A Five Band, Kilowatt Capable, Switched L-Network, Remote Antenna Coupler
Instant bandswitching, a perfect match on five bands, what's not to like?
When I put up Sky Hook-I it was my only antenna and I needed a remote antenna coupler that would match the antenna on any frequency from 1.8 to 29.7 MHz including oddball military frequencies for MARS operations. And, it needed to handle 1.5 KW. And, I needed it right now. The solution came to me in the form of a military surplus Harris RF-601 antenna coupler.
But, eventually, I got Sky Hook-II up in the air. That antenna is a much better fit for the low bands and for MARS so the vertical monopole became my 40 through 15 meter antenna as initially designed. The wide ranging, bullet proof Harris RF-601 became overkill and I sold it.
The replacement? A home brew remote switched array of L-Networks.
For matching the impedance of single ended antennas like an Inverted L or a vertical monopole you just can't do much better than an L-Network. An L-Network has only two components, only one "proper" tuned condition and, generally, the lowest loss of any matching circuit.
Further, an L-Network can be designed in a high pass or a low pass configuration:
What makes a High Pass L-Net be "High Pass"? Think about the basic AC behavior of a capacitor. What happens when you raise the frequency of an AC signal applied to a capacitor? The impendance goes down. Less impedance means more of the AC signal passes through the capacitor. High pass.
Conversely, consider the inductor. What happens when you raise the frequency of an AC signal applied to an inductor? The impedance goes up. More impedance means less of the AC signal passes through the inductor. Low pass. In the above diagrams you can see the signal must pass through the capacitor in the high pass design and through the inductor in the low pass design. Pretty simple..
Why does this matter? Well, if you have trouble with RF from a local, high power AM broadcast station you could design your matching network in a high pass configuration to attenuate the RF from the broadcaster. Or, if your ancient Class-C boat anchor transmitter tends to emit considerable harmonic energy. Then you could design your matching network in a low pass configuration to help attenuate those harmonics.
Another design consideration is matching network loss. Generally, inductors have much more loss than capacitors so your design may favor high pass or low pass depending on calculated loss. Other considerations may include what components you have on hand, their physical size, etc.
Oh, and one more consideration. The high pass L-Network has one advantage that I really like. Since the inductor is the shunt component it holds the antenna element at DC ground. So, no static build up. There is something about that fact that I find appealing..
The first order of business? Determine the impedance of the antenna at each proposed frequency of operation. I'm a CW guy so I chose to measure the antenna impedance at the CW end of each band, 40 through 15 meters. My tool of choice for this sort of thing is the Array Solutions AIM-4170C Vector Impedance Antenna Analyzer.
Here is what the AIM4170C told me about the antenna impedance on each band:
| Band |
Antenna Resistance |
Antenna Reactance |
| 40 Meters | 19 Ω | -J 74 Ω |
| 30 Meters | 114 Ω | +J 114 Ω |
| 20 Meters | 752 Ω | -J 7.6 Ω |
| 17 Meters | 152 Ω | -J 282 Ω |
| 15 Meters | 91 Ω | -J 179 Ω |
The next order of business was to calculate the values of the required capacitor and inductor for an individual L-Network for each band. I also needed to know the voltage across and current through each of the components. I intend to run full legal power of 1.5 kilowatts (except on 30 meters) and there is no sense choosing components that are either too small or too large.
Some quality time spent with the ARRL Antenna Handbook and a spreadsheet and I was in business:
| Band | Capacitance | RF Voltage | RF Current | Inductance | RF Voltage | RF Current |
| 40 Meters | 205 μμF | 855 Volts | 5.5 Amps | 1.11 μH | 938 Volts | 13.5 Amps |
| 30 Meters | 158 μμF | 773 Volts | 5.5 Amps | 8.4 μH | 864 Volts | 1.1 Amps |
| 20 Meters | 61 μμF | 1434 Volts | 5.5 Amps | 2.2 μμH | 1485 Volts | 5.3 Amps |
| 17 Meters | 50 μμF | 1348 Volts | 5.5 Amps | 1.1 μH | 1402 Volts | 8 Amps |
| 15 Meters | 55 μμF | 1070 Volts | 5.5 Amps | 0.7 μH | 1138 Volts | 8.7 Amps |
Not too bad; In the worst case scenario, the capacitors need to have a voltage rating greater than 2000 volts and a current rating greater than 5.5 amps. The inductors need to have voltage rating greater than 1500 volts and a current rating greater than 15 amps. Those numbers seem quite reasonable!
This is a good moment to say a word or two about circuit losses. I mentioned earlier that L-Network losses can be different between a low pass and a high pass design. I favor a high pass design simply because the shunt inductor keeps the antenna element very nearly at DC ground. But, it seemed prudent to spend a few moments looking at circuit losses. At 1500 watts input, the following network losses were calculated:
| Band | High Pass Loss |
Low Pass Loss |
| 40 Meters | 48 Watts | 41 Watts |
| 30 Meters | 7 Watts | 20 Watts |
| 20 Meters | 34 Watts | 33 Watts |
| 17 Meters | 45 Watts | 29 Watts |
| 15 Meters | 39 Watts | 22 Watts |
As you can see, the Low Pass configuration does provide somewhat less loss on some of the bands. For example, on 15 meters the High Pass configuration results in seventeen more watts of loss than the Low Pass configuration. As a rule of thumb, one could say that if the series component current flow is less than the shunt component current flow then one should choose a Low Pass design to reduce circuit losses to a minimum. Or, one could do as I did and ignore the additional loss on some bands and go with a High Pass design -- thinking that the advantage of a grounded antenna outweighs the relatively minor additional loss.
And, while we're on the topic of circuit losses, one more thought. It is often pointed out that in an L-Network the losses in the inductor are usually greater than the losses in the capacitor. I thought I'd examine that assertion. Take the 15 meter L-Network which in my case exhibits 39 watts of loss. Calculating the losses in the individual components revealed that the capacitor provided 4 watts of that loss while the inductor provided 35 watts of the loss. Enough said..
So, with all this data in hand and the component values seeming reasonable I decided the project was a "GO" and began working on a design. Let's start with the RF Deck. We'll look at the controller later.
The Switched L-Network RF Deck (click for full size view)
It's really quite simple. The RF path starts in the lower left corner with an RF input jack which is connected to the RF In buss. If an L-Network relay is energized the RF flows through the L-Net to the RF Out buss and on to the antenna. Pretty basic stuff..
For the RF input jack, I use an N-Style bulkhead connector (dual female) that is also a lightning arrestor. The RF flows to an input buss that connects to each of the six relays that select one of the five L-Networks or the bypass circuit. The bypass circuit is provided so I can connect directly to the antenna for making measurements and for use as a general coverage receive antenna. The five L-Net relays are DPDT so that they switch the input and the output of an individual L-Net to the RF In and RF Out buss respectively. Additionally, when a relay is not selected it not only disconnects it's associated L-Net from the two RF busses but also shorts the L-Net inductor and capacitor and grounds them both. Finally, the RF flows through K7. This final relay serves to connect the RF Out buss to the antenna whenever the coupler controller is turned on. When the coupler controller is turned off K7 disconnects the antenna from the RF Out buss and grounds the antenna.
The rest of the RF Deck circuitry is to connect the control relays to the control cable connector. I have two Harris remote controlled antenna couplers (an RF-382A and an RF-351) and various home brewed remote couplers so I've settled on using the same mil-spec 14-Pin controller connectors that Harris uses as well as N connectors for the RF connections. This allows me to swap any of my couplers into a given antenna circuit at any time.
And, that's about it for the RF Deck. Now, for the coupler controller circuit:
The Switched L-Network Coupler Controller (click for full size view)
The coupler controller is also very simple. There is S1 (not shown), the power switch energizes the 24 VDC power supply (also not shown). There is S2, the bandswitch that routes one side of each L-Net selection relay coil to ground, energizing it. And, there is S3 that routes the relay control lines to either the bandswitch or to a DB-9 connector. Eventually, I hope to have a rig that has bandswitching outputs that I can connect to the DB-9 so the antenna coupler will "follow" the rig automatically as I change bands.
And, that's it! The whole thing is really quite simple. But, you ask, what about construction?
Ever hear the cliche that says: "The devil's in the details"? Well, it's true. The electronic design was easy and could be done from the comfort of my favorite armchair. The physical design was a bit more challenging although still lot's of fun.
The first order of business was to start assembling parts. I had a good selection of variable capacitors, insulators, vacuum relays and related stuff in the Junque Box. But, not the big DPDT relays I knew I would need for switching the L-Networks. Some time ago, I had a Palstar BT-1500 balanced feedline antenna coupler and I had noted that Palstar used large AC switching relays in their RF circuits. So, I thought I would give that a try. Those big power relays tend to be a bit pricey so I started watching for them on ebay. Sure enough, after a bit of patience a set of five 30 Amp AC switching relays showed up for a decent price and they ended up winging their way to me.
Next, I need some kind of material for the ground, RF input and RF output busses. I found some really nice 1/8" by 1" by six foot long copper bar material at a machine shop supply company, Enco, that I frequently do business with. I had some stand-off insulators that were 1" wide so I figured the 1" wide copper bar would make a tidy installation. I chose 1/8" thickness because I thought I might need the structural integrity. I think 1/16" thick would have been just fine and would be somewhat less expensive.
I looked at the various coils I had accumulated over the years and measured most of them to see what I had to work with. It turned out that none of them were especially close to what I needed. And, most of them were wound with what looked to be 14 AWG wire. Remember the L-Network loss calculations from above? I wanted to reduce inductor losses so I decided I needed coils wound with larger wire. The next time I found myself in a big hardware store I took a cruise through the wire department. There I found a 25 foot long piece of 6 AWG commonly used for grounding household electrical panels. Oh, it's beautiful stuff! I grabbed a roll and headed home to do some more calculating to see if 25 feet would be enough.
I had read somewhere that for highest Q a coil should be roughly as long as it is wide. So, armed with my calculated inductance values from above I set about determining how many turns at various diameters would be appropriate for each of the five coils I needed. After a bit of fooling around with a spreadsheet I decided that 2" diameter coils would be suitable in all cases. A bit of experimenting with the 6 AWG wire led to the discovery that winding coils on a 1-3/4" form resulted in 2" diameter coils. So, winding my own coils looked like it was going to be easy and I found I would be able to get four of the coils wound with some wire left over. (The 30 meter coil would be made from some B&W 2" coil stock I had in the Junque Box.)
And, what to put the coupler in? I happened to have a couple of military surplus transit cases stashed away. I'd picked them up from ebay for a good price and it looked like one of them would be about the right size. I like these roto-molded transit cases - they're plenty durable, water and air tight and seem to survive pretty well in the harsh Montana climate.
What about a base? I didn't want to use metal. There is no sense surrounding my soon to be high-Q coils with Q robbing sheet metal. Plus, there's a long standing tradition of building Ham gear on breadboards so I thought, "Hey, why break with tradition?"
These days breadboards are called "cutting boards" and they're no longer made of wood. And, this is a good thing. Cutting boards are now made of various synthetic materials including nylon and UHMW plastics. These materials exhibit excellent dialetric characteristics, are easy to machine and look good. I found a nice 5/8" thick cutting board at a local discount kitchen store and snatched it up. The 5/8" thickness seems better than the normal 1/2" thickness. It is more sturdy and tapped holes offer more threads to mating screws. And, best of all, it was $10!
Now, armed with all the parts I set about the task of making everything fit. The first thing to do was to trim the cutting board to fit in the transit case and to begin laying out component locations and drilling and tapping mounting holes. Some basic measuring tools and a drill are all that's required.
With the end trimmed off the cutting board and mounting holes drilled and tapped I started mounting up the various components. Below you can see the RF Input buss with three of the relays already mounted. I soldered short pieces of wire to the RF Input buss to attach to the left-most normally open connector on each relay. Next to the RF In buss are three standoff insulators that will hold the RF Out buss. The Ground buss will run along the top of the relays, connecting directly to the common relay contacts. Finally, near the bottom edge of the cutting board you can see the RF Output relay (K7) just to the left of center and the Bypass relay (K6) near the lower right corner.
OK, all the L-Network switching relays are in place.
In the picture below you can see that I've mounted the Ground buss. It is screwed right onto the common posts. I used longer screws than those that came with the relays and copper washers between the relays and the ground buss. On the right end of the ground buss you can see the female N connector -- it's mounted directly to the ground buss and the center conductor drops down and connects to the RF Input buss underneath the relays.
In the picture below you see the RF Output buss mounted atop the three stand-off insulators. There is a short length of stranded 12 AWG between each relay and the RF Out buss.
With the relays and RF busses all in place it was time to wire up the relay control lines. I ran a short jumper between the DC control connections on each relay and the relay mounting screws. That brought the DC control connections to the bottom side of the cutting board where it was easy to then run control wiring. I ran the control wiring orthogonally (always wanted to use that word) to the RF busses to minimize coupling between the RF and the DC lines.
Now, it's time to get to the nitty gritty of this whole thing. The actual L-Networks. Shown below is the 40 meter L-Network going into place. The capacitor is out of my Junque Box and I've just wound the coil.
And, voila! The 40 meter circuit is in place!
And, now all five L-Networks are in place. This thing is starting to take shape!
Here's a slightly different perspective. I've wound the coils to match my data and I plan to spread or squeeze them for fine tuning. The left-most circuit is for 15 Meters. Then, left to right, it's 17, 20, 30 (the small coil) and finally the 40 Meter circuit on the right.
Here's what the coupler looks like mounted in the bottom half of the transit case. At the left side you can see the feed-through insulators that carry the RF to the antenna and the ground system. On the right side is the 14 pin control cable connector and the bulkhead N connector / lightning arrestor.
Here's a different view. I need to wire up the DC control lines from the relays to the terminal block and to put in a jumper of LMR-400 Ultra Flex between the bulk-head N connector to the N connector on the ground buss. And, then the RF deck will be finished!
With the RF deck done it's time to turn my attention to the Coupler Controller. I've decided to use a single pole, six throw ceramic RF switch as my bandswitch. A much lesser switch would have easily done the job but I wanted 60 degrees of separation between each band because it would allow a really cool front panel layout.
I was planning to build a little power supply to provide the 28 VDC for the relays but while shopping for a suitable transformer I stumbled upon a 28 VDC 1.8 Amp switch mode power supply for $11. I couldn't pass that deal up, it was cheaper than the transformers I was finding.
The box came from ebay and only cost a few bucks. It"s the perfect size for the coupler controller.
I found knobs that are a really close match to the knobs on my beloved Drake 4-Line and at less than a buck each I've stocked up on them. They'll look great on this box.
So, parts in hand I began wiring up the Coupler Controller. I assembled the wiring harness for the 14-pin control cable, the Manual/Auto switch and the DB-9 connector before mounting those components. It was much easier that way.
Next was mounting the components to the front and rear panels.
The final wiring is taking place in the following picture. The 28 VDC power supply made for a tight fit. Maybe next time I'll stick to the original plan and build a simple linear supply. It doesn't need to be regulated as it's only controlling relays so the fancy switch mode power supply wasn't really necessary. But, hey! It was too good a deal to pass up..
Final assembly is complete and the coupler controller is in place! Here you see it turned on, in manual mode and the 40 Meter L-Network is selected. There is an LED at the end of each band selection line so you can readily see what band is selected. If I ever get a rig that has band output lines I'll connect it to the controller. At that time I'll use a round knob without an indicator line on it for band switching and just rely on the LED indicators to tell me what band the antenna is set for. But, for now, it's in place and I love it!
Final notes:
- I had the front and rear panels machined and painted by Front Panel Express. They're great to work with and the end result is well worth the price.
- Most of the parts for this project were in my Junk Box or obtained at good prices on ebay. A few items were purchased new. All totalled, I probably spent close to $300 on the entire project.
- The individual L-Networks needed almost no tweaking and the SWR is perfect on all bands. It's fantastic to jam 1500 watts down the line and see nothing coming back on the power meter!
A perfect match on five bands, instant bandswitching, no manual tuning and it'll take GOBS of power. Perfect!