Thursday, 15 August 2019

Perseids and FT8, plus MSK144 T/R 15 seconds and other settings.

EDIT - see comments below, many more European amateurs have recently decided to use 30 seconds for T/R on 2m, while on the other bands 15 seconds is still used. This is stupid and will lose them contacts, but being stupid seems to be the order of the day. I cannot blame anyone for listening on the bands, hearing 30 seconds, and carrying on using it. Bear in mind though it will lose you contacts. However, the two systems do not mix so we need to work this out soon. Meantime I am off 2m meteor scatter. I have plenty of other things to do.
==========================================
 If you want to know how come I am qualified to say all this, the answer is at the end of this posting.

Screenshot of the MSK display at GM4FVM - 3 bands on MSK, all on 15sec T/R of course
There is a sight for tired eyes.

As usual click to enlarge if you wish. There is plenty to see in this one, DK4TG is being received on 2m, F6KBF is clearly using 30 second T/R, nothing much is happening on 6m apart from the washing machine making noise, and I am calling OH1MLZ on 4m. Because OH1MLZ is on the odd period I am having to call him on even, though as he is to the North of me this is OK. There is 100 degrees difference between the 2m beam heading and the 4m heading - meteor scatter is not given to accurate beam headings.

This posting is about the 2019 Perseids meteor shower and a brief round up of other contacts, plus how to select the best MSK144 settings in WSJT-X. However, there is a rant I must get off my chest first

Some amateurs think that using 30 second T/R interval in MSK144 gives them some sort of advantage. It doesn't, it makes things harder for them and everybody else. It is deliberately putting themselves at a disadvantage (which, of course, they have every right to do). Plus it wipes out 50% of my receiving time and makes many of my QSOs fail.

I have been using MSK144 on 15 seconds since it was released in 2016 and it works better than 30 seconds ever could. Today I worked LA4YGA in a total QSO time of 90 seconds - using 30 seconds not we would never have got the signal reports exchanged, never mind the 73s.
Super quick QSO only possible with 15 second T/R timing.
 A complete QSO in a single meteor burst. Oh sure, make life difficult for everybody with 30 seconds, but I prefer to make contacts.

To make sense of this for any non-data mode readers, using 30 seconds (a throwback to previous meteor scatter modes) is like buying an SSB transceiver and running it on AM on the basis that people on either upper or lower sideband could reply. True, but hopelessly inefficient. Running AM would fill up the bands, get in the way of SSB operators, and it wouldn't really work anyway. Yet I know some people do just that, and I remember many did it during the changeover from AM to SSB, and where are they now?

Trying to be King Canute and resist change by telling the tide to go out always ends up the same way. Wet feet and others thinking you are a dork for even trying.

There isn't a "wrong" here of course. Amateurs can do what they want. When they are preventing others from going about their business, that becomes wrong in my book.

In this book of mine, the "wrong" way to use MSK144 is to use 30 second T/R cycles. The "right" way is to use 15 second cycles. Why is 15sec right - because the guidance in the WSJT Home Page and the User Guide tells you so. They wrote the software, they know what they are talking about.

But more on how to get the best out of MSK144 later. I will not mention 30 seconds and the people who use it again in this blog posting. Rant over (for now).

This isn't new, it has just suddenly got worse - http://gm4fvm.blogspot.com/2018/01/the-tyranny-of-default-settings.html

The Perseids meteor shower peaked around 12 August, though as I write this I am still having contacts on its long tailing-off period

2 metres MSK:-
144MHz meteor scatter contacts at GM4FVM, 2 to 15 August 2019
23 QSOs, 10 DXCC entities, 22 squares. Best DX OH6CT in KP24 at 1815km. I have worked Olli previously on 4m, though on 2m this was new square. Special was to work ES3RF as Estonia was a new country on 2m. I had previously worked Gena on 4m and 6m, though once again this was a new square on 2m.

Thanks to GM0WDD for a test and some encouragement, and the same goes for GM4GUF and MI0XZZ who also gave support.

And within two days of my contact with Pasi, OH4LA, I had a QSL card through the post from him (another new square by the way!):-

Needless to say I replied by sending my card direct. The reply was sent by post, in the postbox within 15 seconds of course. Whatever you think of them, QSLs have a certain appeal.

Note too that Pasi is using 800W and a 17el 9m boom antenna with elevation, and I am using 250W and a 7 element 3m boom yagi fixed towards the horizon. I am sponging off the efforts of others here, but the results speak for themselves.

4 metres MSK:-
70MHz meteor scatter contacts at GM4FVM, 2 to 15 August 2019
17 QSOs, 7 DXCC, 14 squares. Best DX was S52OR at 1583km.

SM7CAD was once again a really welcome contact, given that only one frequency, 70.1375, is available in Sweden. Curt worked me split, and the QSO was quick and easy. Well, quick and easy after I finally sorted out how to do split with WSJT-X. WSJT-X makes it easy to do CQ on a calling frequency and move once called on to a working frequency, but that is not true split. As I couldn't figure it out, the last time I worked Curt I had to go and work him simplex on 70.1375. This time I got it right. Curt was very patient and helpful throughout my confusion. It only took a day for muddle-headed me to sort it out, and 4 minutes for the QSO.

At this point I feel I should make a point contrary to an editorial comment in one of the commercial radio magazines. The editor said that the increasing popularity of data modes could be a bad thing as it may result in the authorities reducing band allocations if we show we can fit into smaller spaces. Maybe, but what are the chances of getting spot frequencies instead, like the Swedish 70.1375? After all, commercial FM channels 12.5kHz wide are slotted in all over the place. 

I would like to put in a bid for one on 40MHz and another on 90MHz, please. Both the standard commercial 12.5kHz wide type. That would leave room for three spots 2.5kHz wide, one each for FT8, MSK144 and JT65 for moonbounce. The rest of the space is for CW or other narrow band modes.

Surely we can be fitted in. "Microbands".

I never heard anybody say that CW might be bad for amateur radio because it is so narrow, so why single out data modes?

6m metres MSK:-
Next to nothing heard. I do only have the quarter wave vertical, but there is very little meteor scatter on the cluster this year.

Even some tropo and still some Es:-
VHF FT8 contacts at GM4FVM, 2 to 15 August 2019
This is also covering the same dates so there is a little duplication over what I reported here on my last posting. I kept the dates the same to show the difference between these other modes of propagation and meteor scatter.

There is a mix-up of stuff here, from ON4POO on 2m tropo to CU2AP on 6m Es. However it shows that there is other life while meteor scatter showers continue. There are 53 QSOs in 20 DXCC covering 47 squares. The contact with SV1DA may look like 6m, but once again double hop Es opened up on 4m, SV1DA appeared, was worked, and instantly disappeared. That is the second time that has happened. Still no sign of 5B4 on 4m. Nor 4X4. Bah.


MSK144 Settings
To get the best out of meteor scatter using MSK144 you really need to read the user guide and tweak the settings to suit your station. This will involve adjusting them to the maximum you can taking account of the processing power available in your computer.

I have a more powerful PC than the one I use for radio purposes. However, it makes noise on RF. Therefore I stick with my old second hand PC, which has a processor reclaimed from a former computer on which the power supply blew up, destroying everything except the processor (!). The specification for the computer is (for what it is worth) Intel 4 core i5-2500 @3.30GHz with 8GB of RAM running Windows 10.

I say "for what it is worth" because there is no simple guide as to what is the best computer. If it works, stick with it, I suggest. The best you can afford (from eBay in my case).

Mine is pretty modest by modern standards. The processor is old and as you can see from the top of this post, it can still run three instances of MSK144 on the top settings. At the same time it runs a fourth instance with WSPR, and three display screens, plus plays the cricket commentary from the internet and makes the tea. No, Mrs Doyle makes the tea.

So if mine can do all that, something fairly simple should run MSK144 alone. I say this because, although most computers can run MSK with the best settings, the WSJT-X download is set with defaults somewhere near the capabilities of a Sinclair ZX80. There is probably enough processing power in my watch to run with the default settings which come with the download.

The people behind WSJT have to be cautious, but they do seem to think we use very slow and low power computers. So your modest machine can probably run the best settings too and you will do much better if you increase them above what comes pre-set. The WSJT-X user guide tells you how to do it.

Here are the recommended settings in the WSJT-X user guide, which are higher than the ones pre-set in the download:-

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.

They even show what this would look like:-
Then there is the bit about running the highest values you can:-

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.

In my case, with an old processor and four instances of WSJT-X running at once, all are on deep decode and FTol 200, and the figure in the "Receiving" box is about 30%, and of course it is green.

Next what to do to make quick QSOs if we are running 15second T/R or less (of course we are running 15 seconds T/R, as recommended in the previous paragraphs and the image):-

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.

Despite stories about the end of civilisation as we know it, ticking Auto Seq does not make automatic QSOs. It won't even respond to a station who replies to your CQ. In MSK you must click on the station to reply otherwise nothing happens. All Auto Seq does is finish the QSO you started (if you decide to let it do that).

Then there is some interesting stuff - did you know that the MSK message is contained in just 72 milliseconds of transmission, which is constantly repeated? So much shorter T/R periods are possible, e.g. 7.5 seconds?, though we do not seem to be ready for that yet. You can make it even smaller (20 milliseconds) by ticking Sh.

For operation at 144 MHz or above you may find it helpful to use short-format Sh messages for Tx3, Tx4, and Tx5. These messages are 20 ms long, compared with 72 ms for full-length MSK144 messages. Their information content is a 12-bit hash of the two callsigns, rather than the callsigns themselves, plus a 4-bit numerical report, acknowledgment (RRR), or sign-off (73). Only the intended recipient can decode short-messages. They will be displayed with the callsigns enclosed in <> angle brackets, as in the following model QSO

You can read these hashed messages enclosed in <> if you see long or compound callsigns in use, e.g. GM4FVM/P. You can read them even if you are not one of the stations involved if you tick the SWL box. If WSJT encounters something out of the ordinary it will often appear in the <> format and using SWL may help you to see it.

I have never used Sh, perhaps because I have never got over the idea that everybody should be able to see what I am doing, even if they don't know about Sh. Silly, but that is just me.

What surprises me is that if they reckon 15 seconds with minimum 100 FTol are the right settings, why don't they set those as defaults? If it was me, I would set 15 seconds with no scope for variation (as in FT8) and put the choice of period into the "Advanced" section. But I am a bear of small brain.

Summary
So there you go - key points, T/R = 15 seconds (obviously, it is best), FTol = at least 100, better still 200 (as the other stations are often off frequency), decode = Deep if your PC an take it (small benefit but why not have it), and Sh on 2m if you really feel you need it (but the other station might not understand what is happening - best kept for skeds?). Simples.

Best advice - read the WSJT-X User Guide. Don't believe what I say, and certainly don't believe what you hear on KST Chat and other dark misleading corners of the internet.

Joe Taylor, K1JT, who wrote the software, has a Nobel Prize in Physics to his name. If anyone else professes to be an expert on WSJT ask them if they have their Nobel Prize certificate handy.

As for me, I am still waiting for the summons to receive my Nobel Prize. But I stress again, don't believe me, read the guide.

I may not have a Nobel Prize, but I do have this:-
Add caption ... "No further comment required."

73

Jim

GM4FVM

6 comments:

  1. Do agree 100% with your rant - different periods for same mode is a PITA no matter what band. MSK 15 sec, FSK 30 secs seems to be the norm.

    ReplyDelete
    Replies
    1. Thanks Tom. You are quite right.
      So long as we are all on the same period it should not cause interference, but now we have both running at the same time. With MSK144 people are losing so much running it on 30 seconds.
      Running MSK on 30 seconds is like having a six speed sports car but only using the lowest three gears because your Grandad's 1959 Ford Popular was only a 3-speed.
      I learned to drive in a 1959 3-speed Ford Popular, but I wouldn't want a 3-speed gearbox now.
      Best wishes.
      Jim

      Delete
  2. Hi Jim,

    Another great post, wish you had written this before the Perseids ;-)

    I will put my hands up and say that I used 30 secs periods for a few QSO’s thinking it was more efficient due to the switching. How wrong I was. I shall revert to 15 sec periods.

    As for the settings, well being a typical man I have neglected the instructions. I’ll take a closer look when I have time.

    Last point, how do you put your QSO’s on a map? I would like to do similar.
    Regards,
    Gav

    ReplyDelete
  3. Gav
    I am a great one for stating the rules. However in amateur radio not everybody follows the rules, nor even cares about them.
    It does not matter that 15 seconds is just better, if the majority on Europe on 2m are using 30 seconds then we shall have 30 seconds there (if not everywhere else).
    Rather like Betamax versus VHS video tapes (before your time). The Betamax video tape system was better, but VHS won out in the popularity contest. Actually the Philips V2000 system was better still but did not get a look in.
    If some faceless person has decided to forget what is best and decide to use 30 seconds on 2m in Europe then so be it. Other bands all seem to be 15 seconds, hence my posting. I wrote to the VHF editor of RadCom and got a patronising reply which suggested that in Region 1 things are different. I snapped back that the meteors fall in Region 1 the same as they fall everywhere else.
    So I think it might benefit you to ignore me and stick with 30 seconds until the dust settles.
    The simple solution would be to have two frequencies, one for 15 seconds, and one for 30 seconds. Or one for those who like efficiency and one for those who cannot be bothered.
    Anyway, as three stations have decided to beam at me and blast me out with 30 seconds, I have to stop meteor scatter on 2m. It has been a great 3 years on 15 seconds, but the playground bullies have seen me off. Of course, I am not willing to blast them back the way they blast me. I have never had an MSK contact on 30 seconds and I can see no reason to change to something which works less well (that is a bit pig headed Jim?). 15 seconds is good enough for the other bands.
    I am not the Radio Police and it is not my job to adjudicate on this one.
    The maps comes from a wonderful site provided free by DL4MFM. It is called Log Analyser 3.2. I simply output my log as an ADIF files to store on the PC and then upload them to the site. Then I do a screengrab of the appropriate section.
    It is great what amateurs can do when they co-operate and DL4MFM has done us all a favour.
    http://tools.adventureradio.de/analyzer/
    I will put up something about Log Analyser on the blog shortly. I did it ages ago but my links could do with updating. I will put a link in the sidebar soon.
    73 Jim GM4FVM

    ReplyDelete
    Replies
    1. We had Betamax, sadly I remember it well, going down the video store and being able to choose from about 3 titles while my mates had 15 million (I may have exaggerated the amount of titles but I had it tough, if in better quality).

      Thanks for the link, DL4MFM indeed done a marvellous job in creating his site, a great benefit to VHF folk.

      Delete
  4. This comment has been removed by the author.

    ReplyDelete