Wonder Loop on 40m

Last night I had a bit of time to try the Wonder Loop on 40m. Again, I used 5W of PSK31 so I could use the Propagation Reporter Network to see where my signals had been received.

Once again, the Propagation Reporter map does not give a full indication of how far my signals travelled, as there were only 10 monitoring stations on 40m at the time and some of them may not even have been in Europe.

My first contact was with Billy, 2E0WJC in Leeds, who was also running QRP. It was a solid contact with almost perfect copy. Later on in the evening I worked Gordon, G4TZX near Dover, for another good chat. I was then called by Frank, DL7LAX in Leipzig. There was some QSB but we managed to complete a contact. Finally I worked Kari OH2LRG/4 who had difficulty believing that I was using an antenna inside the shack. Most stations I have worked using the Wonder Loop are amazed that I am using an indoor antenna and want more details, so I end up having a chat instead of just the usual exchange of details.

I am very pleased with how well the Wonder Loop works on this, the lowest band it covers. For Kari and others, I should be starting work on the description of the Wonder Loop very soon now.

Com0com no go

I had an email from Dave Freese, W1HKJ today. He is going to try to resolve the issue with Fldigi and the Eterlogic Virtual Serial Port Emulator and I have volunteered to help test any changes. Dave asked whether I was using the commercial (but free) VSPE or com0com which is an open source alternative. I had come across com0com some time ago, but it wasn't immediately apparent whether it did what I wanted, nor was it immediately clear how to use it.

I thought that while I waited for Dave to come up with something I would try com0com, because it would be nice to have a free open source solution to the problem of many programs wanting to access one radio. The VSPE is freeware (at least, the 32-bit version is) but it might not remain so forever. So I downloaded the com0com setup file and installed it.

Com0com is essentially a null modem emulator. That is, it sets up pairs of virtual serial ports that can be used to connect one program to another on the same computer without requiring two real serial ports and a cable connecting them. Because that's all it does, I had first thought that it was no use for my purpose. But there is a companion program called hub4com, available from the same site, that can be used to connect the ports together in various ways. You can connect a real serial port to two, three or more virtual serial ports, connect programs to the other ends of those serial ports, and by that means all those programs can talk to the device attached to the real port. At least, you could if it worked. But for me, it didn't.

Unlike the VSPE ports, com0com doesn't allow more than one device at a time to connect to one of its virtual ports. So if you want three different programs to be able to talk to your radio at the same time, you need to create three pairs of virtual ports. This is done by running the command line setup program setupc.exe and entering the following three lines (note that each line is split in two by your browser):

install 0 PortName=COM11,EmuBR=yes PortName=-,EmuOverrun=yes,HiddenMode=yes

install 1 PortName=COM12,EmuBR=yes PortName=-,EmuOverrun=yes,HiddenMode=yes

install 2 PortName=COM13,EmuBR=yes PortName=-,EmuOverrun=yes,HiddenMode=yes

This assumes that you want your three ports to be COM11 to COM13. You can change those port names and create more or fewer ports as you wish. You are actually creating virtual serial cables, and the "other" ends of these cables are also serial ports, which are named by default CNCB0, CNCB1, CNCB2 etc. It's useful to use these distinctive names for the other ends so they aren't confused with the ports your applications want to connect to.

An annoying feature of this software is that for each port you create - that's two ports for each virtual cable - the Windows Add New Hardware Wizard starts and has to be told to install drivers. The reason for this is that the com0com drivers are not "signed". This is one of the ways that Microsoft discriminates against free software and amateur developers. Signing drivers requires an expensive Authenticode certificate. This is free software, so there is no revenue to pay for one. So the drivers aren't signed. You can curse Microsoft every time the Wizard starts up. I did.

There is a graphical setup utility called setupg.exe. You can probably use it to create the ports instead of the command lines shown above. I used it to check that I had set the ports up correctly. The result should look something like the screenshot below.

In order to use these virtual cables to connect your logging software, digimode programs and so on to your radio, you must connect them together using the hub4com program. As I stated earlier, you can download this from the com0com site. It comes in a separate zip file and should be extracted into the same folder as the com0com program. You can then open a command prompt, go to the com0com program folder and enter a command like this (again it is one line):

hub4com.exe --route=All:0 --route=0:All --baud=38400 --octs=off \\.\COM1 \\.\CNCB0 \\.\CNCB1 \\.\CNCB2

This is supposed to connect the ports so that all data from CNCB0 to CNCB2 (which are the other ends of COM11 to COM13) goes to COM1 and your radio, and all data from the radio sent to COM1 goes to CNCB0, CNCB1 and CNCB2, and from there to COM11-13 and your programs. It also sets the baud rate to 38400, which is what I am using with my K3, and turns CTS off. It's probably advisable to put this command into a batch file, so you can easily re-run it (or make changes to it) without retyping. Obviously if your radio is connected to something other than COM1 you should change that part of the command.

Unfortunately, it doesn't work properly. At least, I couldn't get it to work properly. I did get KComm and Fldigi to connect (separately) to my K3 on COM1 via virtual port COM11. But a lot of the time nothing happened at all. And I never got two programs connected simultaneously.

I ended up using RealTerm (another SourceForge project) for testing, since it is much easier to see what is happening (or, most of the time, what isn't happening) than trying to use ham radio programs that just do nothing if they don't get what they wanted. I found that I could type commands - such as mode commands (MDx;) - while connected to any of the serial ports, and see the K3 respond to them. But output from the K3 was only sent to the first virtual port that connected to it, whichever that was. I also found that if I connected to a port, sent commands and received responses, disconnected, and then connected again, sometimes there was no communication on the second connection.

I spent a lot of time trying to find out what I was doing wrong, but it appeared to me that everything was set up correctly and the software was not performing according to the documentation. So I have given up on it. I am describing what I did here in the hope that someone cleverer or more persistent than me will find out what the problem is, get it working properly, and tell me.

It would be good to have a free, open source way of enabling two or more programs to talk to a radio simultaneously. But I just don't have the patience to spend any more time on it.

Fldigi and VSPE conflict

This is another entry for the "I hate computers" section. I had noticed that Fldigi for Windows did not appear to be remembering the "use software PTT" setting, so the first digimode transmission of a session always went no further than the sound card and I had to open the configuration settings and set it again. Thinking that this was probably a bug that had been fixed by now, I went to the Fldigi website and downloaded the latest version, which was 3.12.3 at the time or writing. After I installed it, the program froze on start up and never even appeared on-screen. I had to use Ctrl+Alt+Del to terminate it.

Some people probably think that because I can write programs this kind of thing is meat and drink to me. In fact I find it just as annoying and frustrating as anyone else. Although my experience means I have more ideas about what to try, I certainly don't immediately know the answer. And the onset of senility means that I often forget something obvious which means I waste a lot of time searching in the wrong place for the cause of the problem. I wish computers would "just work" like any other appliance!

So after half an hour that I would have preferred to have spent testing my loop antenna I eventually discovered that the problem was caused by the fact that I am using the Eterlogic Virtual Serial Port Emulator. If you have not come across it before, this is a bit of free software that allows more than one program to connect to one serial port at the same time. It has become an invaluable component of the computer side of my radio station as it allows my logging program, KComm, to be connected to my K3 at the same time as my remote tuning knob program, KTune, Fldigi or any other digital modes application.

VSPE is transparent to every application I have used it with so far. Even the Elecraft firmware update utility works through it, though obviously not at the same time as any other program. But something seems to have changed in Fldigi 3.12 so that it freezes when trying to access the virtual serial port, though it works perfectly when given direct access to the real one. Leigh, WA5ZNU has reported this to the developers so hopefully there will be a solution. Meanwhile I'm sticking with Fldigi version 3.11.

Some 30m test results

Last night, between the end of one BBC Promenade Concert broadcast and the start of another, I managed to spend half an hour on 30m PSK31 with 5 watts and the Wonder Loop. The map below shows the results from Propagation Reporter just before I closed down.

Not as impressive looking a map as the one on 20m this afternoon, but it was only for half an hour, and 30m is not as busy a band so there are fewer monitoring stations. However, conditions seemed better, received signals were stronger and I received replies after fewer CQs. I made two contacts into France and was also called by a station in the Czech Republic who disappeared after I sent my first over so it didn't count for a QSO.

Something that went unremarked in this blog at the time is that on 25 July a 2 watt WSPR signal into the Wonder Loop on 30m was received by Myles, VK6ZRY in Western Australia (who, incidentally, was also using a magnetic loop antenna!)

I think these results show the Wonder Loop radiates a good enough signal on 30m to make plenty of CW or digital QSOs. Next I will try 40m.

PSK31 contacts

Today I thought I'd try to make some PSK31 contacts using the QRP Wonder Loop. Because the mouse pointers go berserk on both of my laptops when in the presence of a strong RF field, I had to use the main shack computer and therefore the K3. But I turned down the power to 5W p.e.p, which is what I would run if I was using the FT-817 or the K2, although the tuning capacitor I used could easily handle 10W, and perhaps even more.

The picture shows the map displayed by Propagation Reporter after I had been on the air for three hours or so on 20 metres during the late morning and early afternoon. For those unfamiliar with the display, the small markers show the stations that had been received on the Wonder Loop and decoded by Fldigi. The large markers with a time interval show the location of stations that heard my transmissions, and how long ago they last reported me. These stations would have needed to receive "DE G4ILO G4ILO" without errors in order for their software to submit a report, so the large markers represent stations with whom it would have been possible to have a QSO. The concentration of large markers in Western Europe may in part be due to the fact that fewer people in Eastern Europe have a broadband internet connection, without which propagation reporting is not very practical.

It has been more than a year since I last operated QRP PSK31. I got the 100W PA for my K3 because I was starting to find QRP a bit frustrating after the many years (it seems!) of poor conditions. So I had forgotten that PSK31 with 5W can be a bit hard going.

I called CQ most of the time because that gives the best chance of being spotted on the Propagation Reporter network. Although a couple of dozen stations heard me (as you can see from the map) only two replied to my CQs. I made one other contact, with DB7HH, which was in reply to his CQ. Normally when working QRP I would search out other stations and call them, which would probably have resulted in more contacts.

I think the Wonder Loop performed well in this test. I think the difficulty in making contacts was mainly due to the poor conditions. There were no really strong signals during the period of the test. Although the results might not show it, I felt that the loop worked well enough that it would not be too bad if it was the only antenna I had. In fact, it's working better than I expected.