pa0r

I had problems when I ran several programs like firefox, gMFSK, gprsdrive, Glade, together. It would just lock up the laptop (256 MB, 1.2 GHz). Somehow I had messed up the swap space when I upgraded from UBUNTU 0.4 to 0.5.

Swapon did not work. It said 'Unvalid parameter /dev/hda3'. cfdisk revealed that /dev/hda3 was available as type 82 (Linux swap partition). 'Free', however, showed no swap partition. And 'top' showed Swap:0.

The problem has been solved; the swap partition was not initialized. After a 'mkswap /dev/hda3' and 'swapon' I now have 1 GB of swap space available again.

After 3 days x 8 hours of testing I have decided to release the next PSKmail beta from the lab. I have not been able to crash the server anymore. I have also been able to add a status messaging system to the client. Now it is time to tweak the timing. This will need a lot of testing, as there are a lot of variables. The varicode makes blocks of different length. The automatic block length corrector makes blocks of 8 - 128 characters. We don't want the server and client to end up sending synchronous polls to each other. Don't laugh, it does happen.... If you make the idle times of both client and server self-adjusting they will get into step with each other, transmitting polls at the same time.

If somebody has a strategy to prevent this, I'd be delighted to implement it. Meanwhile, I will start thinking... For the moment the best strategy seems to be to fix the server times and make the client choose from a set of times that will prevent deadlock. This idea is implemented in version 0.1.4. But the times have not been optimized yet, I have to do that empyrically.

I visited the server at the TU Eindhoven this afternoon. I measured the antenna SWR at 4:1. Which is not too good at the end of a 2-wavelength cable. When I connected the antenna to the transceiver I must have fixed the connector (it is often disconnected when there is a contest over the weekend), as when I came home the signal had improved from S0 to S7. Which if I calculate correctly is a difference of 36 dB.

The system worked well with the S0 signal, and I had throughput even when there was S9+ QRN.

My connectivity problems are solved, and I can throw the MFJ pactor 1 controller away.

Today I have been testing the newest feature of PSKmail - the dynamic block length adjuster - it really is fun to watch the server adapt to the channel conditions whilst they are changing. The length of the payload block now changes from 16 characters to 32 to 64. Even 128 and 256 character blocks are possible, they can be used on a channel with little QRM, but it is much more fun to watch the shorter blocks... For a 'normal' HF channel 16 - 64 is near optimum, as the arq protocol overhead of a 64 character block is only 12.5% and still only 50% on 16 characters (overhead is 8 characters per block).

The PI4TUE test server has an intermittent contact somewhere in the antenna or in the coax, as the signal jumps from S7 to S0 during daytime, and from S3 to S0 with heavy QRM during the evenings. PSKmail does not mind. Even on an S0 signal the block length settles on 64 chars during the daytime, whereas in the evening (S8-9 QRN) the block length goes down to 16 payload characters.

I found another bug today. The NEW MAIL panel does not send quotation marks to the application... I will have to find out if that is a 'GTK2 feature' or a PERL quirk.

I guess I will be able to release the changes somewhere around sunday night.

We are momentarily in the field day location of our local ham radio club, and as there are hardly any activities I could spend the afternoon testing various antennas for use on the camper.
Firsst I made sort of an half sized inverted Vee which is linear loaded and folded back onto itself for 60% of the length.
With this antenna, which remains inside the footprint of the camper I could work my own server at home on 80 meters (26 km) and the PI4TUE server in Einhoven on 30 meters. The 80 meters signal was S7 max. with steady QSB down to S2, The signal on 30 meters was only S0 but workable with a lot less QRM.

Then I used the 20 meters wire for a straight horizontal quarter wave ntenna. This gave less QRM and a steady S8 on 80 meters. Good going for such a simple antenna. he wire was only 2 meters high, so I did not need the fishing rod mast!!

- via PSKmail -

Antenna day. I had bought a 2m/70cm whip for the camper, and today was one of those tropical days where you cannot do anything proper anyway. My son helped me to install the whip on the roof of the camper with an Aluminium bracket. I soldered the connector to the cable, and managed to short circuit it in the process. So I had to go back to the shack to get my reading glasses, get the connector off, and on again. This time it measured open circuit. I can now (just) work the local repeaters on 2mtrs and on 70 cm with the FT897. Neat.

Next was the 80 meter NVIS antenna. Our last trip to the boat has been succesfull as far as the PSKmail tests were  concerned. Of the 10 days we were in the north, we could sail only 2 days... the rest of the time we had to wait in harbor for a new cooling water pump for the diesel engine. Enough time to prepare a 20 meters long, horizontal antenna for the boat. It worked very well. Feed point is on the bow of the boat, where we have a ground connection to the hull (5 tons of steel). From there the antenna wire goes up to the top of the mast (8 meters high, wood) and then down to the stern. It is linear-loaded to resonate on 3595 kHz. With this antenna I could bridge the 162 km distance 24/24 hrs on 80 meters. Cool.

So I wanted this to work on the camper as well. Problem is, the camper is 3 meters shorter (it is only 6.2 meters long), and it has no convenient mast right in the middle. The fishing rod antenna at the rear works fb for dx, but does not do anything but catch wind on the shorter distances.

I took the same 20 meters of wire and folded the wire back to the mast at the rear. This way it looks a bit like a half-size end fed inverted vee. It does load up fine on 80, but we could not test it properly as it became dark. So we have to put off the tests for the day after tomorrow, as tomorrow the camper needs its yearly checkup by the government (some sort of extra tax you have to pay every year..., whilst making sure your garage gets a chance to keep its turnover at a proper level).