After noting a potential interaction in my introduction article for the TH1vn on 15 meters I slightly repositioned the overhead run of coax between the tower and house. The problem is now gone. I doubt that what I did is responsible for this good outcome so I am back to dealing with a likely intermittent connection. However I don't plan to take immediate action -- it can wait a few days.
With everything stable I decided to plot the SWR of the antenna. The numbers come from the bridge in the KX3. The measured values are close to those measured by my Daiwa cross-needle meter, a meter that has served me well in the past. It's just easier to read a digital display. I have no idea which meter is better, or perhaps they're both wrong.
I used the line loss calculator provided by VK1OD to convert the SWR measured at the transmitter end of the coax to the estimated SWR at the antenna feed point. The transmission line is ~35 meters of RG-213/U. This is nearly twice as long as I need but I am not willing to start chopping up coax for a temporary antenna. Transmission line runs at my previous station had to reach the top of a 19 meter high tower that was some distance from the house.
As you can see the SWR does not consistently favour any one mode. Since the SWR is below 2:1 at all frequencies I plan to operate I am happy with this outcome.
I did not plot the estimated transmission line losses. When the SWR is high the loss can grow quite large. On the SWR charts the high losses correspond to where the SWR is very different between the transmitter and antenna ends of the line.
- 20 meters - 0.9 db across the band
- 15 meters - 1.1 db up to mid-band then gradually rising to 1.4 db
- 10 meters - 1.3 to 1.5 db below 29 MHz then rising to 2.2 db
This is a tri-band antenna and there are other bands. I makes sense to use an antenna tuner to match the antenna on other bands. The only ones of particular interest to me right now are 17 and 30 meters. It ought to work well on 17. On 30 it is too short to work well, even though it can be readily matched.
First, though, it is instructive to determine the mismatched line loss. The transmission line is long and the SWR is high. Here are some numbers to ponder, after which I'll say something about measured performance.
- 17 meters - SWR is 5.7, which implies (by use of the VK1OD calculator) a feed point SWR of 16 and a line loss of 4.2 db.
- 30 meters - SWR is 7.4, which implies a feed point SWR of 19 and a line loss of 3.1 db.
I made one European QSO on 30, just as a test. He had difficulty copying me. It subjectively seems to hear and transmit worse than my eavestrough antenna. I did not bother with an A-versus-B on-air comparison.
On 17, where I expected it to do well, it receives very poorly when matched, much worse than the eavestrough antenna. I tried calling a couple of stations, but without success. That can't just be due to the 4.2 db line loss. (In comparison, the eavestrough antenna has zero line losses since there is no transmission line.) I tried two different tuners, getting identical results. I then turned to EZNEC to see if I could find some indication of why it's so poor.
There is no easy way to model the antenna due to the traps for 10 and 15 meters. While inactive on 20 they still add reactance and loss. Modelling would require some effort, including details of the trap design. Neither is worth the trouble. Instead I modelled a 20 meters dipole built with aluminum tubing and placed 10.5 meters above average ground. I haven't bothered to include far-field radiation patterns in this article since they are just those of dipoles on the respective bands and therefore of no particular interest.
At 18.1 MHz the model dipole shows an SWR of 19.3. That this value is close to the calculated SWR of 16 (see above) may be nothing more than coincidence. However my goal was not to see if the SWR agrees but to get a rough idea of the radiation resistance. Regardless of the reactance value (which is of high) a low resistance (the 'R' in Z = R + jX) might indicate additional losses. Instead the R value calculates to 157Ω.
It is still quite possible that the traps are lossy at 18 MHz since this band did not exist when the antenna was designed. The tuner might also have unexpectedly high losses, though this is unlikely since I got the same results with two different tuners, one of which has high-power components.
I may just choose to shrug my shoulders and skip operating on 17 meters for the summer. I can be pretty happy restricted to good operating on 30, 20, 15 and 10 until the early fall. The SWR is 2.2 at 50.1 MHz so I may yet work some 6 meters before the end of this Sporadic-E season.