Fritz offers the argument in his QEX 600m article [1] that 600m ground-wave regional (0-400km) communications links have potential advantages over NVIS (Near Vertical Incidence Sky-wave) and VHF troposcatter propagation modes.  He mentions that NVIS requires use of multiple bands and VHF troposcatter requires use of high gain directional antennas which precludes building a fully meshed peer-to-peer network.

This discussion uses propagation prediction software to understand NVIS and VHF troposcatter coverage and reliability for regional communications links followed by a similar analysis for 600m ground-wave links.  The discussion provides propagation analysis for both high reliability ARES public safety support links (90% reliability and 10dB fade margin) and casual Amateur Radio grade links (50% reliability and no fade margin).

NVIS Regional Networking during Solar minimum (2008) and Solar maximum (2012).

We first consider HF NVIS, (Near Vertical Incidence Sky-wave) which uses horizontally polarized antennas only a small fraction of a wavelength above ground to develop very high angle radiation patterns.  High angle radiation from these antennas, one required per band, is returned to the earth by ionospheric refraction to a regional footprint out to >400km.  Antennas are installed at as close to optimum 0.175λ heights as possible for each required band.

Resulting differences in optimum heights for each required band dictate use of separate antennas, usually inverted-V center-fed dipoles.  L.B. Cebik, W4RNL (SK), provides a lengthy discussion of the importance of NVIS antenna height in QEX Jan/Feb 2007 [2]  Three or four inverted-V dipoles each raised to optimum height can be fed with a single coaxial feed line, a 4:1 balun and balanced ladder-line between dipole feed points.  See discussion of similar antenna and feed arrangement via http://showcase.netins.net/web/wallio/LPDA.html.

Cebik’s discussion includes high angle radiation patterns and isotropic gain predicted by NEC-based antenna modeling systems.  This optimum gain, 6.5dBi, will be used for all NVIS propagation modeling that follows.

Combinations of multiple variables require hundreds of propagation modeling predictions:

Distances (4)             100km  200km  300km  400km

Bands (4)                   160m  80m  40m  30m

Solar cycle (2)           2008 (solar minimum)  2012 (probable solar maximum)

Reliability (2)              90% & 10dB fade margin vs. 50% & 0dB fade margin

Time-of-day (24)       Predictions for each hour of 24-hour days

We can significantly reduce the number of required predictions for combinations of these variables by creating graphs that cover entire 24-hour days.  But even then we would have 4*4*2*2 = 64 graphs for a complete analysis.   To keep readers awake I will discuss one multi-band solution and then use only 400km Best Usable Band solutions from two sources.  

We will use the WinCAP Wizard implementation of IonCAP to create prediction graphs showing Best Usable Band reliability and SNR across 24-hour days.  These prediction graphs are then complemented by online IPS LAMP predictions.       

IonCAP predictions by:

WinCAP Wizard V2.0 (V5.0.10 currently available)                     http://www.taborsoft.com/wwizard/

IPS LAMP predictions from:

IPS Radio and Space Services of the Australian Government   http://www2.ips.gov.au/,

HF Systems                                                                                       http://www2.ips.gov.au/HF_Systems,

Prediction Tools                                                                                http://www2.ips.gov.au/HF_Systems/7/1,

Local Area Mobile Prediction (LAMP)                                           http://www2.ips.gov.au/HF_Systems/7/1/9

These sources are used for four-season NVIS propagation predictions for  400km links across 2008 (solar minimum) and 2012 (probable solar maximum).  NVIS antennas with 6.5dBi optimum gain are used at transmit and receive stations with 100W of transmitter output power to predict SNR and RELIABILITY link performance.

Our conservative goals for vital ARRL Amateur Radio Emergency Service (ARES) applications require 90+% RELIABILITY and 10dB fade margin over 10dB SNR required for BPSK-31 10E-5 error rate (20dB SNR total).

Multi-band Prediction Example

400km link during January 2008 (solar minimum)

(Click on graphs for full size versions)

160m	80m	40m	30m	Best Usable Band
>38dB SNR >90% Reliability	>54dB SNR >96% Reliability	>17dB SNR >37% Reliability	>5dB SNR >2% Reliability	>59dB SNR >97% Reliability

IonCAP (WinCAP Wizard) predicts that both 160m and 80m exceed our >20dB SNR and >90% RELIABILITY goals during January 2008.  We would not need to use either 40m or 30m to meet our link margin and reliability goals for our entire regional distance range of 0-400km.  Regardless, the Best Usable Band (BUB) graph for Jan08 combines performance of all four bands, 160-30m, to predict excellent SNR and RELIABILITY.  Note that 40m is the best band for a few mid-day hours.

We will now gather IonCAP Best Usable Band predictions for four seasons of 2008 (solar minimum) and 2012 (solar maximum).  Our first entry will be the January 2008 BUB graph above.    

(Click on graphs for full size versions)

	January	April	July	October
IPS LAMP Predictions RED 160m Yellow 80m Olive 40m Dark Green 30m 2008 (Solar min) IonCAP Predictions for 400km
IPS LAMP Predictions RED 160m Yellow 80m Olive 40m Dark Green 30m Light Green 20m WHITE None 2012 (Solar max) IonCAP Predictions for 400km

Further analysis finds that 80m performance would make use of 160m redundant during all of 2008 and solar minimum.  This would allow use of a 13.5m (44ft) mast which could be safely and dependably deployed by a small team.  It is important to also note that during all four seasons of 2008, ARES NVIS reliability 0-400km would require only single band operation on 80m.

Single band NVIS operation would not be effective during the four seasons of 2012 and solar maximum but use of 160m would not be crucial.  This again would allow use of a 13.5m (44ft) mast for the feed-point of 80m, 40m and 30m inverted-V dipoles.

ALE, Automatic Link Establishment which continuously selects the best bands to maintain a link, works very well for military, government and other dedicated frequency services but it is problematic for our Amateur Radio operations.  However, what could be called MLE, Manual Link Establishment, which includes operator awareness of frequencies in use by other stations, is known to be operationally viable given similar multi-band antennas.

Given dipoles for 80m, 40m and 30m, each at optimum NVIS height above ground, all fed via a single coaxial feed line, both IonCAP and IPS LAMP software predict an NVIS station would provide >90% reliability at >20dB SNR to other similarly equipped stations 0-400km 24-hours a day.  Manual Link Establishment would be used to find the Best Usable Band with just a few minutes of link testing.

Oh, and what about casual operating(?)  Reliability is >50% and SNR is >10dB when using just 1W of transmitter output power.

Regional VHF troposcatter communications links using directional high-gain antennas

As proved in every VHF contest, troposcatter links provide a dependable mode for Amateur Radio communications.  Grid squares worked by a modest contest station on 2m during the ARRL 1989 September VHF Contest are highlighted on this map:

This was all troposcatter communications out to >400km using 100W and 10dBi antenna gain at 15m height.  This performance has been typical during numerous VHF contests when there were no extraordinary propagation openings.

VHF/UHF over-the-horizon troposcatter propagation software can be used to understand coverage of communications links which could be at  regional distances.  This software uses terrain map data sets to factor in height advantages and shadowing from intervening high ground.  Setup includes antenna heights and gains, transmit power and transmission line losses, geographical location coordinates of interest and desired link reliability for a required receive signal level and/or SNR.

Troposcatter propagation predictions by:

Radio Mobile Deluxe V6.7  (V8.9.9 currently available)               http://www.cplus.org/rmw/english1.html

Two prediction examples are of interest, >50% reliability with >10dB SNR and >90% reliability with >20dB SNR.  Setup for these predictions includes 15m antenna height (above terrain height); 10dBi antenna gain; 100W transmit power; 3dB transmission line loss; and station location coordinates out 100km, 200km, 300km and 400km from a master station.

(Click on maps for full size versions)

50% Reliability	90% Reliability

Predictions for >50% reliability show close to solid coverage out to 300km and then broken coverage out to 400km and slightly beyond.  This prediction matches anecdotal impressions from decades of 2m SSB/CW operation.  The 90% prediction is a very different animal with solid coverage out to only100km and then broken coverage to ~150km.

These predictions require directional antennas which preclude a fully meshed peer-to-peer network.

600m (~505KHz) regional ground-wave communications

Regional 600m ground-wave links do not use sky-wave propagation via the ionosphere (as does NVIS) or troposcatter propagation (as do VHF links).  Ground-wave propagation (aka, surface-wave for LF-MF) starts with vertically polarized waves at a very low radiation angle which are coupled to surrounding terrain for frequency and ground-conductivity dependent distances.  This propagation mode is used by the AM broadcast industry to reliably cover large service areas with dependable signal strength.  Lower frequency ground-waves suffer less path loss than higher frequency waves which significantly favors our experimental 600m band.

Software is available to predict SNR performance over variable frequencies and distances but we first need to understand that very short 600m antennas relative to wavelength (591.3m at 507kHz) yield very poor efficiency and therefore very low <<0dBi gain.  Radial field quality, both in radial length and number of radials, also has a significant impact on antenna system efficiency as does capacitive and inductive loading.

Before dealing with radial and loading issues we can get a feel for antenna efficiency and gain by following John Belrose’s discussion in QST [3].  We can calculate and graph electrical heights in degrees (Gv) and radiation resistances in ohms (Rr) for a range of monopole heights; then complex impedances; then efficiencies and gains.

(Click on graphs for full size versions)

Gv & Rr vs Height	-jX & Rr vs Height	Eff% & Gain vs Height

(Gain[dBi] = 10log10(Eff%/100))+5.15

We can then use antenna modeling software, EZNEC implementing NEC2 (http://www.eznec.com/), to find close agreement between Belrose calculations and antenna modeling.

(Click on graphs for full size versions)

Vertical monopole over perfect ground	Vertical monopole over radial field	Vertical monopole with umbrella loading

NEC2 modeling yields a reasonable comparison to Belrose calculations for a very short radiator over perfect ground.  We then find reasonable comparison between NEC2 outputs for perfect ground and a modest radial field (with heights 10-15m).  Confidence gained from these two comparisons allows us to take one step further by adding umbrella loading thereby significantly improving complex impedance and radiation resistance values.

Results from the umbrella model can be folded back into Belrose calculations for efficiencies and gains at heights 10-15m:

(Click on graphs for full size versions)

Now that we have gain vs height values we can use SNRD software from Frtiz Raab, W1FR [4] to calculate and graph SNR vs. distance values for three ground conditions, POOR, AVERAGE and VERY GOOD, and 10m vs. 15m antenna height (both with top loading umbrella).  

(Click on graphs for full size versions)

SNR vs Distance for 50% Median Noise Level Poor, Average and Very Good Ground Er & S/m 10m and 15m Monopoles	SNR vs Distance for 90% Median Noise Level +10dB Poor, Average and Very Good Ground Er & S/m 10m and 15m Monopoles

We find that AVERAGE Er and S/m ground conditions are predicted to support reliable 50% Median Noise Level 10dB SNR links out to ~200km with 10m antenna height and ~250km with 15m antenna height.  These average ground conditions are predicted to support very reliable 90% Median Noise Level 20dB SNR links out to ~75km with 10m antenna height and ~90km for 15m antenna height.

A few final thoughts

Antenna deployment issues are considered to be roughly equal.  NVIS requires three inverted-V dipoles at different heights on a 14m mast.  VHF troposcatter requires a 10dBi Yagi on a 15m mast.  600m ground-wave requires a 10-15m monopole with loading umbrella and radial field.  As I imagine antenna erection these antenna system requirements are probably more than a single operator can safely and dependably deploy.

As I imagine station operation, all of these alternatives may be considered roughly equal semi-automatic systems. They would all require occasional operator management to maintain reliable communications links.    

When comparing predicted NVIS, VHF Troposcatter and 600m ground-wave propagation for regional communications we find that NVIS is the clear winner – when considering only a single point-to-point link.  When using 3-bands, optimum antenna heights and 100W, NVIS is predicted to provide highly reliable links out to >400km 24-hours a day.  A significant downside with NVIS is the manual operational requirement to occasionally change operating bands to yield optimum SNR and reliability.  Occasional band changing in a fully meshed network with widely varying peer-to-peer link distances would be operationally problematic.

Medium-wave 600m ground-wave omni-directional coverage would allow fully meshed peer-to-peer communications links but over a smaller regional coverage area than reliably supported by NVIS techniques.  VHF troposcatter use of high gain directional antennas supports only a master-slave star network design.  NVIS has an additional advantage of supporting communications links to relatively compact mobile stations.

Notes:

[1]  Frederick, “Fritz”, Raab, W1FR, The ARRL 500-kHz Experiment: WD2XSH, ARRL QEX Jul/Aug 2007, pp:3-11

[2]  L.B. Cebik, W4RNL, Antenna Options: Narrowband NVIS Antennas, ARRL QEX Jan/Feb 2007, pp:55-61

[3]  J.S. Belrose, now VE2CV, Short Antennas for Mobile Operation, ARRL QST Sep 1953, pp:30-35 + p.108

[4]  Fredrick, “Fritz”, Raab, W1FR, SNRD