Group events like ARRL Field Day present a problem: how do you have a bunch of rigs transmitting in a small space without interfering with each other? The answer might be to use a multiplexer.

For RF, a multiplexer (diplexer, triplexer, quadplexer, etc.) is a device that divides up a signal by frequency into separate bands and routes each signal band to different devices. It can be used to allow multiple rigs to work with a single antenna, or a single rig to work with multiple antennas. For this project, the focus is to allow multiple rigs to transmit or receive on different bands at the same time, all into the same antenna – including multiple rigs transmitting while other rigs are receiving, all simultaneously.

This sounds like highly-welcome magic: much less QRM for each station, and we only need to put up a single antenna? Where do I sign? :) But as desirable as this would seem to be, there isn't a lot of information on how exactly to do this.

The purpose of this document is to outline my learning process as I attempt to assemble such a system for myself for use during group events such as Summer/Winter Field day. I find it very useful to record my thoughts as I go along: you can only be a beginner once, and once you know what you're doing it's very easy to forget those subtle details that might have been really confusing the first time you found them. Hopefully it might be useful to others as well.

Here are my current outstanding questions:

• Did NF4RC actually start by ordering the VA6AM kit, or did they roll their own from the beginning?
  • I'd like to know before I order the kit: is that a good first step?
• Does the $140 Canadian triplexer kit include everything you need beside enclosure, magnet wire, and UHF connectors? (Of course, tools, solder, test equipment, etc. not included.)
• Can I actually successfully tune this thing with a basic VOC multimeter, NanoVNA and TinySA?
• Does the custom NF4RC bandpass PCB differ from the one included in the VA6AM kit? If so, how and why?
  • For now, I think I'm sticking with the kits, but it would be nice to know as this develops beyond the kits.

It seems like this should be a pretty straightforward thing to do. After all, there is equipment that is sold for just this purpose. So just Google for one, buy it and wire it up, right? That's exactly what I thought…

Turns out, no. First of all, most of the multiplexers I've seen are not for amateur radio. And those that are don't usually seem to be for HF: there's a lot more for dividing up various VHF/UHF/etc. frequencies than there are to divide up among HF bands. And when you do find a multiplexer for amateur Hf frequencies, you then have to find one that handles the actual specific bands you want it to handle. And then one that handles the power level you want… And then you find that the perfect device for you is sold by a Russian company, which in the US in 2024 is unlikely to be able to handle your order…

Eventually, you might find one that meets at least some of your needs. Here is a three-band example for a start. So if this works for you, now we can just buy it and wire it up, right?

Still no. It seems that more elements are needed. While the multiplexer by itself does what it says it will – divide up a signal by frequency – the amount of division it will do is limited. Any analog filter, including the filters in a multiplexer, isn't going to eliminate 100% of a signal. And increasing the amount of attenuation that a filter provides can create other side-effects, so in general you only filter as much as you need. And in applications where users are not transmitting and receiving at the same time, you don't need that much filtering. For such applications, ~30dB of attenuation is usually sufficient. And that's what the average multiplexer tends to target. But when you are transmitting and receiving at the same time, you need more like 60dB. So how do we get the extra attenuation?

We add that additional attenuation by adding bandpass filters for each tap on the multiplexer. That is a device designed to filter everything but a narrow range of frequencies. Which is exactly what the multiplexer does, but it does it for only a single range, and attenuates all other frequencies. If you have a bandpass filter that gives you 30dB of attenuation outside of your desired frequency, and a multiplexer that give you 30dB of attenuation, you get your full 60dB of attenuation!

I hadn’t realized the need for the bandpass filters. I thought the simple multiplexer did everything you need. Of course, when I first started this process, I only considered the cost of the multiplexer: a few hundred dollars. For 4 people, that's a pretty good deal! But once you realize that you need a bandpass filter for each tap, that really blows the budget: each bandpass filter costs almost as much as the multiplexer!

Now, my understanding is that if you have a big enough budget, this is a solved problem. If you want to operate on 6 bands (160/80/40/20/15/10), you absolutely can. You just need:

• Split Multiplexer to divide into two frequency groups: Above 40m and Below 40m
• 20/15/10 Triplexer to divide into individual 20/15/10 channels
• 160/80/40 Triplexer to divide into individulal 80/40 channels
• A bandpass filter for each and every band: 160/80/40/20/15/10
• And, of course, about $1800, not including all the coax you need to wire all of that together.

Maybe that explains why we don't see as many of these in use as you might have guessed when you first learned about them. Having said that, for a club to spend $1000 to $2000 on a device that makes group events so much easier certainly seems reasonable, or at least possible. But for just me… Even just half that for a three-band setup is more than I've spent on my rig, antenna, power supply, digital interface and coax combined. Ouch.

Are there other options?

Turns out I'm not the first person to get to this point and try to find a less expensive option. North Florida Amateur Radio Club (NF4RC) has been working on this problem for some time. They got a suggestion from the Gainesville Amateur Radio Club that pointed them in the direction of Multiplexer and Bandpass Filter Kits from VA6AM. But they didn't just stop at the basic kits. They've gone well past that to go all the way to their own custom-designed 6-band multiplexer!

The good news is that they've created a fair bit of documentation about their journey. The bad news is that the people creating that documentation are experts, and unfortunately, often experts simply don't know what other people don't know. Also, the documentation is often created after a project has developed to a certain point. And finally, much of their project sees to be part of face-to-face lab sessions. That means that the documentation is often an encapsulation of details rather than an actual tutorial. When you're an expert (or have access to experts), that's plenty. For me… at the moment, it doesn't seem to be enough.

Fortunately, they have an active mailing list: hopefully I can lean on those same experts a bit as a slog my way through. :) And hopefully, I can flesh out those details into a more approachable tutorial. That's my goal for this project.

Be aware, while the kits allow you to save a bit of money on each device (something like 40-50%), it still ends up being a good chunk of money, and with the kits you're going to be substituting a good chunk of work. It seems the soldering and basic assembly is straightforward. The biggest hurdle is going to be the tuning. This requires a good amount of equipment and expertise to do effectively, and it seems to be vitally important. In NF4RC's project, their initial results provided half the attenuation it was supposed to in certain areas. They were able to tune and adjust to get to something like 90% of expected results; but if they didn't know what they were doing, their project would not have worked.

NF4RC also has some really good lab-grade test equipment. I've got a NanoVNA and a TinySA. Is that good enough? I think so, but I'm trying to get confirmation!

The multiplexer portion seems to be straightforward: A very filter for each tap. It seems to start with an inductor to act as a low-pass filter on the lowest frequency tap, a capacitor to act as a high-pass filter on the highest frequency tap, and an series-LC circuit to act as an initial bandpass filter for the tap(s) in the middle. Behind each of those is multiple series-LC circuits in parallel to act as bandpass filters for each range. Here is the schematic for the 20/15/10 triplexer with updated component values as found by NF4RC.

In addition to the multiplexer, you also need bandpass filters between the multiplexer and the rig for each tap/band. These seem to be similar to the bandpass filters already in the multiplexer: multiple series-LC filters in parallel, along with an occasional additional capacitor, inductor or resistor in the circuit. These seem to provide the bulk of the the isolation between your band and everybody else.

From what I can tell trying to put together the various pieces of documentation, NF4RC started with the VA4AM triplexer kit. At some point fairly early in the process it seems they decided they wanted to expand this to a quadplexer, and later expanded it to a 5- and even 6-band multiplexer. As part of that, they designed bandpass filters for each band using a common circuit board that would allow them to add the specific inductors and capacitors for that specific frequency. Right now, I do not have the fundamental knowledge and understanding to be able to create these designs. NF4RC has provided a good amount of documentation for their designs, but my inexperience does not reach to the point of turning those documents into an actual device. Given that Winter Field Day is a few weeks away, I think I’m going to have to punt on the idea of expanding the standard triplexer. As it is, I’m already pretty overwhelmed at the idea of tuning a fairly well documented kit, let alone something without that same level of detail.

Gordon Gibby, KX4Z, has documented the theory of operation for the multiplexer as part of adding a 6th band to their multiplexer. He explains this, including specific examples for the 160m and 80m bands, in this writeup for the NF4RC club. My simpleminded reduction of his document is this: a simple series-LC bandpass filter is added to the common antenna connection (or, in the case of the highest or lowest band, a simple cap or coil as high- or low-pass filter). This is to 'reduce the parasitic loss from the other bands' filters'. Past that, there are two series-LC circuits to ground in parallel. One is tuned for 1/2 the target band, and one is tuned for 2 times the target band. The purpose of this is to dump those frequencies to ground, blocking those frequencies, and in concert with the series-LC filter above, frequencies even farther away from the target frequency. But those individual series-LC L and C values are not chosen at random. They are chosen such that each of them can be considered a single L and C component of their own. The lower trap provides a specific L value and the upper trap provides a specific C value that combine to become a series-LC filter, tuned for the desired frequency. That provides a path for our desired frequency through the filters rather than being shorted to ground!

It seems that beyond that parallel pair of series-LC circuits (and the virtual series-LC that creates), there is also an additional series-LC circuit, likely to provide even more bandpass filtering. In the case of the highest and lowest taps, the additional circuit is simply a high- or low-pass filter with a single cap or coil. This is very similar to the initial filter that is connected to the common antenna, though with different values.

Beyond that, it seems like further developments were variations on that theme: figure out the parameters necessary for adding the simple series LC circuit to create a tap in the multiplexer section, then figure out the parameters necessary for the two parallel series-LC circuits to ground and the final series-LC to create the final bandpass filter for that new tap/band. And they've done this out now to six band: 160/80/40/20/15/10!

Theoretically, there's no real limit to being able to add additional bands (though see my question at the bottom of this section). However, in practice, there are some physical limitations. One issue is that even toroid coils don't 100% contain their magnetic fields, and having harmonic bands within close proximity in a single enclosure can create interference. VA6AM found that 40m in particular creates problems with the shorter harmonic frequencies. His solution was to create two different triplexers, one for 20/15/10, and one for 160/80/40, and to connect each of these to a "split diplexer" designed to simply divide the frequency range in half, rather than create individual narrow frequency taps.

The advantage of this is that this allows 40m frequencies to be kept in a separate enclosure from the higher frequencies and give you better performance. The disadvantage is that it's a *lot* more stuff: three different enclosures and a lot more more UHF fittings and coax jumpers. But another advantage is that it allows this process to be done piece-by-piece, without having to modify previously built and tested devices. It also allows you to stay with existing kits for the most part – except for the fact that there isn't a kit for the 'duplex splitter'!

Leaving that last part alone for now, the rest is pretty appealing. I can start by building the 20/15/10 kit as-is to get three simultaneous bands. Missing 40 is a pretty glaring weakness, but it greatly simplifies getting started. Adding the duplex splitter would then allow operating on 40: use the low output of the duplex splitter by itself for 40 (or 80 or 160, just not simultaneously), and then have 20/15/10 simultaneously with the triplexer. At the moment, I don't have an antenna that can really do 80/160, so that covers what I would need. But when you get to the point of wanting those bands, you can then add a second triplexer tuned for 160/80/40.

The way this path allows step-by-step growth with clear points that allow a device to stay finished is appealing to me. So that is the path I think I will use moving forward. And that starts with 'simply' building the triplexer (and the additional bandpass filters).

My question regarding additional bands: does the parallel-series-LC-to-ground/virtual-series-LC-to-rig trick require a certain minimum distance between frequencies? For example, seeing as the series-LC-to-ground traps are tuned at 1/2 and twice the desired frequency, what happens if you add taps at frequencies inside of that range? For example, adding 17/12 to 20/15/10? Or does the initial series-LC bandpass filter minimize interference enough for them not to interfere?