At SARCNET we always have lots of projects on the boil. These projects, all inspired by our school Amateur Radio clubs, have been started, but have not yet been completed and documented on their own separate web page. We sometimes need to stop and re-evaluate our approach. We are always interested in what our readers think, too. So, any suggestions or ideas for new projects are always greatefully accepted.
Heavy-duty 300 Ohm open-wire feeder - Mk1
We needed to make our own 300 Ohm, heavy-duty, open-wire feed line for our folded-dipole and full-wavelength resonant loop antennas, which have a feed impedance of around 288 Ohms. We had been using 300 Ohm TV ribbon, but it is very light-weight, breaks, doesn't handle much power and its SWR is adversely affected when rain collects on the webbing between the conductors. So we decided to design our own open-wire feeder. An on-line impedance caculator indicated we could make our own 300 feeder if we maintained a spacing of 9.8mm with a conductor diameter of 1.5mm. We used Electra Cables SR1015 1.5mm^2, house-hold electrical cable.
Conductor: Plain annealed copper
Insulation: PVC 90
Voltage: 0.6/1 kV
Format: 7 x 0.50mm
Insulation Thickness: 0.8
Outer diameter: 3.1
Standard: AS/NZS 1125, AS/NZS 5000.1, AS/NZS3808
Standard Packing Length: 100m, 200m, 250m, 500m
Operating Temperature: 90°C
Our first attempt to make our own feeder was to use cable ties installed in an inverse manner, so that the cable-tie blocks work as a natural spacer. It is easy to make and works well provided you select the right cable ties for the job.
300 Ohm Open Wire Feeder Mk1
Heavy-duty 300 Ohm open-wire feeder - Mk2
Our second attempt to make our own feed line was to use adhesive (dual-wall) heat-shrink tubing. Although harder to manufacture, using a hot-air gun and small, flat pliers, it looked more professional, would be easier to pull through holes and might have the advantage of surviving prolonged UV exposure - something black cable-ties are not noted for, in our experience.
300 Ohm Open Wire Feeder Mk1
4:1 Guanella Balun
We built this 4:1 (200Ω:50Ω) parallel transmission line current balun - a classic Guanella balun. Noting the expert balun advice from Owen Duffy, VK1OD, we chose to use two toroidal ferrite cores, Amidon FT-140-43, each with six bifilar windings, to made a dual-core, current-balun. We tested it with a NanoVNA and it was essentially flat from 1-60MHz. It also passed all five VK1OD Balun Tests. Instead of enamelled copper wire we used figure-8 insulated copper cable. It was easy to wind, required no core insulation and the red and black colour of the cable made it easy too match up the ends. On the unbalanced 50Ω side, the red wires are both soldered to the centre pin of a SO-239 socket, while the black wires go to the shield. On the balanced 200Ω side, the red wire from one toroid is soldered to the black wire of the other. The remaining red and black wires are output. This type of balun is used to feed a folded dipole or an Offset Centre Feed (OCF) dipole.
4:1 Parallel Transmission Line Dual-Core Current Balun
6:1 Balun for 300 Ohm, open wire feeder
Next we needed a 6:1 balun to feed the 300 Ohm balanced feeder from our radios with 50 Ohm unbalanced output. We started with the usual 4:1 (Guanella) balun. Very pleased with the above results: Just for fun we added two extra turns to the balanced side and retested it: It remained flat from 1-60MHz, but matched an output impedance of 300 Ohms - Perfect for our needs.
6:1 Parallel Transmission Line Dual-Core Current Balun
While developing baluns, we realised that we needed a test setup to consistently measure their frequency response. For a while we were using one of those handy BNC-To-Banana-Terminal adapters at the end of a coaxial cable. But we found, when putting two of these adapter cables back-to-back, the SWR was quite high: Sometimes as much as 1.4:1 @ 60MHz.
It was now time to put our new NanoVNA to good use and to find out what was going on. It is first necessary to calibrate the NanoVNA and to completely verify the test setup.
- We calibrated the NanoVNA using the open, short and 50 Ohm loads provided.
- We checked the frequency response of an SMA-SMA, RG-58CU, coaxial cable connected to the 50 Ohm load. It was flat as expected: SWR 1.02:1 (-40dB ref coef) from 1 to 60MHz, as shown in brown below.
- We cut the SMA-SMA cable in half and soldered two short wires to the open ends to make a pair of coax-to-wire test cables. The short wire ends on these test cables would be used to connect to the balun.
- However, we found that the short wires were inductive and they had an effect on the measurements.
- To check the test cables themselves, we connected the wire ends together. One SMA connector was attached to the primary port of the NanoVNA and the other was connected to the 50 Ohm load. The result was an SWR of 1.02:1 (-40dB ref coef) @ 1MHz and 1.14:1 (-24dB ref coef) @ 60MHz as shown in blue below. The latter being somewhat dissapointing, but unavoidable under the circumstances.
NanoVNA-F Vector Network Analyser
Coaxial cable to short wire test cables
Frequency response of the test cable (blue) vs coaxial cable (brown)
Experimental Balun #1
While playing around with parallel transmission line baluns, we wanted to try winding them with some enamelled copper wire. This has the advantage that you can fit more turns on a toroidal core than you can with twin flex. We started experimenting with a single-core 1:1 Balun. This one uses the classic Ruthroff "polarity reversing transformer" with a "crossover" to make the input and output appear on different sides of the toroid.
However, when using enamelled copper wire, we have to protect the insulation from being scraped off by the sharp ferrite material. Traditionally, Teflon tape is used. We tried it, but we found it too flimsy. When winding tightly around the core, the wire punctured through the Teflon tape to the ferrite. Instead, we then tried some Kapton tape. This worked a treat, but it was a little messy to apply as can be seen in the photo. There is a good article on using (or not using) enamelled copper wire in transmitting baluns by Owen Duffy, which recommends using Teflon sleeving. Unfortunately we didn't have any.
We used some 1.25mm diameter wire on an Amidon FT-140-43 ferrite core. We cut 1080mm of wire and bent it in half to make one piece, 540mm long (a trick to make sure that each winding is the same length). We estimated that would provide 25mm of excess lead for connection.
Also, to ensure the parallel transmission line maintained a consistent, minimum, separation we used small pieces of heat-shrink tubing (Balun-guru Jerry Sevick W2FMI used this technique, but with fibreglass tape). We called this balun #1. The SWR was 1.24:1 @ 60MHz. Even correcting in some way for the test cable (maybe 0.07:1 less), this was not really as good as we had hoped.
Experimental Balun #2
Examining the finished product, we had an idea: Why not just use heat-shrink tubing to keep the wire together and insulate it from the ferrite? It was much easier to wind than balun #1 so we called this balun #2. Unfortunately, it did not work as well as balun #1 as the permittivity of the heat-shrink tubing must have reduced the impedance of the transmission line too much. The SWR of balun #2 rose to 2.09 at 60MHz!
Experimental Baluns: #1 above and #2 below
Balun #1 frequency response
Balan #2 frequency response
We used an online open wire feeder impedance calculator: The characteristic impedance of the line was 166 Ohm without the tubing and 83 Ohm with the tubing (assuming a permittivity of 4). It seems the best characteristic impedance was somewhere around 125 Ohms. Also the calculator says the tubing decreases the characteristic impedance. Owen Duffy says: "Sleeving both conductors may be useful, and if the sleeve is of sufficient thickness, the enamel is no longer important. Unfortunately, thick sleeving increases Zo, and so has a usually undesired side effect." After reading that we are not sure which is right.
We wondered if the same was true for the small pieces of heat-shrink tubing we used in balun #1. We started to remove them and the SWR started reducing. So, we removed them all and the SWR was reduced from 1.24:1 to 1.03:1 @ 60MHz with a peak SWR of 1.09:1 @ 30MHz. That was great! We called this balun 1A. But we noticed that the SWR value could easily be affected by slightly moving the wires, changing their spacing.
We came up with the following ideas to improve the construction of balun #1A:
- Add a drop of super glue to each winding to help stabilise them - A bit messy, but it works!;
- Dip the ferrite core in some insulating coating (?) before winding;
- Cover the ferrite core with 36mm heat-shrink before winding.
Balun #1A frequency response
We have to say that our experiments, so far, into making baluns using enamelled copper wire were not as successful as the using figure-8 insulated copper wire, which was much easier to work with and produced better results.