In the discussion during the week, the common wisdom seemed to be that it was better to switch at the transformer secondary, ie the low impedance side. Well, I thought it rather surprising that the isolaton across a relay would be bad enough to cause problems on 160m, so I decided to do a test of my own.
I wound a couple of 9:1 impedance transformers on the FairRite binocular cores that I use for all my Bevs, and actually measured the isolation (in a 450 ohm system) across a sample of the relays that I use at my remote site. These relays are nothing special, except that they have gold flashed contacts to avoid the need for wetting current. Using my N2PK VNA, I found the isolation to be 53 dB. Now, in my opinion, that is more than enough, as it is not going to influence the RDF of a Beverage in the slightest. Even for the guys relying on a 30 dB null off the back, to get rid of some particular source of interference, it isn't a problem. As I said in a previous post, the advantage of switching at the high Z side, is that you only need a single ground point at the feed end. Now there could still be problems of interaction between the Bev wires themselves if they have to approach the relay box, but I do not believe that this would be a problem for two wires arriving from opposite directions, and may even be OK for 4 wires arriving at 90 deg to each other. Since I had the VNA out and running, I decided to measure the effect of using a common ground while switching at the transformer secondaries. Note that this is DEFINITELY not recommended practice, and the feed end transformers should be grounded using separate grounds spaced well apart. From experience, a typical single ground rod may exhibit quite a variable impedance to ground depending on soil type and humidity. I decided to assume a value of 100 ohms as a rough average. This particular test required me to wind a third transformer to simulate the feed from the idle antenna. With the secondary of the idle antenna open, the isolation (crosstalk due to common impedance coupling) was only 27 dB, compared with the 53 dB across the relay. If I shorted the secondary of the idle transformer (also "recommended practice" in other applications), the isolation actually dropped to 19 dB. This is because the choking effect of the transformer magnetising reactance had been removed. Now, figures like this will DEFINITELY affect the performance of the antenna. My conclusions are as follows:- 1) For just two antennas arriving from opposite directions, one can safely switch at the high Z point, and enjoy the advantage of only having to insert a single ground rod. 2) This MAY be true for 4 wires arriving at 90 deg to each other. 3) More wires would need investigation 3) If you MUST switch at the low Z point, you MUST use a separate ground for each antenna. I am not sure if the contesting.com site has a space for file uploads. If so I would be more than happy to post pdf's of my VNA traces. Can someone tell me? 73, Greg, ZL3IX _______________________________________________ UR RST IS ... ... ..9 QSB QSB - hw? BK
