Thursday, July 01, 2010

APRS on HF

A post about the French ban on D-Star on QRZ.com has predictably become a platform for everyone who dislikes D-Star to reiterate their views on why it should be banned altogether, not just in France. These reasons boil down either to its use of a proprietary codec or the fact that it involves linking radios using the internet.

I admit to toying with the idea of making a post on the lines of "what if internet linking of radios was banned worldwide? Wouldn't that be a real catalyst for innovation, to try to replicate internet facilities using entirely amateur radio RF based methods.?" But further thought as to what would be the effect on APRS if it were denied the use of the APRS-IS internet backbone made me realize that the effect would be catastrophic. The amount of traffic would be far too great to be carried over any HF network, while the number of hops needed using a VHF network to achieve worldwide coverage would also be far too many and far too slow. You might be able to do something using a sophisticated network of amateur satellites, but that would be far too expensive. While some uses of the internet do devalue the use of radio itself, in my opinion, many of them make possible things that could not otherwise be achieved. The internet is an integral part of APRS just as it is probably an integral part of whatever D-Star is meant to achieve.

Which led me to the question of what exactly is the point of APRS over HF radio? I understand the purpose of the VHF APRS infrastructure, which is to capture the messages from local APRS stations and pass them to APRS-IS. But given that it would be impractical for messages from one side of the world to be conveyed reliably to the other using RF, what is the point of APRS networks on HF?

I started off the day trying the latest version of Cross Country Wireless's APRS Messenger software. This is an interesting product in itself, in that it enables APRS traffic to use various PSK data modes, which are arguably more reliable on HF than the 300baud packet most people use. Unfortunately there are not many users. I switched to using 300baud packet and my screen quickly filled up with callsigns from all over Europe. But when I looked at aprs.fi to see its map of stations received by my station I saw only one.

The reason, I surmised (since I am far from being an expert on this) is that my client software, APRSIS32, is doing "the right thing" and not forwarding most of the messages I received as they had already propagated by the maximum number of hops. The one station whose messages I did pass on, F6KPH-4, had against it the note "Seriously bad path." Following this was the explanation: "This station is transmitting packets with a configured path of over 3 digipeaters. This causes serious congestion in the APRS network and errors when plotting the station's route on a map. Please consider using a path of WIDE1-1,WIDE2-1 or WIDE2-2, or even WIDE1-1,WIDE2-2 if you are moving very far away from an iGATE."

I don't really understand this WIDEx-x business, I just do what I am told. But I think I get the gist, which is that APRS messages sent over HF should be configured to take no more than three hops. If they don't reach a gateway by then, they will be lost, which is just tough luck, because the HF channels can't handle the congestion that would be caused by messages being rebroadcast more than three times.

Which brings me back to the question of what exactly is the point of APRS over HF for the average amateur? It is certainly interesting to see what you can hear using your own equipment. But if it is impractical for an RF based network to ensure that a message could get from Sydney, Australia to New York, New York without touching the internet, what useful purpose is achieved by transmitting and receiving APRS on the HF bands?

6 comments:

Theodore said...

I must confess to being "old school" in my digital working. PSK31/Olivia etc, real people talking to each other.
APRS is an interesting idea, and I understand the "emcomm" ramifications as well as the position reporting over your walks.
But, and I am surmising, what percentage of people really use APRS, Dstar and all these complicated internetworked applications.
If we get a solar flare, asteroid or whatever EMP inducing event may come along, we will be catapaulted into 1850 when the cellphone tower batteries run out.
Rather than filling our days with ever more complicated ways to use a radio, perhaps the simple joy of talking to someone may be better both socially and for emcomms.
I can just hear the braaap braaap of unanswered packets bleating out to non existent digital networks.
Maybe, instead of figuring out where we are with aprs we should be thinking of where we would like to be. Just a thought.

G4HYG said...

Hi Julian, you always raise some interesting questions.
For those who don't know me I'm the author of APRS Messenger.
I've been interested in HF APRS for some years especially when I was working away from home in an area with no VHF APRS coverage. It was a comfort to my family to see where I was using a webpage I wrote to display my APRS location.
After that I had a few ideas about APRS over digimodes bouncing around my head for a year or so and it was only when I saw the rescue effort for the Haiti earthquake victims on TV that I sat down and started coding.
The program so far has only been "advertised" on the APRS special interest group mailing list and a couple of other Yahoo groups but it's interesting to note that most of the regular users come from the more remote areas of the US, Canada and Australia.
The recent development of the program adding the QPSK modes is allowing several amateurs to run reliable long term tests comparing the different modes over HF paths. It's true that there isn't much activity in Europe yet but it's not really necessary to use HF if you have access to good VHF APRS coverage.
In answer to your question why use APRS over HF the best answer is that APRS isn't just a mobile tracking system. You can send APRS messages with acknowledgements to confirm reception without the remote operator having to be sat at the keyboard. Internet gateways can take those messages, send them to the internet and even forward them as an email with a correctly formatted message.
It's also open source without any closed codecs!
73,

Chris, G4HYG

Julian said...

Thanks for your comments. Chris, I agree that APRS is more than just a mobile tracking system, but when I try to use it for more than that, more often than not the system seems to be broken. It will be easier to explain in another blog post, which I will do shortly.

Dave said...

I think it would be neat if someone modified a protocol like WSPR for APRS use in HF. It wouldn't take that many more characters to be sent, and even though 'packets' might take minutes to send, you could fit loads of them within a few hundred Hz, and recover them in poor SNR conditions.

oh7lzb said...

aprs.fi doesn't really know about HF or VHF, or which station operates on which.

But yes, it knows that a 7-hop path of TRACE7-7 does not make sense on either VHF or HF. If everyone did that, the network would, in most places, die. The places which wouldn't die is the ones filtering such long paths out in digipeaters. So, the advice is to use a maximum of 3 hops on VHF.

I would guess that on HF you should use even shorter paths than on VHF, because the transmission speed is much slower, and the coverage is larger. Unless you have really few users there, in which case it doesn't matter so much as the risk for collisions is small. A WIDE7-7 path on HF calls for a couple of roundtrips around the world for that packet, and that can't possibly work out with more than a couple users, when collisions multiply with the repeating.

Steve said...

I can think of one useful use of HF APRS. It would be to fill in gaps in mountainous terrain, where valley to valley communications are important. Sometimes these national parks are beauty spots etc and as such prevent hilltop vhf repeaters due to planning restrictions. The ideal frequency would not be 10 Mhz. It would be much lower medium wave to 5 Mhz and low dipoles would do the job.
Steve
G4AZB