United States Air ForceibSy - PDF Document

Presentation Transcript

  1. AFCRL.4&.0138 MARCH 1969 ENVIRONMENTAL RESEARCH PAPERS, HO.28 AIR'FORCE CAMOkeDGE RESEARCH LABORAT~ORIES L G. HAMM~ FIELD, BEOPOPO, MMSACMIETTS I1O A o Proosed Inde forMeasurin Ionopheic, Scintillation HERBERT E. WAH'ITNEY CH E ST ER MA LI K JUN4 )1968 OFFICE-OF AEROSPACE RESEARCH United States Air ForceibSy C L A RI N H 0U S-

  2. AFCRL-68-0138 MARCH 1968 ENVIRONMENTAL RESEARCH PAPERS, NO. 284 JU IONOSPHERIC PHYSICS LABORATORY AIR FORCE CAMBRIDGE RESEARCH LABORATORIES PROJECT 4643 L G. HANSCOM FIELD, BEDFORD, MASSACHUSETTS A Proposed Index for Measuring Ionospheric Scintillation HERBERT E. WHITNEY CHESTER MALIK be released to the Clearinghouse, Department of Commerce, for sale to the general public. OFFICE OF AEROSPACE RESEARCH United States Air Force I

  3. I IIII Abstract Signals from radio star sources and satellites are regularly used for study- ing ionospheric irregularities. Amplitude and phase deviations can be imposed on the signals from these sources as they traverse the ionosphere. A parameter frequently used to describe the magnitude of this scintillation effect is scintilla- tion index. An experiment was designed to correlate various methods of making scintil- lation measurements; observations were made at 40 MHz using the ionospheric beacon, S-66. It is shown that when the law of the receiver detector is known, a conversion method allows comparison of data and statistics on scintillation index. A simplified method of scaling scintillation index is described. The accu- racy of the simplified method is determined by a comparison with measurements of scintillation index by machine computation. Ii iii PRECEDING PAGE BLANK

  4. t7 Contenti 1. 1 INTRODUCTION 3 EXPERIMENTAL RESULTS 2. 9 SCALING OF SCINTILLATION INDEX 3. CONCLUSION 13 4. 15 REFERENCE Illustrations 1. 2. 136 MHz Scintillations from the Early Bird Synchronous Satellite Detector Characteristic of 3 R-390 Receivers Illustration of Scaling of Scintillation Index for Three Detector Characteristics Measured Values of Scintillation Index from Linear Voltage Receiver vs. Measured Values of Scintillation Index from Linear Power Receiver Measured Values of Scintillatioi index from Receiver with AGC Characteristic vs. Measured Values of Scintillation Index from Linear Power Receiver Measured Values of Scintillation Index from Linear Voltage Receiver vs. Measured Values of Scintillation Index from Linear Power Receiver 2 4 5 3. 6 4. 7 5. 7 6. v

  5. flluwo.ons 7. 8 Measured Values of Scintillation.Index from Linear Voltage.Receiver vs. Measured Values of Scintillation Index from Linear Power Receiver Measured Values of Scintillation Index from Receiver-With AGC Characteristic vs. Measured Values of Scintillation Index from Linear Power Receiver' 8. 8 10 11 12 12 Graph for the Conversion of Pmax - Pmin to ScintillationIndex Comparison of Scintillation Index Measurements Comparison of Scintillation Index Measurements Comparison:of Scintillation Index Measurements 9. 10. 11. 12. Table 1. Conversi)n Table: Pmax- Pin (dB) vs. S. I. (%) 10 II I1 i i Vi vi It

  6. I. A-Proposed Index for Measuring Ionospheric Scintillations I. INTROIDUCTION Studying the'effects of ionospheric irregularities often involves measuring the aiplitude and phase variations imlosed'on a signal as it traverses the iono- sphere. Many workers have used signals from radio stars aiid satellites to record the changes in amplitude known-as scintillatioiis. Several measures are used to characterize the depth of-scintillations; each of these measures is called a "scintillation index" (S. I.). One measure is to use a scale from 0 to 5 and by visual inspection, without actual m3asurement, assian a value to the record. Other ways involve scaling lthe'deviatiop,pf the signal amplitude from the mean amplitude. are thus possible, depending on whether mean deviation or root-mean-square deviation is used and whether the record is proportional to voltage or power. the probability distribution of the amplitude deviation is known, it is possible to ,relatethe four measures of scintillation index (Briggs ane Parkin, 1963). it is difficult to relate these measures theoretically because/the probability dis- tribution of the amplitude deviation is not in general known, experimental compar- isons'show that they are proportional to each other. These relationships and additional considerations are given in the cited reference. Four measures If While (Received for publication 18 March 1968)

  7. 2 'However, 'in most cases, only a relative measure of scintillation index is required. For a statisticalanalysis that involves a long period of time, such as a study of the seasonal dependence of scintillations, itis important.that-a standard method of data scalingbe used. The notation that has been adopted by-AFCRL and the JSSG (Joint Satellite Studies Group) is: P P -P +P'. . max Scintillation Index = nnn nin max where P ais excursion of the scintilletions, -and Pmin is the power amplitude of the third level up from the minimum excursion. The use of this expression is demonstrated in Figure 1. The chart, record is calibrated in dBm which is converted to relative power levels for calculation of S.1. Over the period 1730 to 1,745 E.S. T. an index of 40 percentis obtained. the power amplitude of the thirdLpeak down from-the maximum max .63-112dbm 6I 1120 1730 1745 E.ST. Figure 1. 136 MHz Scintillations from, the Early Bird Synchronous Satellite S. 1. ZPma r Pm P 4 0 Pmax + Pmin However, even though only relative values are Usually required in a study of ionospheric scintillations, problems do arise in the comparison of data taken by various workers. Some measures of S.1. may have been scaled from chart divi- sions rather Lban by using an-accurate calibration. The S.1. scaled from the chart divisions under'anassumption thatthe chart display is-either linear in power or voltage can lead to errors; in reality the chart amplitude could be somewhere in

  8. 3 I between, because it depends on the law of the receiver detector. Also, statis- tics on satellite scintillations may be in the formof a voltage-index distribution based on a voltage calibration. For comparison of data, it is-useful to have a, conversion-relationship between power index, S. I. (p), and voltage index, S. I. (v)" If dat- was obtained from a receiver with AGC and the slope of the detector durVe is known, scintillation, index values based on chart divisions can be converted for comparison on a S.(p) basis. Because of the emphasis on relating statistics on ionospheric scintillations to the problems of communication engineers, the following method was developed for cbnverting data to a common base, S. I.(p)- The amount that a signal fades is important to communication engineers and is:related to scintillation index by the following: Signal, fades (dB) = 10 log Il-S. 1-S.,I. (v)]" (. = 20 log For example, if S. I. (p) = ,0. 5 = 50 percent, then the'signal fade is 3 dB. Since the signal fade in dB would be the same on either a linear power or linearvoltage display, the relationship is formed: S. I. (p) = 2 S. I. This conversion method can be extended to cbver receivers that are not linear with input power or voltage, such as a receiver with AGC, but do have-a constant slope to~the detector curve over the fading range, by the following: S. I. 2 (v) (v) x 1-s. 1. S. 1'P) = 1 wher S.I is the scintillation index measured in chart divisions with a receiver that has a detector law exponent of x. 2. EXPERINENIAI, RESUI.TS To determine the accuracy of the described method for converting S. I. values to a con receivers were gain adjusted so that a 46 MHz satellite signal received on a com- mon antenna and a comijion frequency converter, whose output had been equally divided into three parts, would detect the identical signal in three different modes linearpower mode, linear voltage mode, and the AGC mode. Asshown in pnon base, the following experiment was made. Three R-390 Collins Kp I

  9. 4 Figure 2, the detector characteristics 6f each receiver had been carefully mea- sured to determine over what signal range the, receivers would be essentially linear with power or voltage or maintain a constant detector law. IO REC. *1 APPROX. LINEAR WITH INPUT POWER. REC'*2 APPROX. LINEAR WITH INPUT VOLTAGE (DET. OUT 1- 0 ) .. RE. e3 A.G.C. CHARACTERISTIC - >/ ro o RC. 01 I II I io- ° °0 0 10 POWER INPUT -MW Figure 2. Detector Char;cteristic of 3 R-390 Receivers By careful adjustment of the receiver gains and the chart recorder gains to maintain a full-scale deflection for the same input level, it was,ppssible to simultaneously record-identical signal excursions in the three mfodes. Each channel was accurately calibrated so that scintillation index could~be measured for the three modes for identical signal excursions either using chart-division scaling or the voltage and power calibration. in Figure 3. For this analysis identical signal peaks were used for~the calcula- tions of S. I. A sample chart record is shown I I

  10. 5I 1 I 7 AMI - • Figure 3. acteristics Illustration of'Scaling of Scintillation Index for Three Detector Char- Scintillation index was calculated for-various degrees of scintillation foi, all three modes. These index numbei-s were then plotted against each other in vari- ous ways to determine the practicability and relative statistical accuracy of corn- paring (I) resultsof equipment having dissimilar characteristics, (2) scintillation measurements derived when calibrations were not taken and a detector charac- teristic law was, assumed, and (3) conversion of scintillation measurementsito, a common standard. The results of the experiment to compare scintillation index measurement are shown ia Figures 4 through 8. Figure 4 shows the plot of~the index obtained by using the signal power calibration for the linear voltage receiver (middle chan- nel of Figure 3) vs. the index for the linear power receiver (lower channel). If perfect accuracy was obtained in the calibratioa and scaling, each channel would indicate the-same S. I.; however, there is a mean error f6r the plotted points of about + 2. 2 percent.

  11. 64 37.I . . -14) -Ion 401P 24 - Figure 4. Measured Values of Scintillation Index from linear Voltage Receiver vs. ;Measured Values of Scintil- lation Index from Linear Power Receiver. bration for both receivers 4.. Power cali- Figure 5 shows the plot of the index obtained by using the signal power cali- bration for the AGC -receiver (upper channel of Figure 3) vs. Athe indeA for the linear power receiver. There is a mean error for-the plotted points of + 4 per-t ~~~~v ECWR (LvY.- ! ' ~cent. Figure 6 shows the plot of the index obtained by using the signal voltage calibration for the linear voltage receiver vs. the index for the linear power = 2 S. 1. (v receiver. The solid curve represents the relationship-S, I.,(p) The scatter in the points shows a mean dispersion of about 4 percent. cent. S.. 1. Figure 7 represents the same data as, Figure 6, but in this case the Voltage index was scaled using chart divisions. Figure 8 shows the plot of index values taken from the AGC channel using chart divisions. The solid curve represents the relationship S. 4.5 ((GC)) (I-S.gr 7rp t The exponent of 4. 5 represents themean slope of the AGC I. - * 4 detector characteristic as shown in Figure 2. 1i

  12. 7 mo *56 72 so * Ir. 64 AG - %4~~ 40!- 011 // ' . | • , a, * Oi 4/ h 0 M SO 72 64 56 40 40 3n, 24 16 0 *RECEIVER) PWRSCIMMILAIM0 IfAXX S (UNEFR POWE Figure 5. Measured Values of Scintillatin Index from Receiver with AGC Char- acteristic vs. M easured Value~s of Scintillation Index from Lnear Power Receiver. Power calibration for both receivers 724 as 72 56 2 600 0 0 9. -a S64- 16 81 96 00 $a 24 i2 P 4 0 IEX N 24 6 A POWER SCINTILLATIN (LNEAR POWER RECEIVER) Measured Values of Scintillation Index from Linear Voltage Receiver Power calibration for linear power Figure 6. vs. Measured Values of Scintillation Index from Linear Power Receiver. Volt- age calibration for linear voltage receiver. receiver

  13. 82 64- 56. > 48- 4z40 32 -SP) 2 SI(,,), SIV, 24- A ~ C b . POWER 0& 6,64 A2 6,0 O- 96, j0, 232 SCITILLATION WDE6K ~LINEAR POWER RECEIVER) Figure 7. Measured Values.of Scintillation Index from Linear Voltage Receiver vs. Measured Values of Scintillation index from Linear Power Receiver. Chart division calibration for linear voltage receiver. Power calibration for linear power receiver 4 S56- * S1 X40- 24 . - 0a1'24 32 46 4856 64 7 0B r 0 POWER SCINTILLATION INDEX%. (LINEAR POWER RECEIVER) Figure 8. Measured Values of Scintillation Index from Receiver with AGC Characteristic vs. Measured Values of Scintillation Index fr~om Linear Power Receiver. Chart division calibration for AGC receiver. Power calibration for linear power receiver

  14. 9 XI SK-.WlNl ( F-SI.NTII.I.UIO. INIW.X It was pointed out in a preceding section tlat:scintillaticn index was defined as mevel an . At first the measure of S. I. was made by measuring the two levels and -then computing, either by machine or by hand, the resulting S. I. The first step was to mark the Pmax and Pmin levels on the chart for each time period that S. I. was to be read. Pmax was arbitrarily chosen as the third peak down from the maximum and Pmin as the third minimum up from the deepest fade. In the case of the loi altitude satellite, S.J. was read eithMr at the peak of the Faraday cycle or once per minute if Faraday periods were ob- scured by heavy scintillation. For the synchronous satellite signals, scintilla- tionindex was measured in each 5 minute interval. The time constant of the recording system was chosen to be short compared with-the fastest scintillation. The second step was to read Pmax and P Mi as a relative power level based on an accurate amplitude calibration of. the chart deflection. If the calibration had been made in decibel steps, it was converted to an appropriate numerical ratio. Scintillation index was either computed manually, using the formula, or Pmax and Pmin were punched on cards for machine calculation of S. I. and further analysis. A simplified method of measuring S. I. has evolved which is now implemented for all S. I. measurements. It again depends on an accurate amplitude calibra- tion in decibel steps of the chart deflection. the same manner as described in step 1 above. The dccibel change of P P maxand Pmin are determined in - max Pmin is then read, using the receiver amplitude calibration curve. The mea- sured values in decibels of Pmax - Prm are then converted to S. I., using the curve shown in Figure 9. If further analysis is required, then either the decibel values of P - P or the values converted to scintillation index can be max min punched on cards 'for machine processing. The conversion graph shown in Figure 9 was determined by assuming equal percentage, changes from the average level for Pmax and Pmin' changing those values to a decibel change-and then -summing them for the total decibel change as, shown in Table 1. For example, based on an arbitrary average level of 1. 0, a S.1. = 50 percent corresponds- to a P min= 0.5 = 3 dB, a P max= 1.5 1. 77 dB, and, therefore, a total change for P As a check on the accuracy of the simplified method compared with the previously used method of machine computation, several values of scintillation index were measured using both procedures. The calibration of an S-66 record at 41 MHz was chosen because the average level for low altitude satellites changes over a far greater range than that from synchronous satellites. - P of 4. 77 dB. max min

  15. 10 00 z 4 z 2 3 4 -51 7 .o 2 ,o 3D 40 PMAX- PNIreD Figure 9. Graph for the Conversion of Pmax - Pmin to Scintillation Index. Pmax - Pminn is the peak-to-peak excursion of a scintillating signal and is mea- sured in decibels based on an amplitude calibration of the chart record Table l. Conversion Table: Prmax - P min (dB). vs. S. I. (% -dB +dB Pmax - P (dB) S.I.(%) .09 .22 .46 .71 .97 1.25 1.55 1.87 2.22 2-60" 3.01 3.47 3.98 4.56 5.23 6.02 6.99 8.24 10.00 13.00 20.00 30.00 .09 .21 .41 .61 .79 .97 1.14 1.30 1.46 1.61 1.76 1.90 2.01 2.18 2.31 2.43 2.55 2. G7 2.79 2.90 2.99 3.00 .18 .43 .87 1.32 1.76 2.22 2.69 3.47 3.68 4.21 4.77 5.37 5.99 6.74 7.54 8.45 9.54 10.91 12.79 15.90 22.99 33.00 2 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 99 99.9

  16. The results of this comparison are shown in Figures 10, 11, and i2. The measurements of Pma iwndB and their conversions to S. 1. are plotted against the normal method of machline computation of S. 1. Figure 10 corresponds to the case of measuring S. 1. for weak~ signals, that is, where the average level is about 114-of the chart width, Figure 11 for average levels of about 1/ 2 of the chart width, and Figure 12 for strong signals with an aveiage level about 3/4 of the chart width. The deviations of the points from the dotted line show the lack of perfect agreement between the two m~ethods of measurement. Tile greatest differences are for the small signal case and are due to the-compression of tile calibration curve at small signal levrels which results in greater, iraccuracies of reading the value of the signal. - P : OO . Z90- 40 -' 70 06 0 - a40- 220- / di 0 10-20 S.I. (%) NORMAL METHOD- MACHINE C0MPI-!TATIONt 30 50 60 70 40 90 100 80 Figure 10. Comparison of Scintillation Index Measurements. directly from calibration and conversion curve vs. S..I. measured] with machine computation of Pmax - Pmin/Pmax + Pimin; small signal case-average level approximately 1/4 chart-width (10-15 mm) S. 1. measured

  17. 12 .>,oo ,Ioo z9g_ h/, s- / _7O 0 o * 060- V / '40 / - L 30 - 020 - I 0-I I * 100 90 80 70 60 50 40 20 10 S.I. (%) NORMAL METHOD - MACHINE COMPUTATION 0 S. I. measul ad Figure 11. Comparison of Scintillation Index Measurements. directly from calibration and conversion curve vs. S. I. measured with Imachine computation of Pmax - Pmin/Pmax + Pmin; medium signal case-average level approximately 1/2 chart width, (20-30 mm) >1001- / o " 70- t 60 / / 0o 50 - W / / o 40 w 0 -/ "20 1o /2 90 100 70 80 50 60 -20 30 40 10 S.I, (%) NORMAL METHOD - MACHINE cOMrUTATION 0 S. I. measured Figure 12. Comparison of Scintillation Index Measurements. directly from calibration and conversion curve vs. S. I. measured with machine computation of /max - Pmin/Pmax + Pmin;,large signal case-average level approximately 3/4 chart width (30-35 mm)

  18. If the simplified method is to be applied for analysis where there is a small signal-to-noise ratio and the deflection on the chart due to the sky background temperature is appreciable compared to the signal, the receiver calibration must be correctly applied so that On lyi ie.deflections from the signal are being mea- sured. I. C LUS(:IsION in summary, it is felt that only a relative measure-of scintillation index is necessary to describe amplitude fluctuations caused by ionospheric irregularities, P - and that the-simple measure S. I. = PMa should be used. To obtain the best accuracy when comparing scintillation data, the records should be accurately calibrated on-a power base. If scintillation index is scaled from chart divisions and the-law of,the detector is known, the in- dex values can be converted to a power base with a small decrease in accuracy, as compared to scaling from a sigiial generator calibration. Comparison of data without the benefit of-a calibration or knowledge of the detector characteristic can lead to errors of a factor of 2 or more. This is evident from Figure 3 where the apparent S. I., as indicated by chart divisions, is very small for the AGC mode and la. gest for the linear power mode. The simplified method of scaling S. I. gives adequate accuracy-,or statistical studies. - mi , mi' which was defined earlier,

  19. 15 Reference 'Briggs, B. H. and Parkin, I. A. (1963) On the variation of radio star and satellite scintillation with zenith angle, J. Atmos. Terr. Phys. 25:339. PRECEDING PAGE BLANK

  20. Unclassified -Security Clasadicim ODocmiff CONTUOLDATA. - R sarecs:and indeximg aneuution Rest &r entred a4en ale .. ereU nte is cleas.ifi (Secarily classificaione of lsle. bfof .1 1. OIWGIATHIG AwCTIVITY tCaorusr WI- Air Force Cambridge Research Laboratories (CRP) I- G. Hansconm Field Bedford. Massachusetts 01730 206 REPSRT SECURITY CLASSIFICATION Unclassified 2& GouP 7 3- REPORT TITLE A PRPSDINDEX FOR MEASURING IONOSPHERIC SCINTILLATIONS A. OESCR*TOVE NOTES fTWp of tepart wadimeiu.,iee desi Scientific S. AUTHOISI (Fiamme. mde imi d LImarnm) Herbert E. Whitney Chester Malik Interim- 1& NO. OF RIEPS 1 TOTAL NO. OF PAGES Dco~ ATE REOR March 1968 -7&. 20 AFCRL-68-0 138 OS. 4643-02-01 bPROJECT. TASK. WOUikusilT 6240539F si. c. 000 ELE1MENT Ay01* ubr Neyb ;4b ERP No. 284 681000 4L DOD SUmELEMENTi STRIDUTION STATEMENT 1Distribution of this document is unlimited. -It may be released to the Clearing- house, Department of Commerce, for sale to the general public. 1. -1 - IL SPONSORING MILITAR1Y ACTIVIT V Air Force Cambridge Research Laboratories (CRP) L. G. Hans com FieldZ Bedford, Massachusetts 01730 11. SUPPLEMENTARY NOTES TECH, OTHER 13. ABSTRACT, Signals from radio star sources and satellites are regularly-used for studying ionospheric irregularities.. Amplitude and phase' deviations can be, imposed on the signals-from these sources as they traverse the ionosphere. A parameter fre- queuitly used to describe the magnitude of this scintillation~effect ias scintillation index. An experiment was designed to correlate various methods of making scintil- lation measurements; observations were made at 40 MHz using the ionospheric beacon, S-66. It is shown that when the law of the receiver dete~ctor is known, a conversion method allows comparison of data and statistics on scintillation index. ofA simplified method of scaling scintillation index is described. The accuracy ofthe simplified method is determined by a comparison with measurements of -scintillation index-by miaching optain 17 SFORM 1D NOV as 17 I Unclassified Security Classification