Sunday, 21 January 2018

The tyranny of default settings

The tyranny of default settings.

Sounds menacing, eh?

Not too much to report, apart from a perfectly predictable 10m Es opening on 19 January following an increase in solar activity. The K number rose to more than 4, as predicted four days earlier by NOAA and also two days earlier by Solarham. This duly produced this result:-
10m Es spots on WSPR at GM4FVM on 19 January 2018
10m Es opening in January at the bottom of the solar cycle - nothing unusual to report there then. I had hope for 20 January too, but the enhanced solar activity faded away.

I have confirmed what was limiting my PC performance when using multiple instances of MSK144 mode. It does seem to be a bandwidth thing as mentioned previously as a theory. I now need just one computer. So the back up computer has gone to the further reaches of the shack, despite all the fettling I did to it.

But what has been intriguing me recently has been strange behaviour amongst some amateurs who use the "wrong" settings for MSK. I have distinctly heard four stations using 30 second T/R intervals, one fairly local, but the others in the 750 to 1000km region. If I can hear these stations in both receiving segments then something pretty serious is wrong.

Always providing that my timing is correct, and I use Meinberg software for timing, then is just not be possible to hear the same station in successive receive segments, unless of course they were transmitting in both. Maybe their timing is wrong. Or maybe they are actually deliberately transmitting in both.

I checked my timing by running two sets of Meinberg on two computers, and verified with the Meinberg NTP monitor that both were functioning correctly. They both showed the same time even though they were synching with different hosts.

At least one of these stations is using 30 second intervals because I can see that they also have "Auto Seq" ticked. I can tell because they transmit for 30 seconds, and then if they get a reply, Auto Seq switches them to the other 30 second segment.

This is bizarre behaviour. They are losing 50% of their receive time, making QSOs on average twice as long. They are also occupying both segments, meaning that any station close to them might as well give up. And half of their transmissions will be lost for a DX station using the standard 15 second interval, causing another 50% loss and making any QSO next to impossible.

I have been pondering all this. Why do such a thing when it will only (at best) lengthen your QSOs, or (at worst) lead to many QSOs failing to complete? And it will annoy every other amateur in your area who will be unable to operate while you are on the air.

Then the penny dropped. The last time I downloaded WSJT-X it came with the default setting as "T/R 30 seconds" and "FTol 20". Of course I reset those to 15 seconds and 100, as the User Guide suggests. Perhaps these people are just using the default settings. Perhaps they never looked at the User Guide.

At this point I could don the cloak of the "Old Timer". I could point out that my enforced period as a Short Wave Listener taught me the benefit of listening first. If I had listened first it would immediately be clear to me that other stations were using 15 second intervals.

However, I think it goes deeper than that. My instinct when downloading WSJT-X is to explore it and find out more about it. Maybe other stations just accept the defaults and think no more about it.

OK, every so often every station gets it wrong and transmits on the wrong segment. The "Tx even/1st" button is small and it is easy to forget to change when you turn the antenna. But to use the 30 second T/R constantly is shooting yourself in the foot - you miss, on average, 50% of the replies. And the other station misses 50% of yours, leaving the statistics of randomness to condemn you to fail most of the time.

In between the times when these stations have been wiping both rx opportunities for me I have worked a few stations during the supposed "dead period" (early January to April) for meteor scatter.
VHF meteor scatter contacts at GM4FVM 9 to 21 January 2018
Meanwhile, the stations using 30 second periods seem to be calling CQ a lot, and not getting many replies.

I wonder why.
On a different tack, the issue about strength reports as opposed to meteor ping duration and number of pings has been mentioned elsewhere. Here is a nice clip from OZ1JXY Henning's 4m QSO of today today . Somehow my report of +03dB, although good, hardly goes far enough to describe this ...
That counts as quite a blast in this quiet period.
Enough from me, but for what it is worth I will post the relevant section from the WSJT-X User Guide which I doubt if anyone involved will read. Ah well.


To configure WSJT-X for MSK144 operation:
  • Select MSK144 from the Mode menu.
  • Select Fast from the Decode menu.
  • Set the audio receiving frequency to Rx 1500 Hz.
  • Set frequency tolerance to F Tol 100.
  • Set the T/R sequence duration to 15 s.
  • To match decoding depth to your computer’s capability, click Monitor (if it’s not already green) to start a receiving sequence. Observe the percentage figure displayed on the Receiving label in the Status Bar:
MSK144 Percent CPU
  • The displayed number (here 17%) indicates the fraction of available time being used for execution of the MSK144 real-time decoder. If this number is well below 100% you may increase the decoding depth from Fast to Normal or Deep, and increase F Tol from 100 to 200 Hz.
Most modern multi-core computers can easily handle the optimum parameters Deep and F Tol 200. Older and slower machines may not be able to keep up at these settings; at the Fast and Normal settings there will be a small loss in decoding capability (relative to Deep) for the weakest pings.
  • T/R sequences of 15 seconds or less requires selecting your transmitted messages very quickly. Check Auto Seq to have the computer make the necessary decisions automatically, based on the messages received.


  1. Hi Jim,
    What in your opinion are the pros and cons of MSHV vs WSJT for FT8 and MSK144. I like the MSHV but the majority clearly use WSJT. I'm guessing this may be a case of if it ain't broke dont fix it.
    73 Dave

  2. David
    That is a tough one.
    The big difference is that WSJT-X offers full CAT control, so you can change band and mode without resetting the VFO. MSHV only runs the PTT via CAT, so you have to remember to set the VFO carefully. And so WSJT can do clever stuff with split frequencies, while with MSHV you have to set split on your rig manually and remember to turn it off afterwards.
    I liked MSHV because it is simpler and very robust, precisely because it does not do the fancy stuff. It is easy to set up because there are no tricky controls to set, as it just doesn't do tricky things.
    There is a school of thought which says that third party packaging of WSJT modes is a really a rip-off of somebody else's idea. On the other hand as the WSJT team made their material open source, what can we expect? There have certainly been a few problems with third party packages releasing advances before they were fully developed.
    Performance seems similar, as is processor load. I ran them both simultaneously on MSK receive and found no difference in performance.
    I suppose another difference is that MSHV still includes outdated modes such as FSK441. On the other hand, WSJT-X deletes modes which have been replaced. Personally, I think that outdated modes should be buried, but if anyone wants to stick with FSK then MSHV is the way to go.
    So currently I only use WSJT-X, but I keep MSHV handy just in case.
    As you suggest, "stick with what works" is a good plan.
    73 Jim

    1. Appreciated and well explained thankyou for taking the time to explain.
      73 Dave

  3. Thanks for the useful info Jim, I have no experience with MSK144 and would have used the default settings I guess. I think I might get a look on this mode in the near future. 73, Bas

  4. Thanks for that Bas. I guess we all trust the default settings to some degree. It is worth trying to listen to MSK on 50.280. Because the signals come in bursts it can take a while to find some. I would suggest a sked but you are in a difficult direction for me and the distance is a bit close. But maybe worth a try someday. 73. Jim

  5. Reading your comments with Bas, have you tried scatter by beaming to a common point rather than at each other. I had a few sucessful MS qso's on 2metres years ago with this, I remember being an Italian Stations first G on MS pointing to a common point because he had a big mountain in the direct path.

  6. Yes, Dave that might work. I was assuming Bas did not have a beam, as his entry mentions a vertical. I think that a lot of contacts made by stations using verticals come via back scatter and side scatter type paths.
    The back scatter method raises issues when I am beaming at 90 degrees and I am called by a station, say, 500km away who I know is at 150 degrees. Do I leave my beam on 90 degrees, or do I point directly at the station, given that they are probably too close and over a hill from me?
    Hard to decide that one. If it is a sked then they know what we should do, but if it is a random contact who knows which way they are beaming? It may be that beaming towards them is the wrong thing to do.
    73 Jim

  7. Ahhh, I see, didnt realise Bas was vertical. I guess if you are hearing them ok on random, leave your antenna where it is, unless they turn theirs toward you!
    Still waiting for my antenna.... yawn.
    73 DAve