Propagation on the DGPS/NDB Band (200-400 kHz)

[ChrisSmolinski]

About a year ago wrote Amalgamated DGPS ( http://www.blackcatsystems.com/software/dgps_decoding_software_sdr.html ) which can decode all of the DGPS channels present in an SDR I/Q recording, and I started doing nightly recordings of the DGPS band, which runs from 285-325 kHz, to see what stations come in each night, and of course to try for the most distant DX stations possible.  This is right at the defined (but arbitrary) separation between the Low Frequency (LF) band which runs from 30-300 kHz, and the Medium Frequency (MF) band which runs from 300-3000 kHz. So in theory it should share propagation characteristics of both bands.

Several stations share most of the DGPS channels. Often there are 2 or 3 US stations on a channel, and perhaps 1 or 2 from Canada. Then also stations elsewhere in the world Skeezix and I are the two most prolific DGPS DXers here I think. He gets stations from Alaska and Hawaii on a regular basis, and I get the West Coast on a regular basis, and in the winter Azores/Madiera most nights, and rarely continental Europe. I have never received Hawaii, I do not think Skeezix has heard Europe, and neither of us have picked up stations from elsewhere in the world, with the exception of Panama (which is not that far away).  Actually let me repost the message I posted earlier today, which got me to thinking about this, as it has some distances:

Interesting how you only get three daytime stations, all under 500 km. I need to see what I can get here in the daytime.

I just realized how much closer to Alaska you are. The Alaskan DGPS stations are mostly around 3,000 km from you, Kodiak is 4,167 km. Upolu Point, Hi is 6,285 km, I guess that is your furthest catch?

The west coast stations are around 3,600-3,800 km from me, Alert Bay is 4,016 km. Level Island Alaska (which I have heard) is 4,415 km from me, and Kokole Point Hawaii (which I have not heard) is 7,874 km. Azores is 4,122 km, and Madeira is 5,374 km. Portugal is 5,648 km and Helgoland Germany is 6,233 km, I think my furthest catch - a brief 19 second opening with 3 decodes, on a split channel which helped I am sure. Hmm, Koblenz Germany was also received once with 3 decodes but over 12 minutes. But also on a split channel. 6,385 km.

In many ways, the Alaskan stations are a similar distance from you as the west coast stations are for me. And Hawaii for you is about the same as Europe for me. Also these both are our longest distance catches, and they are very similar in distance. Your path to Hawaii has to travel over a bit of the US first before the Pacific Ocean, I have a shorter distance to the Atlantic, but then for stations from continental Europe there is a bit of land again. You do get Hawaii more consistently than I get Europe, and with many more decodes. A lot of this is a function of what other stations the desired target has to compete with (part of my motivation to build the loop)  

I find it interesting how comparable the distances are, and whether they represent the maximum range possible. There's also the transmitter power (and antenna pattern) of each of the stations. I believe Canada has coverage maps on their site.  How many of these stations have omni directional antennas, how many are purposely directional, and how many are accidentally directional due to local terrain?  I believe there is a west coast DXer that routinely gets DGPS from Australia. Quite a long haul, but if it is only over the ocean, that helps a lot.

I need to read up a bit on MW/LW (DGPS sits right at the intersection of the two) propagation. I wonder what distance is covered by each hop, and is roughly 6,000 km a maximum because it is several hops already, and any further is attenuated too much? Especially for the power level, I believe these stations are typically a few hundred watts. vs the megawatt (or at least hundreds of kW) that the broadcast stations use. NDBs are at roughly the same power as DGPS stations I think, although CW is going to be a more efficient mode. If only decoding of it over the entire band could be automated.

I realized I needed to learn more about LF/MF propagation. Some time ago I read this US Navy paper on VLF/LF/MF propagation, but it was more focused on VLF: http://www.navy-radio.com/manuals/0101-1xx/0101_113-02.pdf

Doing some searching on the web, I stumbled on the BBC Research and Development Department of their Engineering Division, which has over 2K papers online: http://www.bbc.co.uk/rd/search?Type=Publications

You can search, and I found three papers of particular interest regarding LF and MF propagation:
http://downloads.bbc.co.uk/rd/pubs/reports/1973-13.pdf
http://downloads.bbc.co.uk/rd/pubs/reports/1974-03.pdf
http://downloads.bbc.co.uk/rd/pubs/reports/1983-05.pdf

I strongly suggest reading all of them (and maybe checking out some of the other papers, which I need to do as well).

Some things I learned (which may not be new to others):

Propagation in this region is mostly via the E layer. And sometimes, in winter, even via the D layer in daytime. The upper end of the MW band is sometimes propagated via the F layer.

You can get roughly 2,000 km to 3,000 km maximum per hop. Maybe a little more over ocean paths.  That means that our very long distance DX catches of over 6,000 km are either pushing the limits of 2 hops, or are 3 hops. As they are over partly to mostly water paths, I suspect the former? Several of the papers discuss the path losses with each hop, due to various factors. It is easily in the 20 or 30 dB (or more) range per hop. So it really adds up fast. When I look at the signal strength of these DGPS stations, I can see how they are fairly strong with one hop, very weak at two hops, and would be incredibly weak at three. I now suspect this is our limiting factor, and why it is very unlikely we'll be able to get some of the really distant stations. Also from those papers, you (and the station) need to be very close to the coast to maximize the benefits of a hop over water. This likely explains how one DXer I know of who lives in CA is able to pick up DGPS stations from Australia. This likely also explains why MW DXers go to Cape Cod or the Outer Banks for to try for trans Atlantic DX. There's a big difference being right on or a mile from the coast, vs 200 miles from the coast. You can't think of it as "oh, what's an extra 200 miles on a 3,000 mile path?". It's not linear like that.

Which means the chances of me getting a DGPS station from Australia is slim to none. And perhaps the same for Hawaii. Hawaii is about 7,900 km from me. If I check the path, something like 3,800 km is over the US, and the rest, say 4,100 is over the Pacific. There is a really nice graph showing how many hops required for a given distance. (but it could be slightly longer I think for hops entirely over the ocean) That is most likely 4 hops for me. So it is not going to happen. Australia is over 15,000 km away. For someone on the west coast it is around 11,000 km, but the path is entirely over water. So it presumably could be accomplished with 3 hops.

I've picked up Level Island before (at least once, I forget how many times). Obviously the transmitter is on the water (hence the Island part of the name) but the rest of the path is over land. It may well be near the limits of what I can receive from Alaska. Biorka AK on 305 kHz is 4,564 km away.  There are no other US/Canadian DGPS stations on 305, so there must be an NDB blocking it, or some other QRM source, as it is just about the same distance as Level Island. Or I guess the power is less or the antenna pattern is different, or...  

Europe is also at the extreme limit for me. I am probably restricted to those on the coast or islands, but perhaps that is most of them?

[skeezix]

Lots of great info here. I grabbed the PDFs and will read them.

I wouldn't necessarily give up on Australia yet. Kilokat picked up their MW stations recently. Granted it was very weak and those stations (probably) put out more power, which helps. Same with Hawaii.

I don't remember getting Europe, but I think I got Greenland or Iceland. And yeah, Hawaii is nearly a regular in here outside of peak summer. AK comes & goes. There's a few that I find interesting- When I have a bunch come in (as was the case a couple of weeks ago), there are a few that are within a few degrees and not terribly far apart, yet the decodes can vary wildly.

From Aug 1, 2017 UTC 2330 - Aug 2, 2017 UTC 2335

1926     897  294  295  313.0 100  Kodiak, AK                     Alaska               57.621   -152.201 4,167  311
1        896  292  293  310.0 100  Kenai, AK                      Alaska               60.681   -151.35  4,094  316
205      892  284  285  288.0 100  Gustavus, AK                   Alaska               58.427   -135.706 3,200  313
88       890  280  281  305.0 100  Biorka, AK                     Alaska               56.858   -135.546 3,168  310
3094     891  282  283  295.0 100  Level Island, AK               Alaska               56.467   -133.098 3,014  309
48       889  278  279  323.0 100  Annette Island, AK             Alaska               55.069   -131.6   2,901  306

Why such a discrepancy? Site location? Power? Pattern? Different angles? Probably combination of all of the above.

[ChrisSmolinski]

1926     897  294  295  313.0 100  Kodiak, AK                     Alaska               57.621   -152.201 4,167  311
1        896  292  293  310.0 100  Kenai, AK                      Alaska               60.681   -151.35  4,094  316
205      892  284  285  288.0 100  Gustavus, AK                   Alaska               58.427   -135.706 3,200  313
88       890  280  281  305.0 100  Biorka, AK                     Alaska               56.858   -135.546 3,168  310
3094     891  282  283  295.0 100  Level Island, AK               Alaska               56.467   -133.098 3,014  309
48       889  278  279  323.0 100  Annette Island, AK             Alaska               55.069   -131.6   2,901  306

Why such a discrepancy? Site location? Power? Pattern? Different angles? Probably combination of all of the above.

I agree, all of the above. Also, more and more, I am convinced other stations on the same frequency often are a major factor. I went through the various DGPS frequencies, and noticed that for 305 kHz, Biorka Alaska is the closest DGPS station to me. Granted closest means 4,564 km, but you'd thinkI'd have heard it a few times. The next nearest stations are in Brazil. Well, it turns out there's a carrier on 305 kHz, probably the NDB in Ontario. Likewise, checking some of the other DGPS channels reveals the same problem. Now that Millers Ferry is gone, Richmond BC is the nearest station on 320 kHz. And I routinely get decodes from it, but not all the time. Again, there's an NDB carrier there. 317 and 323 have local NDBs that are quite strong. Although I used to pick up St Paul / Alma on 317 before it went QRT, although MN is much less of a haul than Alaska. 308 has an NDB carrier, I presume from GA, but I do occasionally hear Azores, so again, it is possible to overcome these QRM sources. I suspect that destructive interference of multiple skywave signals might be our friend, occasionally nulling out a pest. Perhaps this is the real cause of the brief bursts of a few decodes we observe from a station? It's not that it faded in for that brief period, but rather the interfering station faded, or was nulled, out. In theory we could do this ourselves via two antennas and phasing. But that's something that would likely require continuous tweaking. Maybe it could be automated?

Split channels (0.5 kHz) offer our best bet for DX targets. No domestic DGPS stations, and the NDB problem should be minimal. No carriers, worst case the MCW sidebands from Canadian NDBs are close, at either 0.4 or 0.6 kHz for stations just below or above our split channel respectively. That should be outside of the passband for the DGPS MFSK signal. And I do occasionally pick them up from Europe. But not that often. They may be at the extreme edge of what is possible to receive (distance wise) given the power level of DGPS stations and the number of hops required. I've been pondering any potential additional software tricks to squeeze some more decodes out of the noise, but am presently out of ideas.

[ChrisSmolinski]

I wouldn't necessarily give up on Australia yet. Kilokat picked up their MW stations recently. Granted it was very weak and those stations (probably) put out more power, which helps. Same with Hawaii.

Yes, I saw  that. Amazing. I tried yesterday morning, while Ozy Radio was coming in well on 5045, but could not get even a carrier on any of the .au MW channels. Granted I was using the sky loop down on the lower half of MW, where it is not very sensitive.

I checked, and they all seem to run 50 kW like high end US MW stations. So not the MW level used by Europe and China, but still around 2 orders of magnitude more than the typical DGSP station? I checked the Southern Avionics site, and the DGPS transmitter they sell is 50-1000 watts. I thought I read that most of the US/Canadian DGPS stations are a few hundred watts? I have no idea if that is typical for stations outside the US as well. That extra power certainly gives you a few more propagation hops before your signal crashes into the noise floor. And even then, it's a rare catch. An upside to DGPS DXing is we only need two decodes in a row to declare reception of a station 

[ka1iic]

Some time ago I read this US Navy paper on VLF/LF/MF propagation, but it was more focused on VLF: http://www.navy-radio.com/manuals/0101-1xx/0101_113-02.pdf

Chris,  as you know the Navy manuals are directed more for VLF because of the Navy nuke subs, with one of the transmitting sites at Cutler, Maine.  I made a simple converter for VLF some years ago and the Cutler station was incredibly loud at my QTH 100 or so mile N/W of them :-)

I was always told that these lower frequencies need an 'excellent' ground... note I didn't say 'good' :-)

73
vw

[skeezix]

Yes, I saw  that. Amazing. I tried yesterday morning, while Ozy Radio was coming in well on 5045, but could not get even a carrier on any of the .au MW channels. Granted I was using the sky loop down on the lower half of MW, where it is not very sensitive.

I need to make note of 5045 and try for it here.

I checked, and they all seem to run 50 kW like high end US MW stations. So not the MW level used by Europe and China, but still around 2 orders of magnitude more than the typical DGSP station? I checked the Southern Avionics site, and the DGPS transmitter they sell is 50-1000 watts. I thought I read that most of the US/Canadian DGPS stations are a few hundred watts? I have no idea if that is typical for stations outside the US as well. That extra power certainly gives you a few more propagation hops before your signal crashes into the noise floor. And even then, it's a rare catch. An upside to DGPS DXing is we only need two decodes in a row to declare reception of a station 

Nautel sneaks up to 3kW
http://www.nautelnav.com/solutions/dgps/

However, what the stations really put out... 

I did find something with power levels a couple of months ago. Few hundred up to 2kW-ish. IIRC.

Not sure why this information is so hard to find, not like its national security secret since its a beacon with a well known position whose entire purpose in life is to broadcast position information continuously.

Some good general info, however, I wonder how much has been implemented seven years later.
http://www.gps.gov/cgsic/meetings/2010/hernaez.pdf

They note moving from the Southern Avionics Company to Nautel Vector-D transmitters. Nautel makes nice MW broadcast transmitters too... with C-QUAM.   

[ChrisSmolinski]

The question of NDB's modulation frequency (400 Hz and 1020/1030 Hz) came up in another thread. Basically, the IACO specs are:

3.4.5.4 The frequency of the modulating tone used for identification shall be 1 020 Hz plus or minus 50 Hz or 400 Hz plus or minus 25 Hz.

1030 is well within the generous 1020 +/- 50 Hz tolerance. But why 1020, which seems like a dumb choice, honestly. I can't find anything regarding the history of how it was chosen. I asked a private pilot (who is also a ham), he does not know either. He did point out that the NDBs are typically only used within 10-15 miles, so interference from other beacons is unlikely. He also said:

Current FAA policy is to NOT repair failed NDB's as they are being phased out in the US in favor of GPS approaches which require zero ground equipment. So as they stop working....they're gone - and the geostationary WAAS satellites are replacing the DGPS LF stations as well.

Yes. GPS for everything. I can't see any problem with that. What could possibly go wrong?

Anyway, the question came up (for me anyway) while thinking about how to automate NDB decoding, much as Amalgamated DGPS does for DGPS.  I posted the entire ICAO NDB spec here: https://www.hfunderground.com/board/index.php/topic,37057.0.html

[skeezix]

I asked a private pilot (who is also a ham), he does not know either. He did point out that the NDBs are typically only used within 10-15 miles, so interference from other beacons is unlikely. He also said:

"Current FAA policy is to NOT repair failed NDB's as they are being phased out in the US in favor of GPS approaches which require zero ground equipment. So as they stop working....they're gone - and the geostationary WAAS satellites are replacing the DGPS LF stations as well."

Yes. GPS for everything. I can't see any problem with that. What could possibly go wrong?

They're also decommissioning around 1/3 of the VORs:
https://www.faa.gov/nextgen/update/general_aviation/approach_procedures/


Don't have thinking people that keep pulling out all of these other system and narrowing down to one system that is not all that secure.