AmateurRadio.com

Welcome to the first Linux in the Ham Shack for the new year! 2011 is upon us and we wanted to present our listeners with a fantabulous new show. That’s what you’re going to get. We were able to entice the lead developer of FLDIGI, W1HKJ, to come on the program and give us his personal insights into this best-of-class Open Source amateur radio software.

Yes, we know we talk about FLDIGI a lot in our program. We try not to beat it to death. But if you use this program, there is so much information packed into this interview, you’re going to learn at least a dozen things you didn’t know about FLDIGI, even if you use it every day. So sit back, strap in, and join us on this thrill-packed ride into the dark and seedy innards of your favorite digital mode app. We were impressed; you will be too.

73 de The LHS Guys

Due to having been banned from using the software I have not been keeping up with what is going on in the development of the ROS digital mode. However there have been a few interesting postings about it. In the digitalradio Yahoo group Skip KH6TY has posted the results of some tests conducted with ROS on 432MHz which appear to show that it suffers badly from the effect of doppler shift and flutter experienced at those frequencies, failing to decode over paths where Olivia was successful and even SSB was readable.

This has prompted a rebuttal from the ROS author, which however seems to overlook the problem of doppler distortion encountered by Skip. He has posted a series of comparisons between ROS 2250/8 and Olivia 32/1000 which purport to show that ROS holds up while Olivia prints garbage. He concludes: “The difference between both systems is about 5dbs (3.16 in natural units). This means that ROS8 need 3.16 times less power than OLIVIA 32/1000 to establish a QSO to 150 characters/minute.”

Assuming that this is true, I nevertheless feel that a tradeoff of bandwidth for power or speed is inappropriate in the context of the narrow HF band allocations for digital modes. Most amateur QSOs do not need to go at 150 characters/minute (most people can’t type that fast). On the other hand the 2250Hz wide ROS transmission blocks three channels that could be used for Olivia 32/1000, and even more channels that could be used for a narrower mode. The use of 2250Hz ROS effectively limits the number of people who can simultaneously hold a digital QSO.

Even if it is true that Olivia needs 3 times the power than ROS to get through, Olivia is still a better choice of mode in the real world, because it is easier to increase the power 3 times or to switch to a slower mode than to find extra space within the HF allocations to accommodate the use of such a wide mode.

ROS would be less of a problem if people used it only in circumstances where it would not be possible to communicate using a narrower mode. Unfortunately that discipline does not exist among today’s radio amateurs. People are using ROS to make contacts with others whose signals are strong enough that 30Hz wide PSK31 could be used. This is just selfish, and it is the reason why I feel that such a wide digital mode should not be permitted on HF at all.

Suppose you were a member of a village football team that had practised and played on the village green for years. And suppose that one day you turned up for a game and found a new rugby team using your pitch. You’d be pretty annoyed, wouldn’t you? Even though the village green is common land and so legally there for all to use, the football team would not expect its use of the football pitch established over many years to be usurped in this way. So it isn’t all that surprising that when it is, there’s a punch-up.

Turn now to the amateur bands and this is precisely what has happened to users of the Olivia data mode. Someone has turned up with a new game called ROS that requires a much larger pitch and it is interfering with the Olivia users’ ability to play Olivia.

OK, you say, but surely no-one could object to the rugby team using the football pitch when the football team isn’t using it? It’s a fair point, although as the football team uses the pitch off and on throughout the day they wouldn’t be happy about it. But here the analogy starts to fall down, because not only do the football team (Olivia) and the rugby team (ROS) not speak the same language, but they are also blind so they can’t see each other to ask even if they were able to.

The only way for the two teams to both play on the village green without falling over each other or resorting to fisticuffs is for each of them to have their own, separate pitches. Now could someone please translate this into Spanish and show it to the coach of the rugby team?

As someone who has quite often used WSPR, I have often said that it would be nice to be able to exchange reports with other stations and confirm the “contact”. I felt this simply because the WSPR network is a friendly community and it feels right somehow to be able to do that, just as you would confirm an enjoyable contact on another mode.

Modes like the little used WSPR QSO mode and JT65A on HF offer the chance to make two way QSOs using similar power levels to WSPR. But my recent experience using JT65A on HF, and more recently watching a station in Iceland take half an hour to try to exchange a signal report and confirmation with a station in Brazil using a certain other weak signal mode led me to wonder if in trying to make it possible to make “contacts” using the least possible power we have thrown the baby out with the bath water.

Communicating via moonbounce (EME) on VHF has always been about exchanging the barest minimum of information, because even doing that is a major achievement. But on HF it is always possible to have a proper contact, even if it means using a bit more power or waiting until propagation is a bit better. So why are we endeavouring to use modes designed for EME on the HF bands? What is actually being achieved? Someone’s computer is able to pick a few characters of information out of my barely audible signal, with the help of heavy error correction and the fact that the message format is known. My computer is able to do the same with his. Is this actually a contact?

Yesterday I tried for the first time in several years to use the Olivia digital mode. Before I did, I Googled up some information about it, and came across a Yahoo group containing a post by Waldis, VK1WJ. He wrote: “Yesterday I had a sked with DJ2UK on 20m in JT65A. Bert came in with around -17dB but he couldn’t decode my signal. After a while I saw in SPECJT that he had switched to Olivia 8/125. So I did the same, and we had quite a long error-free QSO. Olivia 8/125 is not exactly fast, but it still beats JT65A hands down. May be we could entice our JT65A friends to try Olivia instead?”

According to Waldis, instead of exchanging a couple of numbers using JT65A, you could have an actual (if slow) conversation using the 8/125 variant of Olivia. But the JT modes are currently in vogue, whereas Olivia – being developed in 2003 – is yesterday’s news. People are raving about the capabilities of a certain other mode that is making a lot of news recently. Have all these people actually tried some of the forgotten modes? Because I think if they did they might wonder what all the fuss was about and whether the newcomer really justifies its use of bandwidth.

What was I saying about reinventing the wheel?

Today I spent a couple of hours making some contacts using the Olivia digital mode. It isn’t the first time I have used the mode. I used it several years ago back when I used the MixW software. I seem to remember that it produced good copy even when using 5W from my K2, but it was hard to find other people to QSO with. But now there seems to be plenty of activity. Olivia even has its own website.

Although using Olivia is much the same as using PSK31, using the same software, it has quite a different feel. As with PSK there are different versions of Olivia. I was using Olivia 32/1000 which on 20m is normally found around 14.106MHz. This mode is 1kHz wide and operators use USB with the suppressed carrier frequency on an exact 1kHz multiple, e.g. 14.105, 14.106 or 14.107. The center frequency of the waterfall is set on 1500Hz, so the audio tones are generated between 1 and 2kHz, which is in the center of the passband with most radios. If you do this and your rig is accurately calibrated you should hear anyone who calls on that frequency without any need for re-tuning.

Olivia 32/1000 prints out slower than PSK31 which makes for nice relaxed contacts. People tend to chat using the keyboard instead of exchanging macros, though they may use a macro for the basics. Olivia uses forward error correction, so you may not see any print for several seconds after a transmission started and it may not finish printing until after the transmission stops. But the copy is usually 100% perfect even when the signal dips into the noise so you can barely see it in the waterfall. I swear I’m not biased but I thought Olivia performed much better than another mode that is no longer mentioned by name in G4ILO’s shack and is more than twice as wide.

My first contact was with Andri VE2AHS in Ottawa which was 100% perfect copy throughout. I don’t believe I would have made this contact using PSK31 at the same power level (25W into my attic dipole) nor do I believe I would have had such good copy using the other mode.

After lunch I was called by Fred OH/DK4ZC. Fred is very interested in digital modes and was apparently involved in trials of Chip – a mode similar to the latest one – with Nino, IZ8BLY back in 2004. He told me the trials were abandoned as the results were poor. He asked me what I thought of the new mode and of course I told him. Fortunately my K3 has very robust cooling and can withstand long periods of operation in continuous duty cycle modes. 🙂

Next I had a 40 minute chat with John, W9CY near Peoria in Illinois. Again, it is just unimaginable to have such a long ragchew with the USA using any other digital mode. John had heard interference from a strange new mode recently and asked me if I knew what it was so the K3 PA got another warm-up as I filled him in on the details. In fact, some fairly strong QRM from the unmentionable mode did appear during one of his overs but although I did lose a few words Olivia survived the interference pretty well, and better than the new mode survived Olivia.

Finally I was called by Adriano, IZ1PSS, who informed me I was his first contact on the Olivia mode. He had a good strong signal so I left him the frequency as I had been on the air for quite a while by that time.

I really enjoyed my session using Olivia – so much so that I intend to spend a lot more time in future using the mode. The ability to hold long rag-chew keyboard QSOs with near-perfect copy over long distances using relatively low power and an indoor antenna isn’t something that I thought I could do. We hams seem rather good at trying to reinvent the wheel when we already have several perfectly good ones that we hardly ever use.

See you on Olivia!

That’s right, everyone. We’ve hit Episode 10. I know it’s hard to believe. We can hardly believe it ourselves. This one has been sitting in the editing room for a while due to life conflicts once again. It also spent a lot of time going through the editing machine. All that said, it’s finally produced, mostly coherent, and occasionally informative.

As always, thank you so much for listening to the program. Please help spread the word about Linux in the HAM shack by tweeting about us, posting on your blogs, telling your fellow hams and just getting the word out. We appreciate it, and we’ll do the same for you if you send us your information.

Tons of feedback in this episode and then we tackle digital modes from RTTY to Throb. Enjoy.

73 de Russ