Recently I have been working on a project to phase two untuned copper loop antennas. These would be set in line, in phase. The figure-8 patterns would be aligned to null a specific station. Then the phaser would be used to either enhance the null, or to null a second station. It hasn't worked.
The main problem has been leakage from the cables used to connect the antennas to the phaser. I did some testing. The receiver system used is that in my truck, outlined here.. A well shielded dummy load was connected to the external input on the system. Tuning through the AM band showed no signals at all, not even a het while in sideband mode. Nice!
A 150' length of Radio Shack RG-58 coax was run out straight in the yard. The dummy load was put on the far end, and the near end hooked to the external input. Signals through the whole band, with a maximum of S-9+20 on a 5kw station about four miles away. Not nice! Various grounds were tried, and a 50 ohm to 50 ohm unbalanced to unbalanced toriod transformer was used. No appreciable change.
Substituted some new Belden RG6, same setup. Same results. Still not nice!
A balun was made, 50 ohm unbalanced to 110 ohm balanced. It was put in a Hammond 1590G cast aluminum box with a BNC for the coax and an RJ45 for the balanced side. The cable was new CAT5 ethernet cable. A termination was made for the far end using another RJ45 jack and a 100 ohm resistor. The scan was done, and signals were down significantly. Locals were down at least 10db, and non-locals like WFAN and WCBS in New York (150 miles away) were inaudible. Progress..
A loop antenna was connected to the far end by winding a half dozen turns around the toroid pickup and connecting to an RJ45 jack. Significant directional ability was noticed, unlike with the coax.
Further work on this is going to happen. I intend to try a center-tapped winding on the balun, and ground that. I also will make a similar balun for the far end in case there is some unnoticed issue out there. I also want to try running some chicken wire over the cable. Perhaps the combination of a shielding effect and the capacitance to ground will help.
A different type of coax that totally prevented signal ingress would be best. However, this is intended for portable and temporary use, so any sort of hardline is out of the question. That use also makes driven grounds impractical.
There have been two suggestions on this. First is to use a quad-shielded variant of RG6, with the appropriate connectors properly installed. Second is to use shielded CAT5 cable. For that to work, I would need to find some sort of connector that would have a 360 degree shield integrity around the RJ45 jack. Just grounding to the outside of the metal box probably wouldn't be as effective. Shielded CAT5 is also less common and more expensive. I think they make special RJ45 jacks/plugs with a grounding tab.
Another real possibility is the use of twinax cable such as was used on older IBM networking, for example. I don't know the percent of shielding over that cable, and that whole scenario would likely be rather expensive. Another balanced possibility is shielded audio cable like Belden 8451. That has an overlapped foil shield and a separate ground lead. Should be around the same 100 ohm impedence value.
Tried using two runs of coax in parallel as balanced line. Didn't work. Made two boxes with two BNCs in, one out. Has a toroid trifilar wound with a winding to each jack, with one balanced side inverted. Effectively a 180 degree 3db splitter/combiner design. It wasn't any improvement on a single run of coax. I will look at this again and test the individual components. I'd have expected some improvement.
If anyone has thoughts on this, or has already done similar experiments, I'd like to know about it..
|Last update: 27 August 2007|