Match & Diplex

This is a bump update from some research for the Radio Project.

I’ve gotten some more clarity on the desired end game for the antenna array. Here’s what I’m looking at:

  • I’m working on an optimized discone for reception in FM broadcast (88-108 MHz), air operations (118-137 MHz), all of VHF and UHF. By, “all,” I mean everything that’s worthwhile listening to and isn’t going to cause any issues. I’m not supposed to — and I have no interest in — snooping in any bands where I have no business. This gives me a wide receive bandwidth of (in round numbers) 80 MHz to 450 MHz. Discone antennas are supposed to have a 10:1 bandwidth so I figure I can squeeze out 80 MHz to 800 MHz and not use all of it..
  • The same discone antenna will be my VHF transmit antenna for 2-meter amateur bands. Discone antennas are advertised (colloquially, on the Internet) as having an, “about 5:1,” transmit bandwidth so with a low end in the 80 MHz range, I should be able to do 2-meter relatively easily. If I design it right, tuning it to 88 MHz instead of 80 MHz as the lowest frequency, I may be able to get it to do a decent job of transmitting on 70-cm amateur bands. Either way, using the discone as a transmitter is mostly a stop-gap bridge to get me on the air and DXing enough to figure out what more I’d want to build.
  • For HF bands, I’m going to build something bigger, along the lines of a dipole. I’m vague about that because I’m not building a 160-meter dipole. There’s got to be something that’ll get me good performance given sufficient ground planes and what not in all the HF bands I want to work in. This will be a transmit and receive antenna and it’ll probably go through several iterations.
  • Once I figure out my VHF and UHF transmit needs in more detail, I’m going to put appropriate dipoles together with the HF antenna. Other than matching, they should be able to coexist if I get the physical configuration right for phasing and what not.
  • Down the road, when I’m comfortable that I have a working and stable platform, I’m going to build some lower-frequency experimental stuff. I’m leaving an empty, “slot,” for a third antenna system that is yet to be determined. That slot may be used for other experimental things as well but having the infrastructure to pop up another antenna — feed lines, a cleared pad, etc. — will be nice. By then, I’ll have experience, I’ll probably have better tools to figure out the electrical characteristics of my experimental antennas and I may very well work towards an upgrade from the FCC.

All of this together means I’m going to have three feed lines into the shack with several wide-band antennas. Those lines will feed several radios:

  • My main ham rig. I’ve got a bunch of equipment but some of it is yet to be determined.
  • The scanner I have right now. I’m not going to build anything for this because I don’t want to mess with any regulatory stuff outside bands I’m allowed to use. I’m pretty sure a scanner out of the box won’t mess with anyone else’s reception of whatever they’re receiving.
  • Hopefully there will be an upgrade to a desktop scanner at some point. Same deal as the current scanner, just a bigger and badder version of the same thing.
  • An FM broadcast radio. The regular deal, listen to music and what not. No reason to tinker with any of that since I can probably get everything I ever wanted and more from Goodwill.
  • Whatever experimental stuff I’m going to cook up. I think there may be an SDR in my future but that’s probably a long way off. It’s nice to have the headroom.

This means that I will have five radios sharing three feed lines. I need to split lots of things. I need to diplex — maybe more — to some of the feed lines as well.

This article (PDF, 718 KB) shows how to build some simple DIY splitters. For some background, this article (PDF, 348 KB) also gives great information. I haven’t processed much of it yet but it looks like splitters are similar to impedance matchers in that they’re mostly LC circuits. In the simplest sense, they’re LC filters, which means I can tune them to the correct impedance with careful work.

Over time, I’d like to design some more isolation and possibly active splitting as well. No definite plans or even feasibility thoughts in the area yet.

This does however give me hope that I can get everything I want, engineered the way I want it, and that I can probably build most or all of it myself. I’m going to have to learn how to wind coils and transformers but I’m not sure how much radio work I can do myself without that anyway. That’ll also be really useful to know for PWM-based work, like power supplies, class D amplifiers, SDRs, etc. so learning that is going to be an important next step.

Matching Antenna & Feed Line Impedance

As part of the Radio Project, I’m attempting to build my own antennas and feeders out of either transcycled or inexpensive parts. I’ve already got some parts on order from McMaster as fill-ins for missing bits for the discone antenna. I’m not in the position to make an investment in high-price feeder right now. Quad-shielded RG6 is easily available in bulk at (relatively) low prices.

RG6 is 75Ω. Just about everything out of the I-don’t-know-what-RF-is market (TVs are in that market; radio gear is out of it) runs in 50Ω. I’ve got some decent 50Ω RF patch cables from the bulkhead (through the exterior wall) but getting from the antenna to the building is going to require some length of good cabling.

The first application for the discone antenna is the scanner. It’s receive-only and I bet it’s pretty well built. More than likely, it would work okay with a 75Ω feeder. At some point though, I’d like to use the same antenna to transmit on both the 70cm (UHF) and 2-meter (VHF) bands. I have no idea what the eventual impedance of my antenna is going to be either.

I ran across an article on homebrew impedance matching transformers (PDF, 222KB) on W8JI’s website. It’s a little above my head right now but I think I’m starting to get some ideas.

For starters, I think I need impedance matching circuitry. I’m pretty sure I need two circuits, one between the antenna and the feed line and the other between the feed line and my gear. The feed line is pretty short and I’m not expecting a long run from the bulkhead to my gear so those are probably not major factors. I don’t have my antenna parts yet so I’m only considering the feed line impedance right now.

I looked at some connectors and they seem to be rated for 50Ω in all the sizes I want: Type N for most of my radio gear and BNC for the scanner. I’m probably going to have to connect the feed line through a matcher outside for a 50Ω bulkhead and whatever other stuff I need to do on the shack end.

There are really easy thing you can buy to do this all for you. Mini Circuits seems to have exactly the right thing except it’s about $3 per piece with a 20-piece minimum order. And it only does 1W. And it’s surface mount.

I’m probably going to have to build my own balun (matching transformer). I may have to build two of them. It looks like I can get fairly good results from the right number of turns of the right kind of wire on the right core. I don’t have any practical experience in selecting any of those correctly (yet).

The turns ratio is the first thing. The ratio of the number of turns is the square root of the ratio of impedances. In my case, that’s the square root of two over three, which is some nice irrational number (thanks, both two and three, for that). I think I can probably get away with something close, like 5:4.

I’m going to guess that the wire is probably less important. From what I remember, there are three things to pay the most attention to in a transformer winding: inductance from the winding, DC resistance and parasitic capacitance between turns. The inductance is easy. It’s the thing I want. More on the dielectric in a minute. The DC resistance can probably be solved by using thicker wire like #18 or #20 instead of #24 or #26 that people usually use for transformers. Thicker wire usually comes with thicker insulation and from what I remember capacitance is inversely proportional to the distance between two conductors. Capacitance is also dependent on the size of the conductors but that’s increasing way slower than the separation due to the insulation so I think I can safely ignore that for now.

W8JI talks about heat being the largest factor for selection of a core, both in terms of how it affects saturation and in terms of the longevity of the device. I think that using larger wire I’ll need a larger core which should dissipate heat better. I have no idea at this point on how to select the right core. I’m pretty sure I want something torroidal because I think that’ll produce less stray RF (interference) and I’ve got a lot of other wireless gear that may not work as well with lots of interference. Plus I think that it’s just generally bad anyway and it’s my job to reduce it.

That leaves me with core selection, which includes getting the right dielectric. More research on this is to come.

Incidentally, I’ve realized that I’m not going to get through this without having an antenna analyzer. I looked at some of that stuff a few days ago and the ones off the shelf are expensive. It’s critical test gear which my gut says I should probably buy, but one that covers HF through UHF looks like it’s around $1,000. That doesn’t even cover lower frequencies like MF and LF (or even ULF) that I’d like to one day tinker in a little. It’s also way out of my price range.

I’m going to need a dummy load too. I did a little research into that too. That seems like it’s really easy and rewarding to build.

Thanks also to Bruce Conti for this article (PDF, 913KB) that explains some things about how to figure out values for baluns as well as how to test them. The notion to go to Mini Circuits to even look at that option came from there too.