both ends of the communication path. The computer program written originally came from Bob Atkins' article in Communications Quarterly (1991), but was translated to True BASIC, then revised several times to achieve the accuracy and detail needed to obtain the results desired for this study.

I've communicated using this mode on 6 meters and have some familiarity with the quirks experienced in its use. But I am sure there are readers who have much wider experience and could profitably add to the accuracy of these results. All known effects were included which would modify and limit the transmission (see Appendix).

Troposcatter signals in the UHF and microwave bands have a similar range if they have a similar transmitter power. While there is increasing loss at higher frequencies, there is increasing gain in the same capture area of the antenna. The two factors balance each other. For the constant 100-W transmitter, Table 1, note the increasing ERP, especially of the cylindrical parabola with its constant 64-square-foot area. Since it is assumed that the receiving antenna has the same area, the increased gain compensates for the increased "loss" due to frequency.

In troposcatter transmission the most power is found in direct line-of-sight from the original beam. However,

### Tables begin on page 25.

as a result of collisions of the beam with molecular substances in the air, some of the energy is scattered. The path with the least angle of scatter, or bend, between the transmitting station and the receiving station will give the strongest signal, and thus the greatest range. Transmission distance versus take-off angle for a given power is shown in Table 2. According to this table, a station on a high hill, with only lower hills on its horizon, will give very distant contacts. The higher the station has to elevate its beam to cross its horizon, the shorter its range will extend. The range drops off rapidly. Note in this table: that the ERP (output of the antenna) is held constant, and the mileages are, as a result, reduced with higher frequency. (I will not guarantee the accuracy of the -1° mileage; it worked out this way on the computer program!)

Table 3 shows the effect of air pressure on the range of the troposcatter signal. The lower the pressure, the greater the mileage. This is probably because the increased pressure results in more obstructions in the same space, hence reducing the power in any one path. As shown, the change in path length for a given pressure change is about constant.

In Table 4, several patterns can be observed that are present on all further charts. There is a peak range, normally close to freezing at low humidity (with

Fig 1—Troposcatter profile.

more range at lower humidity and less range at higher). Note, though, the slight dropoff in ranges below freezing.

Compare Table 9 with Tables 10 and 11 and note how the range changes as air pressure increases and decreases. Higher air pressure means less range; lower air pressure means increased range.

Comparing Table 12 to Table 9 shows the rapid decrease in range with a slight increase in the elevation of the beam. (The elevation value was chosen for author's site. )

In Tables 16-20, note the increasing antenna gain due to decreasing wavelength and constant reflector size (capture area).

These tables also show how high heat and humidity drastically reduce range at high frequencies. In Table 20, note the available distances during dry, cool weather, even at low power.

Comparison of Table 21 with Tables 22 and 23 shows very similar results to those at 432 MHz for varying air pressure. Because of the sharp vertical beam pattern, it is possible that no Earth noise will intrude on the received signal, and the 310-K signal loss might be relaxed. I could not find information to substantiate this, and we have not begun testing this data at our club, so 310 K was retained.

In Table 24, increased elevation of the antenna beam causes considerably reduced ranges from those of Table 21. This is consistent with our findings at 432 MHz.

Note the 10- to 18-mile increase in the ranges of Table 26 compared to Table 25 with the cylindrical parabolic antenna. This is due to the 3-dB higher gain of the dish in Table 26.

The large antenna of Table 27 gives a respectable range even with a low-power Gunn diode transmitter.

Table 29 shows a cold weather sport. Find something better for those summer days! I have been looking for some high-power diodes that would work as

push-push frequency/power doublers for this band—anyone know of a source? If I find one, I'll stick one of our 10-W, 10-GHz power transistors on it!

Receiver noise figure values have a significant relationship to troposcatter distances, as shown in Table 30. Note that for these arbitrary values for the 432-MHz band, the difference of 1 dB in the noise figure is equal to 10 miles. Similarly, 1 dB is about 6 or 7 miles for the 10-GHz band. This is a good reason for trying to obtain the lowest noise figure front ends possible on the preamp.

Conclusions

There is real possibility of a communication range of about 300 miles through all the microwave bands using troposcatter communications. This is

Fig 4—Cylindrical parabolic antenna.

especially true if high-quality communication equipment is used (moon-bounce equivalent). Most of these stations will need to be home brew, with the best of construction techniques and special attention paid to low-noise reception. The worst problems are beyond the control of the amateur, but are addressable: the path needs to be as near to 0° above the horizon as possible (to get the nearest direct angle and, hence, most scatter power); the atmosphere should have the least amount of water moisture in it (low temperature—around freezing, and low percent humidity). Consistent microwave DX communication is there during all but the hot, humid summer days.

The author is a college professor of physics and electronics. He has been chief engineer for WKOl, Ch 43 TV, Richmond/Cincinnati. This was a presentation made to the Midwest VHF-UHF Society meeting, February 26, 1993.

### Bibliography

Reference Data for Radio Engineers, 4th Ed, 11th printing, ITT, 1964 (in my case). The 1987 ARRL Handbook, ARRL, Newington CT. The ARRL UHF/Microwave Experimenters Manual, 1990 (very thorough—almost all of my formulas were in this), ARRL, Newington CT.

The Microwave Newsletter Technical Collection, (G3YGF/ G4KNZ), 1983, RSGB.

Bob Atkins, KA1GT, "Radio Propagation by Tropospheric Scattering," Communications Quarterly, Winter 1991, p 119.

DJ9HO, et al, The UHF Compendium Parts 1,2,3 & 4 (now if only they would bring out a microwave Part 5).

Appendix

To meet the stated limits, several choices were made:

• The transmitters were determined to have the maximum possible power on each band. On the VHF bands, this power was to be a full kilowatt output. [This includes the novel SSB method, used in Europe, SBFM, or the clipping of the amplitude component of the SSB signal so that only the frequency change is transmitted; there is no difference required in the reception of this signal and it is not obvious to the receiving station that this transmission method is used (so they say—the author is adapting a transmitter to this method to try it). The signal thus would be able to use Class C, or better, amplifiers, not be concerned with linearity, and so have high efficiency—75% or better, close to a 3-db gain in power.] The use of a 2C39 amplifier (water cooled) gives good power on 1.3 and 2.4 GHz (see the ARRL Handbook and QST), a 1977 article by W9ZIH ("Video modulated Four-Tube Amplifier for 1270-MHz Television," June 1977, Ham Radio) describes a 4-tube 2C39 cavity amp at 1.3 GHz that would result in a full kW). Some 10-watt transistors are available for 10 GHz, but it is recognized that most microwave depends on the Gunn diode, and 0.1 watt is pretty good output with these. The transmitters should be mounted at the antenna, or have low-loss leads (open-wire twin lead at VHF frequencies).

• The antennas were limited to two types: for the low VHF/ UHF frequencies, the dual-quad Yagi (from Weiner, UHF Compendium) was chosen, using 20-foot pipe for the booms for 50 MHz and 10-foot pipe for 144, 222, and 432 MHz. For the microwave frequencies, a cylindrical parabolic reflector (re: 1961-66 73 Magazine, "Big Sail Antenna" and related articles) was chosen, size 8'x 8'. This allowed high gain (3-db less than the full parabolic) from a sharp vertical pattern, with reasonable beam width (fan pattern) in the horizontal, to permit CQ-type contacts. (Normal amateur communications in the microwave bands—without all the problems of prearranged time and frequency and dish alignment.) Since reflectors are sensitive to X/10 irregularities, the use of the reflector at 10 GHz means that the surface must be accurate on the parabolic curve to 3 mm (1/8 inch) (1 mm at 24 GHz). This means a solid sheet of metal (aluminum) fixed to a rigid set of ribs. This is nearly impossible for a full dish; even a 2-meter satellite dish may not meet this requirement. (These are built for 3.5-GHz. Don't use a perforated dish above 3.5 GHz!) So the simple curve of the cylinder antenna is almost demanded—and gives the benefit of a fan type horizontal beam. (I wonder how aluminum covered insulation board would work? 2 sheets at 4x8?) The feed (horns) for transmit and receive are mounted on the same base, but are separate units. To avoid transmitted power reflecting into the receiving system, a wedge isolator was assumed to be mounted on the reflector directly below the transmitter horn/receiver horn junction (or, with more loss, the transmitter feed can be a pipe com ing up from the reflector between the two systems). This causes a slight loss of directed power to the distance but permits fall-duplex operation. The antenna was considered to be mounted on a stable base, at a height of 30 feet (a tripod made of 3 old telephone poles! for the cylindrical parabolic). Wish I didn't have to move!

• The pre-amps: GaAsFETs can give close to 30X (0.5 dB) front ends at the UHF frequencies, and the new HEMTs may do 1.0 dB (xF at frequencies beyond those in this study. 0.5 dB was chosen as the input noise figure below 1 GHz and 1.0 dB for those bands above. It must be acknowledged that 1.0 dB is a reach, with typical values now nearer 5.0 dB. Preamps are mounted at the antenna and a separate coax feed line run for receive.

•Formulas index of refraction:

t r based on water vapor pressure e := h (9.4051-(2353/t))

with: m = pressure (millibars) h = percent humidity t = temperature (K)

refractive index loss In = 0.2(ns - 310)

(310 is the temperature of the Earth in Kelvins)

scattering angle (degrees) sa = 0.005dmi]es + 2t (t in K)

3-db beamwidth 27000

coupling loss

(slight distortion of results on cylindrical reflector)

(all measurements are in meters) h2= height of hills hi = height of ant d = distance gain of long Yagi gi == 10.1 + 6 logbl/A.

bl = boom length (feet) ibook says 10.1 should be 12.1,1 couldn't do it!)