DiBpersion effects In frequency windows of millimeter range radiowaves
Pelix V.Kivva and Yuri M.Galaev
Institute of Radiophysica and Blectronics, Academy of Sciences of Ukraine, Kharkov 310085,Ukraine

ABSTRACT

Measurements of wide band signal propagation peculiarities have been carried out in 37 GHz range under different season and meteorologic
conditions on 13 km long direct visibility path over rugged country near Kharkov (Ukraine). Dispersion effects conditioned by direct and indirect meteorologic and season factors have been discovered and measured. A method which allows measurements of differencephaee shifts
in received signals has been developed. Quantitative evaluation of atmospheric communication channel coherence band variation has been
carried out.

INTRODUCTION

In development of modern radio engineering systems and radiolocation wide band signals with band width of 1 GHz and more may be used
such systems, along with waveguiding and cable signal propagation lines, open atmospheric lines in which wide band signal propagates in turbulent atmosphere over the rough division surface are also used. In such lines signal power losses may occur, and also Its coherence losses, which lead to limitations of response, accuracy and resolution of radio
engineering systems.

Additional limitations of coherence band may arise because of the

dispersion in communication channel (dependence of the refraction fac-

tor Ti from the frequency &Y). Dispersion characteristics of the open

communication lines operating at frequencies corresponding to atmosphenc gases molecular absorption bands'3 are studied quite well. Some

peculiarities connected with hydrometeors dispersion characteristics

are also considered

As a rule, available results show weak fre-'

quency dependence of the value fl in the atmosphere "frequency win

dows"6.

The aim of the work is development of the phase invariant method applied to the problem of wide band signals of millimeter range propaga-' tion in real media and measurements of the coherence band limitations arising from meteorological and season factors.

1.MEASURKMENT METHOD AND EQUIPMENT

The developed method is based on the measurement of difference-phase

characteristics in discrete spectrum of wide band signal ?,8,9 •

A

spectrum of such a signal may be formed by frequency multiplication or

796 / SPIE Vol. 1968 Atmospheric Propagation and Remote Sensing 11(1993)

O-8194-1204-X/93/$6.OO

Downloaded From: http://proceedings.spiedigitallibrary.org/ on 07/03/2016 Terms of Use: http://spiedigitallibrary.org/ss/TermsOfUse.aspx

by carrier frequency 2i, amplitude or angular modulation. An important

feature of the aignal iB a 8trict link between single discrete compo—

nents of its spectrum. The essence of the method is the following. The

sensing modulated signal with the carrier frequency & and with

frequencies of the lower and upper side components W &-fl. and &' &2=&,+nJ, reepectlvely,( CL is modulation frequency) is radiated into

the channel under investigation. While propagating each signal

spectrum component receives the following phase increment

=n(&)zct

(1)

which are proportional to both their frequencies &) , a radio channel

length 2 , and propagation medium dispersion characteristics /z (2&).

It follows from (1)that under the condition of radiowaves propagat I on in

dispersion free medium ( tz(2il)= const) the dependence '°(ziY) is frequency

linear function, and in dispersion media (t(2ef j const) it is frequency

non—linear function. shape of the signal

6Exaamreplsehsowonf

such dependences and spectrum S(&r) in Fig.l. It is obvious that segments

from Y quency

saexgmi8enctosrrfersopmonXdiangxisto

fre-
(that

ie

ensured by the modulation law) are

equal to

and 9224Oo . In

dispere ion free medium they are
equal to each other, but in dis-

ço2

—— — — — —

persion media they are not. Thus

the measure of unequality of the segments is dispersion measure in
radiowaves propagation channel

ço
i

which may be written in the form

1

of difference of the segments

2y

- &r

—

fV(oP-L7oJ

L ko- w
V2 'ol

'-' ço0 -1w +'u2'

Pig.1.Examplea of phase-frequen cy characteristics of cozmnunication lines and sensing signal
spectrum.

In 8,9 such combination of spectrum components phases of modulated

signal was called "phase invariant" because it does not vary in space

and time, provided that the signal propagates in dispersion free medium.

In case of dispersion medium

conat and

depend on dispersion

value and , and sign ofPis determined by the law (character) of
dispersion value variation (normal and abnormal). To measure the value

is of içO at the receiving end of the radiowaves propagation channel the

carrier oscillation of the received sensing signal 6

separately

multiplied with signals of each side frequency. Further on phase shift

between the results of multiplication having difference frequencies is

measured. This is the value of phase invariant P.

To carry out the investigation transmitting and receiving equipment applying measurement method in 8 i radiowave range has been developed
and produced. Transmitting device carrier frequency 18 37 GHz, modula-
tion frequency is 0.5 GHz. Transmitting device output power is 70 mW.

SPIE Vol. 1968 Atmospheric Propagation and Remote Sensing 11(1993) / 797 Downloaded From: http://proceedings.spiedigitallibrary.org/ on 07/03/2016 Terms of Use: http://spiedigitallibrary.org/ss/TermsOfUse.aspx

To reduce equipment error during the measurementB a special attention

hae been payed to the tuning, correction and calibration of pha8e-fre—

quency and phase-amplitude characteristics of the measuring equipment

unite. Tran8mitting device and main unit6 of receiving device have been

thermoatated. During the mea8urementa the equipment wae calibrated by

the test signal with controlled parameters analogous to the sensing

signal spectrum. The resulting root—mean-square equipment error of the value çP measurements has not exceed 0.5 • The receiving device anten-

na was mounted at an altitude transmitting device antenna -

aoft

30 an

m above altitude

the earth of 12 in.

surface, and the Both antennae

were identical with the parabolic reflectors diameter of 1.1 m• The

measurement equipment included meteorological equipment and rain inten-

sity measuring device.

Experimental communication line has been selected in terms of the following conditions: sufficient extent modelling real lines, absence of considerable reflections from the earth surface, possibility of direct radio visibility under any meteorological conditions. The communication line with the length of 13 km over the hilly terrain has been chosen. The average altitude of the radiowaves propagation channel was 50 m over the earth surface. Measurements of the vertical field struc— ture carried out under different meteorological conditions in location of receiving device and calculation of the field structure made in approximation of the wave diffraction in the zone on tops of the hills have shown that results of the experiment and the calculation agree well. That allow to estimate possible diffraction errors of the value
z0 measurement with the change of refraction condti8ns. Calculati— one showed that the errors did not exceed values of — 1

After propagation in the communication line the radiated sensing

signal was received by the receiving device, in which it was processed

according to selected measuring method. Simultaneously the intensity of

spectrum side components 7 and

of the received signal spectrum

was measured. The measured values of L\4 , 7, , and '2 were directed

to the recorder. Calibration of the equipment and measurement of meteo—

rological elements were carried out not less than one time per an hour

of observations. The rain intensity was estimated by means of a cup—

-shaped rain gauge and by values of signal attenuation in the precipi—

tations. The experiments were carried out for one year under different

meteorological conditions: clear weather, rain, snow, etc. Clear

weather conditions were considered the cases when no hydrometeors fall—

—out was observed in the radiowaves propagation channel during the

experiment. Measurements results obtained as a rule during one day

(24. hours) when one or another atmospheric phenomenon was observed were

considered one experiment. The measurement results procession was

() is carried out to obtain the following temporal characteristics:

4

daily variation of the average phase invariant value,

j 4

çodsii.,

is season
value.

average

daily

variation

of

the

phase

invariant

gD5reompueipnneuddetenacciecnsotreodrfivna1gl9sjt(ooZf)thPoe(b)staeraiensaeoldnizaaacntcdioorfnodsirn•egaDcthoepotefhnedtehanevcmeerstahgeindg(e)preewnsedurelentcseof

798 / SPIE Vol. 1968 Atmospheric Propagation and Remote Sensing 11(1993) Downloaded From: http://proceedings.spiedigitallibrary.org/ on 07/03/2016 Terms of Use: http://spiedigitallibrary.org/ss/TermsOfUse.aspx

(z) was calculated. The calculations were conducted according to the following algorithm
LPjd ()

where 17? is a number of experiments during the season, z is time.

Total number of experiments is 237. Total time of the continuous
measurements of i9 values is 2186 hours. Conducted experimental
investigations have revealed differences in influence of various atmos—
pheric phenomena on phase invariant value and variability.

2 . MEASUREMENTS RESULTS

In clear weather in different year seasons the temporal realizations
of °(z) presented weakly fluctuating dependences with smooth average
diurnal variation. The values of MP measured in summer were as a rule negative, while measured in winter they were positive. In spring and autumn the change of the sign of measured value was observed. Statistic processing of the observation results has shown availability of regular variations of iP values. Dependences calculated for the day time are
presented in Fig.2. Dependencea corresponding to winter, spring, autumn and summer seasons are marked in the figure by digits 1,2,3 and 4. The values of confi— dea dence intervals determined for the probabi-. lity estimation of 0.95 are marked by vertical strokes. The characteristic feature of cis (z) variations is presence of extrema in the interval between 2 p.m. and
£4. p.m.

The experimental investigations allowed to reveal diurnal and seasonal variation of
the phase invariant values, which in terms of selected measuring method is a parameter

characterizing dispersion value and varia-

tion law. The analysis of measurements re-

sults showed that such phenomena as diaper-

sion in atmospheric gases, atmosphere tur—
bulence, diffraction on the underlayer obstacles under conditions of variable

refraction did not allow to obtain a satis-

factory qualitative and quantitative explanation of the experimentaireaulta. Thus, it
was experimentally estimated that in the

lOq,j2 2 4 6 (h)

atmosphere "frequency windows" in millimeter wave range the near-surface atmosphere layer showed the properties of a regular
datiisopnerlsaiwonofmetdhieumd.isTpheersviaolnuehaasnddituhrenavlan-

Pig.2. Average seasonal diurnal measurements of
phase invariant.

and seasonal variation.

SPIE Vol. 1968 AtmosphericPropagation and Remote Sensing 11(1993) / 799 Downloaded From: http://proceedings.spiedigitallibrary.org/ on 07/03/2016 Terms of Use: http://spiedigitallibrary.org/ss/TermsOfUse.aspx

The rain fall-out during the warm

season (summer and beginnning of the

autumn) caused a considerable

2i0n8•r.e.a3s0e0•ofChtaheraicPtevriasltiucesmeuapsurteoment

results obtained under such conditi-

one are presented in Pig.3. Time in

hours is plotted on x axis. The

measurement results in two different

experiments are shown by solid and

dotted lines respectively. The dura-

tion of rainfalls and their temporal

position are marked by rectangles.

Arrows indicate the instants of the

strongest attenuation of the signal

received, caused by rains• Pigures Plg.3. Measurement re8ults of

by the arrows correspond to the

phase invariant during rain fall—

attenuation values (in dB) measured -out.

at those instants. The characteris—

tic feature of the data was a considerable up to 50 mInutes, delay of

the ço value increase as to the beginnning of the rain. In case of

"short" rain (Pig.2, dotted line) the increase of the XP value was

observed after it ceased. The decrease of the A'P value down to initial

level was gradual and lasted for several hours. The experiments results

showed that the presence of rain drops in radiowaves propagation chan—

nel was not the direct reason for the

value increase.

The measurements made during intensive long snowfalls showed that if

the snowfalls were caused by the motion of a warm atmospheric front

then diurnal zçP variations practically did not differ from the same

dependence measured in dear winter weather (Fig.2, curve 1). Such

experiments showed that the presence of snow floce in the radiowaves

propagation channel did not have strong affect on the

value typi—

cal for the winter season. But, provided that the snowfall was0caused

binycrtehaesecoolfd

atmospheric front motion, a considerable, up to 30 , the measured value was observed. The result of an experi—

ment carried out under such conditions is given in Pig.k. The figure is

accompanied by the results of synchronous air temperature measurements (Pig .4 , dott ed line ). Snow falled continuously during the whole experl— ment (12 hours). Time intervals when snowfall was the most intensive

are marked in the figure by rectangles. It is obvious that the beginn—

ing of the ° value increase coincided the beginning of the air

temperature decrease, which later falled down to —12 C.

The considered results of experimental investigations present typi—

cal time variations of phase invariant observed under different meteQ—

rological and season conditions. But data about extremum values of

phase invariant observed during the experiments are also undoubtedly

useful. Such results are not representative but they outline boundaries

of the measured value and show the conditions in which they were obser-

6' ved. The esult of the experiment in which maximum positive value

4 9' =

was observed is presented by the solid line in Pig.5. The

result was obtained at the end of autumn when cold atmospheric front

800 / SPIE Vol. 1968 Atmospheric Propagation and Remote Sensing 11(1993) Downloaded From: http://proceedings.spiedigitallibrary.org/ on 07/03/2016 Terms of Use: http://spiedigitallibrary.org/ss/TermsOfUse.aspx

Toe

accompanied by rain and snow passed.

In the same figure the result of

2

the experiment in which maxmum

negative value of 'P =

was

0

observed is given by the dotted line. The experiment was carried

out at the end of spring in clear

-2

cloudless days with steady hot

weather. During the experiment at0

—4

3.20 p.m. the temperature was +30 C. The wid velocity did not exceed

-6

7a,rz 'PM 0

2 m/s , but from 11 • 00 a .m • to
5 7 9t(fz) 1.00 p.m. the incerase of1the wind velocity up to 5...? rn/B was

Pig.L1.Measurement results of phase observed.

invariant while passing cold

atmospheric front.

Thus, the essential influence of

different atmospheric phenomena on the value and variation law of dispersion found is shown experimenal-

ly.0The range of experimentally obtained z9 values was from -2k to

+64W

3.FREQUENCY BAND CHARACTERISTICS OP CO1MUNICATION LINES

In paper it WaS shown that the envelope of rectangular radio pulse at the output of the dispersion radio channel can be calculated
by the following relation

[s(a)-s(a_!)]2}2, (2)

where C(U) and S('U)are Prenel integrals, U (—zc,1) z- dimen-

sionlees
pulse,

opairsampeutlesre,amCpliistugdreouapt

velocity, 2 the channel

is duration output, T,

of initial signal

detection time at the channel output.

In expression (2) the value Z is determined as follows

'°L / (7- T'ycZ.,—' '1 I'LY)

-1

(3)

where 20;;, is radio pulse carrier frequency.

In expression (3)(w-)is a function characterising the law of dis—

persion variation in the vicinity of the point

which has the

following form

(&)

=

id2
2 d&r2

()]

(14.)

It follows from relation (2) that the pulse envelope mode at the die—

persion channel output in function of U parameter essentially

depends on the relation Z I 1, . The calculations showeçl that the

relation 17

1.6 is the threshold one, but when 2/i, <1.6 the

SPIE Vol. 1968 Atmospheric Propagation and Remote Sensing 11(1993) / 801 Downloaded From: http://proceedings.spiedigitallibrary.org/ on 07/03/2016 Terms of Use: http://spiedigitallibrary.org/ss/TermsOfUse.aspx

pul8e mode at the channel output le distorted so much that its duration
determined by the half value
of its amplitude is more than V Therefore the limitation set on the duration decrease of the input radio
pulse by the dispersing medium may be written in the following form

'-'mri >/ 1.6

(5)

\J•%%/

Let us express Z, by the values. It is shown in paper 9 that the
4'h) piiase invariant of the modulated

in Pig.5.Results of
extrernum values

observations of of phase invariant

oscillation propagating in diapers'-

ing mediUm dispersion

mpaayrabme eetxeprrDessdeedtebrymitnheed

the vicinity of the point 2Y&'

çP

(6)

where

is modulation frequency.

d22 dd22 = ____ —

rc?,-,.2, (2o-) L

I - 18 wave number.

Comparing expressions for

and D from relation (6) we may obtain

x(z)=O,5c427f .

(7)

Taking into account (7), the expression (3) may be written in the

following form

o(7 (11!I2)/

(8)

In this case condition (5) will have the form

z-,7ztz >, 6(jt(coIa2)"2.

(9)

We will treat the value 1/tprzjpz as the maximum value of the
radiowave propagation dispersion channel pass band4fmx. Using condition (9) we may write

Jmxfl (2,56WII)hl2.

(10)

This expression allows to estimate communication lines band properties with the help of the P values measurements results. Applying meaaurements results presented in Pigs.2...5 let us calculate variation of

802 / SPIE Vol. 1968 Atmospheric Propagation and Remote Sensing 11(1993) Downloaded From: http://proceedings.spiedigitallibrary.org/ on 07/03/2016 Terms of Use: http://spiedigitallibrary.org/ss/TermsOfUse.aspx

4J1T?ax values of the experimental communication line. The largest

values of

in communication line turned out to be observed in

clear weather in spring and autumn. Under thee e condit ions J4 .5GHz.

In clear weather in summer and winter z/max changed within 2.6...

14 GHz. After rain fall out in warm season (summer, beginnning of

autumn) 4fmctx decreased to l.6...2 GHz. The motion of cold atmospheric

fronts had the most important effect on the communication line band

properties. Under these conditions the values of

decreased down

to l.O...l.5 GHz.

Thus systematic experimental investigations carried out by "phase invariant method" in the open near..surface communication line of millimeter range showed that in atmosphere "frequency windows" its surface layer manifests the properties of the medium with regular dispersion. The value and the law of the dispersion change have diurnal and season variation and are not connected to the dispersion in atmos— pheric gases and the presence of hydrometers in radiowaves propagation channel.
A technique is suggested for estimation of band properties of coiu— nication lines in which the necessary and sufficient calculation data are the phase invariant values. Estimates of the surface atmosphere layer conducted with application of this technique and the results of experimental investigations showed that surface communication lines of millimeter range possess the widest pass band in clear weather in spring and autumn. The most considerable decrease of the pass band takes place in the motion of cold atmospheric fronts.

The dispersion found in atmosphere "frequency windows" may by the most important factor limiting pass bands of surface communication lines of millimeter range compared with other known phenomena causing such limitations.

It. REFERENCES
1. R.C.Dixon, Spread Spectrum Sstems , A.Wiley—interscience publi—
cation, John Wiley ant eons, New York. London. Sydney. Toronto. 2. L.I.Sharapov,"Influency of atmospheric gasas absorption lines on
the propagation of wide band signals of the short part of millimeter waves range", Preprint No.121 of Inst. of Radiophys.& Electron. Ac. of Sci. Ukr. SSR, 28p., 1979 (in Russian).
3. S.A.Zhevakin, A.P.Naumov, "On refraction factor of the low atmos— phere on millimeter and submillimeter radio waves, magnet permittivity of molecular oxygen", Radiotekhnika i elektronika, Vol.12, No.8,
pp.1339—1342, 1967 (in Russian).
k. T.Oguchi,"Electromagnetic wave propagation and scattering in rain
and other hydrometeors", Proc. of IEEE, Vol.71, No.9, pp.1029—lO'?8,l983. 5. D.E.Setzer, "Computed transmission through rain at microwave and
visible frequencies", Bell Syst. Techn. J., Vol.149, No.8, pp.1873—1892, 1970.
6. R.K.Crane, "Fundamental limitations caused by propagation", Proc. IEEE, Vol.69, pp.196—209, 1981.

SPIE Vol. 1968 Mmospheric Propagation and Remote Sensing 11(1993) / 803 Downloaded From: http://proceedings.spiedigitallibrary.org/ on 07/03/2016 Terms of Use: http://spiedigitallibrary.org/ss/TermsOfUse.aspx

7. Yu.M.Galaev, B.V.Zhukov, P.V.Kivva, "Method of phaae invariant and 1t8 application to the investigation of radio waves propagation
III All-Union Symp. on Propagation of mm and submm waves in atmosphere:

Abstracts of papers, Kharkov, pp.2OO-2Ol, 1989 (in Russian). 8. V.A.Zverev, "Modulation method of ultrasound dispersion measure-

ment", Doklady AN SSSR, Vol.91, No.4, pp.791-794, 1953 (in Russian). 9. V.A.Zverev, "On a new method of ultrasound dispersion investigaa

(in tion", Memorial of A.A.Andronov: proceedings, Moscow, AN SSSR, pp.657-'

—680, 1955

Russian).

10. M.A.Koloaov, N.A.Arinand, 0.I.Yakovlev, "Radiowaves propagation

in space communication", Moscow, Svyaz', 156p., 1969 (in Russian).

804 / SP1E Vol. 1968 Atmospheric Propagation and Remote Sensing 11(1993) Downloaded From: http://proceedings.spiedigitallibrary.org/ on 07/03/2016 Terms of Use: http://spiedigitallibrary.org/ss/TermsOfUse.aspx