I’ve used power combining to achieve about 3-db more total power output for a little over a year now and I have to say that I can’t imagine doing it any other way. 3-db may not sound like a lot until you consider that we are weak signal operators that constantly live near the detection limit, particularly on the more difficult paths. Others prefer different devices for making more power than the IRF540 that is at the center of the GW3UEP QTX-100 style amplifier . That’s OK too. There are a lot of great designs out there. What I like about the GW3UEP amp is that 1) it requires few parts, all of which I bought in bulk back in 2012, 2) it uses an inexpensive IRF540 ($1.49 for the “good” ones!), and 3) offers high reliability and survivability through abuse that is second to none. Is there a downside? Not for me but these amps are not linear so you can’t use them with modes that vary in amplitude or phase like PSK-31 and these amps need a 24+V DC power supply at 5+ Amps capability. Basically I spent about $50 for parts in 2012 which resulted in enough parts to build about fifteen complete amplifiers. I’ve built four to date. So that’s the amplifier talk. Back to combining…
Combining is the act of taking two or more signals that are in all ways equal and adding them together in-phase. Variations in phase relationship or amplitude result in a portion of the power being rejected, typically to a resistive load and wasted as heat. A well isolated two-port system will still see a 50 ohm load resistance even when the second port is inactive or no amplifier is present, however, half of the power from the operating port will be lost to the resistive load as heat.
The process starts with an input signal that is typically from a single source but can come from multiple sources as long as they are phase locked. In my case, my primary signal source for 630-meters is the MF Solutions transmit downconverter. The downconverter output passes through a small attenuator whose output is a T-splitter. The splitter is the connection point for equal lengths of coax to drive each amplifier. And I mean EQUAL LENGTHS. If you are duplicating this project and are unsure of your ability to build two, absolutely identical cables, buy the machine-built cables from a reputable seller. They will likely be identical.
Each identical cable connects to an identical amplifier. The amps should be identical but I tend to be a sloppy builder and realize there are some variations due to my construction techniques. Fortunately 630-meters is very forgiving, compared to sloppy design at microwave frequencies, which will get you into trouble. Notice I am NOT a microwaver!
The outputs of each amp feed each input port of the combiner with IDENTICAL cables. The output of the combiner can be any length but should pass through a low pass filter that is capable of handling the output power. I use a W1VD Kilowatt+ low pass filter that was built by Dave Robinson, WW2R / G4FRE . The outputs of combiners are apparently pretty harmonic rich and that only makes sense given that any harmonics that might be part of the input single will be 3-db stronger on the output. Its garbage in – garbage out!
So what about the combiners themselves? I like the W1VD zero-degree combiner. Its expensive to build (about $200 in ferrite unless your junk box is like Walmart) but the port isolation is phenomenal and I know that my amplifiers are going to see 50 ohms even if one should fail. The RG-142 used on the input ports is available from Pasternack. Power resistors are available from a wide variety of suppliers but let me make a few comments on this topic. We are all conditioned as hams to avoid wire wound power resistors because the inductance can create problems as we increase in frequency. At medium wave, the inductance in a wire wound resistor is very, very small so the impact is quite minimal. I know because I am using a pair of parallel 50-ohm wire wound resistors mounted in heat sinks in my combiner and they look like 0 ohms of reactance on the analyser. Ferrite is available from Amidon and just realize that its probably going to be expensive.
There are cheaper alternatives with less isolation but probably an equal result. Wilkinson combining is common in the microwave industry and I am pretty sure there are quite a few solid state AM broadcast transmitters that use it as a cost effective means of combining a number of smaller amplifier modules into larger total power out. Steve, VE7SL, did a very nice write up on his blog of the Wilkinson combiner after building one and trying it out on the air. This design truly is a junk box combiner. Parties interested in building their own can use this calculator to determine the values of the lumped constant components. Note that it appears Steve is also using a wire wound power resistor in his design.
Finally, when reversed, power combiners can function as power splitters or perhaps better stated, just splitters. One could use a reversed combiner to split a receive antenna among two receivers, for example or take a single power output and split it into two antennas. Perhaps this could be the basis for a phased array, for example.
So there is a lot of versatility with combining (and splitting!) and its not that complicated to accomplish. The hardest part is the exactness required in construction and as has been shown, there is a slight amount of flexibility compared to high frequencies.
A very good technical document about combining can be found here.