HF Circuit reliability prediction with VOACAP and NEC2 antenna simulation

Two separate HF circuits are analyzed. The first, a path from Michigan State University to New Zealand, is shown to be unsuccessful for a reliable voice circuit during 2200-0200 UTC. The second, a path from Michigan State University to the Pentagon's HF station in Arlington, VA is examined and shown to be successful for a voice circuit during 1600-0200 UTC.

MSU<->NZ no success | MSU<->Pentagon success

International HF Circuit--No Success

The reliability of an HF circuit between Michigan and New Zealand in Aug. 2007 is examined.
1kW NZ transmit power and 100W Michigan transmit power is assumed.

It will be shown that voice circuit reliability is expected to be low and efforts will likely be unsuccessful.

We see that the 18MHz and 21MHz bands have at least a 15dB advantage during our hours of interest, which here is 2200-0200 UTC. Thus, we'll focus on 18MHz and 21MHz for subsequent plots.

We see right away there's a problem--the signals are weak, and generally on long paths, low radiation angles are needed, and that mostly excludes low dipoles...

We see that the signal goes from a minimum of -148dBW to -135dBW from NZ to MSU and -159dBW to -146dBW from MSU to NZ.during our time of interest. This means our best reception of NZ at MSU is S3 to S1, and from MSU to NZ is not even S2 to less than S1, based on S9=50µV and each S-unit being 6dB.

This is not the best starting point, because now we will have to have radiation at the optimum elevation angle to complete the circuit at all for a voice-quality circuit.

Our Ansatz is holding--we see low angles of radiation are required.

VOACAP is said to predict a mode of the optimum angle--variations of elevation and azimuth are to be expected.

The 3-element SteppIR yagi claims 7dBi for 18MHz and 21MHz bands. For simplicity let's assume this figure is achieved at one end (assuming yagi height of 50ft.), and that we have about a 15ft. high tuned dipole at the other end.

For the dipole, let's run a quick simulation with NEC2.

We see that at 6 degrees we are already at -7dBi, and less than -10dBi at 3 degrees.
Any gain possessed at the NZ end's SteppIR is canceled by the unfavorable dipole radiation pattern of MSU. For an order of magnitude estimate, we may say the isotropic case is sufficient here.

If we used narrow-band digital modes, such as PSK31 or CW, we take advantage of the following relationship:
This says that our receiver noise level changes with respect to receiver bandwidth--noting the convenient decibel relationship.

Conclusion

While VOACAP can include antenna factors into the predictions, for simplicity we use isotropic (0dBi) antennas for a "first pass." If the circuit analysis justifies a more detailed look, one can be carried out.

In this case, we carried out many of the calculations necessary to sufficiently predict the reliability of an HF circuit. We have found that our weakest transmit link will give less than an S2 signal, and well possible below S1. This means the HF circuit will be difficult if indeed possible at all to maintain voice communications, in an RF-quiet rural environment.

Since the MSU end is known to be in a dense area with expected noise levels of at least the suburban order, it is unlikely that this NZ/MSU HF circuit will be successful at the given times for voice communications.

Domestic HF Circuit--Success

The reliability of an HF circuit between Michigan and the Pentagon in Arlington, VA in Aug. 2007 is examined.
1kW transmit power is assumed for both ends.

It will be shown that voice circuit reliability is expected to be good and efforts will likely be successful.

In this case, the second plot a simplified version of the first with only the bands of interest. We quickly see that the 7MHz and 3.7MHz bands will be of interest. 1.8MHz and 5.5MHz may also be useful, if the path is good enough for 50W ERP allowed on 5.5MHz or if we have enough of an antenna on 1.9MHz (not likely the case at MSU or the Pentagon).

These plots are very interesting, and show the effect of 3.7MHz and 7MHz bands "lengthening" and "shortening" for this domestic path. We notice that in the afternoon, the E-layer of the ionosphere blocks access to the F-layers and so to make our refracted signal come down in the desired spot, we have to use a lower or higher radiation angle. Similar effects are shown for 1.9MHz and 5.5MHz bands, with the 1.9MHz band being "covered" by the E-layer until late evening/early morning, which explains the radiation angles needed.

Here is a 40 meter dipole, 25ft. above a building taken as 60ft. diameter circle of steel-reinforced concrete, about 45ft. tall. This approximation will be used to represent the supporting structure for antennas at each end.

We see that if the antennas were full-size, we'd expect about 6dBi, owing to "ground gain" and other alterations to the dipole's free space pattern. We may feel comfortable continuing to use the signal strengths predicted by the isotropic case.

Going back to our earlier plots, we see that median signal strength of at least -90dBW is expected during our time of interest. This is equivalent to about S9+10, or assuming an S8 background noise level, about 20dB S/N, which is adequate for SSB voice communications. This is approximately equivalent to 54dB SNR/Hz, the unit used by VOACAP. Let's see what VOACAP computes.

VOACAP predicts for 20dB SNR (54db SNR/Hz) a 70-75% time availability for this circuit.
VOACAP predicts for 10db SNR (44dB SNR/Hz) a 85-90% time availability for this circuit.

Thus, we see that we could expect at least a minimally usable voice circuit 85% or more of the time. If we allowed at least an hour window for this contact, we might expect at least to get a basic call sign exchange and signal report, and likely also be able to hold an interesting QSO.

Conclusion

This HF voice circuit between the Pentagon and Michigan State University is predicted to be a success at the desired times, using the 40 meter band.