Can you hear me now?

I live in the Seattle area, but my QTH is out in the boonies near Olympic National Park. Although the country is beautiful (and wet!), the cell-phone service varies from marginal to lousy. When I look at the phone's "S-meter" I sometimes see one or two bars, so I make the call. But all too often, the signal drops and the service is lost. If it were analog, I probably could hear the person I call, but he might not hear me. How frustrating. But, I am a ham radio operator, so I should be able to figure this out.

Cell-phone frequencies are in the 850- and 1900-MHz bands in the U.S.A. and although I am no cell-phone expert, the principles should be the same as those we use in ham radio. Obviously, the antenna gain of my little handheld is less than that of the big directional antennas on the cell towers. And their transmit power must certainly be higher than mine. Hmmm . . . Maybe if I just move around, or hold my phone at a different angle. "Can you hear me now?"

It seems clear that the two sides of a cell-phone circuit are not the same. All that made me wonder about HF radio reception. I know from experience that sometimes my ham radio contacts cannot hear me, although I can hear them quite well. It's time to investigate this interesting phenomenon.

HF Signal Reciprocity

One would think that ionospheric radio propagation would be reciprocal. That is, the signal strength in one direction should be the same as in the reverse, or reciprocal, direction. In HF ray-trace theory, the distance is the same and the ionospheric control points-the points where the wave is reflected (or more properly, refracted) back to the ground-should be the same. But I decided to run some test circuits using the ACE-HF System Simulation & Visualization software to see if signal predictions were the same in both directions. (See my review of the new ACE-HF V2.05 model in the May 2006 issue.)

ACE-HF permits one to predict both Signal Strength and Signal-to-Noise Ratio (SNR), and I started with signal strength. I set up a couple of sample circuits from my station, using North-South and East-West directions. I used default isotropic antennas with +6 dBi gain at each end, set the transmit power at 1000 watts and assumed SSB communications with a Required SNR (RSN) of 48 dB-Hz and a 50% Required Reliability. The Normal absorption model setting was selected (see my review of ACE-HF's new Absorption Model in the July 2006 issue of CQ Magazine.)

I used the month of August for the analysis because in the northern hemisphere summertime atmospheric noise is higher, and I found that the predicted Smoothed Sunspot Number (SSN) for July 2006 was about 18. After quickly reviewing several ham-band predictions, I settled on the 40-meter band for this analysis. Man-made noise was set at the default Rural level for all the tests.

Fig. 1 - 40-meter signal strength for a circuit to the South:
(click to view full-size version)



For a North-South path, I specified a 3634-km circuit from Brinnon to a maritime station at 15° N latitude, south of my location. Figure 1 shows the predicted signal strength. As expected, the signal is highest at night and decreases as daylight approaches.

By using the new ACE-HF User Mode switch, I then changed from the Ham mode to the Shortwave Listener mode, which simply reversed the circuit so that my station became the receiver. I then repeated the prediction and found the chart to be nearly the same. But to be sure, I used the ACE-HF Bands display to find exact hourly values, as shown in the table of figure 2. These values show agreement within 1 dB as the direction of transmission was changed, so I concluded that for this circuit, 40-meter signal propagation was indeed reciprocal (and the differences were probably round-off errors).

Fig. 2 - Reciprocal 40-meter signal strengths for N-S and S-N paths:

Hour	0	1	2	3	4	5	6	7	8	9	10	11	12	13	14	15	16	AVG
N to S Sig.	15	30	37	41	42	42	42	42	42	41	41	41	40	40	37	31	17
S to N Sig.	15	31	37	42	43	42	42	42	41	41	40	40	39	39	37	31	18
Difference	0	1	0	1	1	0	0	0	-1	0	-1	-1	-1	-1	0	0	1	-0.6

Now I tried a circuit from Brinnon, WA to Bathurst, New Brunswick-a city at about the same latitude as that of Brinnon. This East-West path was expected to show signal strength differences with direction, because at most times-of-day the terminator in August was sweeping through the path and the control points were slightly varied. The results of this comparison are shown in the table of figure 3, where the average difference was less than one-half dB over the 17-hour time period. Again, signal strength was judged to be essentially reciprocal.

Fig. 3 - Reciprocal 40-meter signal strengths for W-E and E-W paths:

Hour	0	1	2	3	4	5	6	7	8	9	10	11	12	13	14	15	16	AVG
W to E Sig.	28	31	33	34	36	37	40	42	40	39	43	39	31	23	6	-5	-18
E to W Sig.	27	30	33	34	36	37	40	42	43	42	42	39	31	24	6	-4	-17
Difference	-1	-1	0	0	0	0	0	0	3	3	-1	0	0	1	0	1	1	-0.4

I knew that one shouldn't judge signal reciprocity by only two examples-results will vary with frequency and by path characteristics. But it doesn't matter anyhow, because signal strength is not what determines circuit connectivity. Instead, it is the received SNR level that controls our ability to make solid contacts.

Circuit Reciprocity

Think of it this way: Imagine sitting in a crowded hall, waiting for the concert to start. You can easily hear a conversation between people a few rows behind you and you can speak to your neighbor in a whisper. The ambient noise in the hall is very low. But now the curtain rises, the conductor appears and the audience begins to applaud-the noise level has risen considerably. You must raise your voice-raise your signal level-to be heard by your neighbor. The noise part of the SNR equation has gone up, so your signal level must also be higher.

The same thing happens when we listen to our radio. A distant signal that we heard at S7 yesterday can't be heard today even though we had scheduled a QSO, because now we are in the middle of a thunder and lightning storm. The noise has risen-but not the distant signal-so the SNR has gone down. In radio communications, SNR is the name of the game, not signal alone.

Unless you are in the middle of an industrial area where high man-made noise levels exist, the principal limitation on received SNR is atmospheric noise that comes from lightning flashes. When you are very near a thunderstorm, your receiver may be almost blocked by interference from lightning. We can illustrate this by again using ACE-HF to simulate a circuit. This time, I specified a circuit from a ham in Keflavik, Iceland to a station in Miami, Florida, as shown in figure 4.

Fig. 4 - Keflavik to Miami circuit:
(click to view full-size version)



I chose this circuit purposely because August thunderstorms concentrate in the Caribbean, and in central North and South America. Thus, as one approaches the Polar Regions atmospheric noise levels diminish and noise at Keflavik should be lower than at Miami.

Let's see what ACE-HF has to say about reciprocal SNR predictions. Figures 5 and 6 show comparative SNR vs. time-of-day charts for this circuit. The first is for the transmitter at Keflavik and the second reverses the circuit. Reception at Miami is marginal, but when the circuit is reversed, a significant SNR increase is predicted. The only thing we have changed is the receiver's location. The better SNR is due to the lower atmospheric noise level at Keflavik.

Fig. 5 - Keflavik to Miami 40-m SNR vs. TOD:
(click to view full-size version)



Fig. 6 - Miami to Keflavik 40-m SNR vs. TOD:
(click to view full-size version)



To quantify this effect, we will again use a table, as seen in figure 7. The SNR difference is striking! For the 13 hours examined, where the predicted SNR was above the "red" chart areas, the average SNR difference at Keflavik was nearly 15 dB. To put this in perspective, to achieve a comparable SNR at Miami would require raising the Keflavik transmit power from 1000 to about 32,000 watts! The effect of atmospheric noise on connectivity is indeed powerful. And, for this circuit at least, connectivity is certainly not reciprocal!

Fig. 7 - 40-meter SNR for Keflavik and Miami reception:

Hour	1	2	3	4	5	6	7	8	9	10	-	22	23	24	AVG
SNR at Miami	44	44	43	44	48	49	51	49	44	38		41	43	45
SNR at Keflavik	59	59	60	61	62	65	64	62	59	53		52	58	60
Difference	15	15	17	17	14	16	13	13	15	15		11	15	15	14.7

Of course, not all HF circuits have such large reciprocity differences. Those where their terminals are in equivalent noise regions would enjoy similar SNR levels, other conditions being equal. To examine this effect more completely, I set up an ACE-HF Circuit Group chart, as shown in figure 9.

The Circuit Group chart permits one to see simultaneous predictions for as many as 18 circuits and is a favored ACE-HF tool for use in contesting and DXing. For example, to achieve one's Worked All Countries award, you might set up 18 circuits from your station to the missing areas. ACE-HF will then compute all 18 circuits for all ten ham bands and all 24 times-of-day. You can then watch the chart-it advances automatically every hour-to see when various bands will be open. Or, you can advance the time setting in order to plan your next call.

Fig. 8 - NW7US 40-meter coverage at 08 UTC:
(click to view full-size version)



In our case, however, I wanted to use the chart to test SNR reciprocity, so I used the chart to specify nine circuits from my station in Brinnon. Then, for each circuit, I used the ACE-HF Ham/SWL user mode to reverse each circuit. The results for each station are shown one after the other in the chart. To get uniform results around my QTH, I first ran ACE-HF area coverage maps to find the best hour for uniform coverage on 40 meters around my station. That coverage at 08 UTC is shown in figure 8.

Fig. 9 - NW7US circuit SNR for reciprocity test group at 08 UTC:
(click to view full-size version)



Returning to figure 9, the Circuit Group chart shows the SNR values for the nine circuit pairs. The green cells show SNRs that are above the RSN, the yellow cells are for values within 10 dB of RSN, and the red cells show SNRs below that level. The best frequencies for each circuit are shown by the blue-colored cells.

Focusing on the 40-meter values (Ch 04), the circuit to Bathurst is reciprocal within about 2 dB as expected. But the circuit to Barrow on the north coast of Alaska shows a difference of 10 dB! Honolulu is quieter than Brinnon by 4 dB, which is to be expected-because the Pacific Ocean is more "peaceful", I suppose.

Comparing the other circuit pairs reveals other reciprocity differences that can be explained by the expected location of thunderstorm centers around the world. At 08 UTC it is nighttime over North and South America and that is when thunderstorms usually occur. For example, reception at Caracas, Venezuela is expected to be noisy, and the chart shows that the 40-meter predicted SNR is 8 dB worse than at Brinnon, WA at that time.

System Factors Affecting HF Reception

There are many system factors that affect HF reception, and all must be understood and evaluated if we are to make accurate simulations. As we have seen, software such as ACE-HF will sort out the signal strength and SNR predictions automatically, but it is up to us to properly specify the system if our predictions are to match our on-air experiences.

A brief checklist of system factors that will affect reception is given in figure 10. As you review the list, think of how you can determine those factors for both ends of your circuits.