PSK Demodulation

43

5.3 Determination of BER

In our case the delay between the input stream, i.e. the sequence of bits generated by the

pseudo random generator and the output stream (demodulated bit stream) was quite less.

So we can simply use a XOR gate followed by a counter to count the number of errors.

Obviously the setup should be run for small enough time so that the counter does not

overflow.

5.4 Results

• Initially we were making our own PRBS generator using LFSR’s (Linear

Feedback Shift Registers). But the PSK modulation and demodulation kit

available in the lad had its own PRBS generator. So we used the PRBS generator

available with the kit.

• Since some of the modules in the PSK demodulator kit were not working, so we

could not set up the entire circuit as intended. Right now we are working on how

we can get the signal, which will be obtained, at the receiver side in case of fading

channel. For this we are trying to multiply the demodulated output of our

Rayleigh Fading channel simulator with the modulated signal.

44

6. PROBLEMS FACED

• The desired frequency of operation was 10.7 MHz, but the operational amplifiers

available had maximum possible bandwidth of 5 MHz.

• At frequencies in the range of few MHz or higher there was lot of interference

due to radiation. We used all the probes available in the lab, but still interference

could not be avoided. This was the reason why we could not properly test the

output of the balanced modulator chip and the phase shifter.

• The spectrum analyzer available in the lab is capable of showing the frequency

spectrum only up to 50 KHz. After that we had to find the amplitude response

using CRO to obtain the power spectrum density.

• There was only one RF source available in the lab, that was capable of working

up to 20 MHz. But the output of the source was distorted in the frequency range

of about 5 MHz to 14 MHz.

Because of all the above-mentioned problems we shifted our frequency of operation

down to 1 MHz. In the case of bred board there was a lot of interference due to radiation

etc. even at this frequency. But in the case of PCB this interference was almost zero.

Another problem was that the demodulator part of the PSK modulator and demodulator

kit available in the optical communications lab was not working.

45

7. FUTURE WORK

1. The power spectrum of the output can be observed using the new power spectrum

analyzers, which are going to be available in the communications lab soon. These

new spectrum analyzers can work from a frequency of 400KHz to 1GHz as

compared to the previous spectrum analyzers that work only up to 50KHz.

2. BER detection can be done pretty easily, with the BER detection equipment,

which will be available in one or two months in the optical communications lab.

3. Rician fading can also be obtained with slight modification in the circuit.

46

APPENDIX 1

Program for A/D Conversion and Sampling using 80196 microprocessor

ORG 4000H AX EQU 1CH AL EQU 1CH AH EQU 1DH BX EQU 1EH BL EQU 1EH ES EQU 24H AD_COM EQU 02H ADRESL EQU 02H ADRESH EQU 03H USERADC EQU 0D2H INTPEND EQU 09H INTMASK EQU 08H PBCD_BCD EQU 22F0H LED_DISP EQU 21B0H MAIN: LD USERADC, #4100H CLRB INTPEND LDB INTMASK, #02H EI LDB AD_COM, #18H LOOP: SCALL AD_DIS SJMP LOOP ORG 4100H AD_INT: LDB 38H, ADRESL LDB 39H, ADRESH LDB 3AH, 38H ANDB 3AH, #07H ANDB 38H, #0C0H SHRB 38H, #06H EI LDB AD_COM, #18H RET AD_DIS:

47

LDB 0EFH, #0AH LDB 0EEH, #0DH LDB 0EDH, #14H LDB AH, 39H LD BX, #6000H MUL AH, #02H ADD AX, BX MOV BX, AX LD AX, [BX] INC AX ST AX, [BX] RET END

Explanation of the Program: The program samples the analog input available at Channel

1 of Connector C6 at 10 KHz and converts it into a digital output of 8 bits. The number of

samples corresponding to each digitized value is stored in memory locations starting from

6000H as a 16-bit value. The total number of samples taken is 10,00,000.

48

APPENDIX 2

Matlab Program for comparison of Fading Channel Simulator Output with Standard

Rayleigh Distribution

clear; fid = fopen(‘input.txt’,’r’); for I = 1:256 s = fgetl(fid); a1(I,1) = str2num(s); end fclose(fid); a2 = ray (0.7); a3 = ray (0.8); a4 = ray (0.9); for i =1:256 b(i,1) = sqrt((i-0.5)/256); end plot(b,a2,’r’,b,a3,’g’,b,a4,’b’,b,a1,’k’); function out= ray(sigma) for I=1:256 out(I,1)= uint16(1000000*(raylcdf((I-1)/256,sigma)-raylcdf((i/256),sigma))); end

49

50

REFERENCES

1. ‘Radio Propagation and Cellular Engineering Concepts’ by K. Feher.

2. ‘A Multipath Fading Simulator for Mobile Radio’ by Gaston A. Arredondo and Advert

H. Walker- IEEE Transactions on Vehicular Technology, Vol. VT-22, No. 4, Nov.

1973.

3. ‘A Computer Generated Multipath Fading Simulation for Mobile Radio’ by John I.

Smith- IEEE Transactions on Vehicular Technology, Vol. VT-24, No. 3, Aug. 1975.

4. ‘Mobile Radio Communications’ by R. Steele and Hanzo – IEEE Press.

5. http://www.engr.sjsu.edu/filt175_s01/sp2001/proj_sp01c/mai_trang/Filt_pass_last.htm

6. http://dbserv.maxim-ic.com/appnotes.cfm?appnote_number=700

7. http://www.tele-ip.com/fadesimulator.html

8

9

. T. Eyceoz, A. Duel-Hallen and H. Hallen, Deterministic Channel Modeling and Long

Range Prediction of Fast Fading Mobile Radio Channels, IEEE Communications

Letters, Vol. 2, No. 9, pp. 254-256, September 1998.

. A Systematic Approach to the Design and Analysis of Optimum DPSK Receivers for

Generalized Diversity Communications over Rayleigh Fading Channels, IEEE

Transactions On Communications, Vol. 47, No. 9, September 1999 1365, Mahesh K.

Varanasi, Senior Member, IEEE.