开云体育

A group of us ran a test yesterday with Winlink VARA-FM on 144.950. The path was 14km with a single distinct peak in-between 145m above the northern site and 60m above the southern site. Initially we could send a 90kbyte file P2P between the two locations in about 2 minutes running between level 7 and 9 in FM wide. North was using a commercial J-pole about 8 feet off the ground, while south was using an Arrow 3-el Yagi on a pole (I don't know the height but was it was elevated). We recalibrated before almost every test.
?
A third site (west) was 9km and 13km from north and south respectively. Calibrations yielded SNRs of 23 to 27 dB in all cases.

North and south could initially exchange files at good speed as stated above. We then tried north to west and found that west could not send to north while north could send to west at good speeds. The former would start at level 9 and end up at level 1, with no data transferred at the higher levels.? The latter would send the same file about 3 minutes.

We then elevated north's antenna to 8m, and tested again. It started to rain, which may be significant. We found that south to north did the same thing as the west to north test, starting at level 9 and dropping to level 1, and sometimes disconnecting. North could send to south without issue. We changed back to the original antenna configuration with no improvement. The elevated antenna showed the best signal strength and SNR in calibration.

South tried sending to west and had the same problem.

While with the variable terrain and probably knife-edge diffraction, we're surprised at the extremely variable results without corresponding variation in signal SNR, and also the extreme difference between forward and reverse directions. We're going to review and set up another test, but we would appreciate any suggestions, comments, or similar experiences from others that might account for this?

73,
Chris VE3NRT?

_._,_._,_

---

?Can you tell me where you actually did these tests??? I.e. locations or actual coordinates of the three sites.? I'd like to model it on a? topo map program.? Having spent decades dealing with mobile data on VHF in mountainous terrain (southern California / Los Angeles) for both amateur and public safety radio, I'm curious.? But some questions and random thoughts....

1)?? What kind of radios were you using???????? How much power??? What kind of IF bandwidth??

2)?? What kind of interfaces between the radios and VARA computers? ? Were they connected to the radios via a dedicated flat-response input such as the 6-pin mini-DIN "data" port, or just a mickey-mouse mic & speaker hookup??

3)?? What kind of terrain was this??? Urban areas with multi-story buildings? Suburban low-rise? Open grass lands/farms?? Dense forest? or what?

4)?? You didn't mention what kind of antenna was at the 3rd site.??? The Arrow antenna is of course directional. Was it aimed directly at the 2nd station with the omnidirectional j-pole???? How far off the beam of the Arrow was the 3rd station???

5)??? High-speed data is incredibly intolerant of multi-path reflections.? When signals arrive both directly from the sending station,? and over longer paths after being reflected by tall buildings, water towers,? bare (un-vegetated) hill sides, etc off to the side of the direct path,? they arrive at the receiver at slightly (nano-seconds or micro-seconds) different times.? [Hills with grass and trees tend to absorb VHF/UHF signals while bare rocky hill sides tend to reflect and scatter them.] ?? What should be successive data symbols in the transmission wind up overlapping each other due to the time delays on the longer indirect paths, smearing and mangling the data.? [This is exactly the same phenomenon as "ghosting" in classic analog NTSC broadcast TV. You don't see it in digital TV because the screen just goes black when the multi-path-induced errors overwhelm the built-in forward error correction of the digital signal.]

When the error rate starts overwhelming the built-in forward error correction of the transmission, the VARA program responds by slowing down the transmission rate so that the overlap becomes a smaller percentage of each packet. Which of course reduces the net data throughput.

6)??? Under normal conditions, the direct-path signal should be overwhelmingly stronger and should "cover up" any indirect-path signals. But you seem to say even the direct path was grazing a hill-top, which can radically reduce the signal strength (like 20-30 dB less!) compared to a true literal optical line-of-sight path.? The result is that the direct path signal might not be any stronger than indirect-path signals reflected from objects off to the side.? So NO decisive capture by one version of the signal.?
?? The only real solution to this problem is to use high-gain HIGHLY-DIRECTIONAL antennas at both ends of the path. Note that? more power or omni-directional gain antennas WILL NOT fix this problem - the RATIO of desired to undesired signals doesn't change.?

6) ?? I don't know for sure, but I suspect that VARA's reported initial data throughput starts out as a "wishful thinking" value, that is then corrected downward to more realistic (lower) values as experience with NAKs/ACKs accumulates during a transmission.

7)? ? Further, after being diffracted by grazing a hill top, signals tend to become stratified - formed into thin layers vertically with alternate strong and weak layers.? Sometime moving the receive antenna only a few feet/meters vertically will change the receive signal level 10 or 15 dB. Counter-intuitively, sometimes moving the antenna LOWER will improve the signal! ? ?
? ? Not to mention that shallowly grazing a non-line-of-sight edge can actually change the polarization of the signal. I.e. a signal arriving at the edge vertically can leave the edge as a horizontally-polarized signal.? (A lot of the "mobile flutter" on distant (non line-of-sight) mobile signals at VHF/UHF is actually constantly-changing polarization.) ?

In the late 1970s, I worked at Collins Radio in Cedar Rapids, Iowa. At the time, repeaters were not ubiquitous as they are today.? People routinely used 100-150-watt-plus amplifiers on 2M mobiles to work simplex 50-100 miles (80-160 Km)? across the gently-rolling terrain of Iowa to fixed stations with 8-element or more beam antennas. ? One normally expected mobile flutter and rapid fade-outs/fade-ins as mobiles moved down Interstate-80 at 70 MPH / 110 KMh.? I used two 8-element KLM? beams vertically side-by-side fed in phase with a divider harness to produce vertical polarization.?
??? I then tried mounting the two antennas one leaning 45 degrees to the left and the other 45 degs to the right (i.e. 90 degs apart) and fed them with a phasing harness with an extra 1/4-wave of coax on one leg. ? This creates circular polarization instead of planar vertical-only.? The effects were spectacular - the mobile flutter and fading on long simplex paths totally disappeared!?? Note that you DON'T need to have CP antennas at both ends of the path - the normal vertical-only mobile whip at one end of the path will work as long as the other end of the path has a CP antenna.??

I had the same experience with a two-meter repeater covering a narrow canyon road in Los Angeles. The narrow winding rocky-walled canyon road was a nightmare of multi-path phase distortion and rapid-fire fluttering when the repeater at the summit used the usual vertical gain antenna.? I switched the repeater to a circular-polarized? crossed-yagis? antenna intended for satellite tracking pointed down into the canyon.?? Again, the results were night-and-day - the flutter and spattery audio phase distortion on mobiles in the canyon completely went away.? ? ? ?? ?? ??

---

Stephen H. Smith??? wa8lmf (at) aol.com
Skype:??????? WA8LMF
EchoLink:? Node #? 14400? [Think bottom of the 2-meter band]
Home Page:?????????

APRS-over-VARA igates now operating on 30 & 60 meters
??

"Studio B" Ham Shack on Wheels
??

-- APRS over VARA? --
??

?