Remote symmetric matchbox

Here is my second attempt to create a symmetric matchbox for my QRP purposes. I quickly realized that it is quite annoying to run to the antenna to retune after a frequency change. So I decided to create a remote matchbox based on my previous manual one.

There are some commercial balanced matchboxes. The single remote one (HamWare AT-615B) is VERY expensive and uses a lot of control wires.

When I saw that I decided on three main design goals for my prototype:

• no micro controller on the antenna side
• immunity to interference
• simple wiring

So I started thinking about the control interface and realized that if UTP 5e is good for 100 meter long lines up to gigabit speeds it must surely be good for low frequency as well. UTP gives me four differential pairs to use. SPI then needs four wires and can control a shift register directly. This combined will allow me to reuse my stash of UTP 5e cable and connectors and when I do not have any more UTP is quite cheap anyway.

Here are the first steps I took.. schematics and PCBs. I could have used ground plane construction, but I was a bit worried about the capacitance it would create.

I have already mentioned that I am using SPI as the control protocol and here is how I envision it to work:

• A micro uses SPI to send 24 bits over to the tuner. Each bit controls one bipolar relay, together they control three sets of eight relays each.
• Each SPI signal is converted to a differential pair (5V/0V levels) to prevent interference and transfered in that form to the tuner side controller. Each signal is using one twisted pair in the UTP cable.
• At the tuner side the pair is decoded to SPI signals again and those are connected to three 74HC595 latching shift registers (24 bits).
• The shift registers hold the control value and pass the bit values to 74ACT04 invertors that drive the bipolar relays (FTR-B4).
• The ACT family can source or sink more than 20mA per leg. That is more than enough to switch the state of my relays.

Transceiver side controller - SPI to UTP

Tuner side controller - UTP to relay controller lines

The controller boards use AM26C32CD receivers and ST26C31B drivers. I know I am wasting one perfectly good receiver/driver, but I have quite lot of those in my stash :)

The characteristic impedance of UTP twisted pair is 100 ohms, the same as the output impedance of the ST26C31 driver (see Table 6 in the datasheet). For that reason there are 100 ohm terminating resistors just before the receivers.

Coil switching board

Capacitor bank board

Coil and capacitor switching circuits are not special at all:

• Coils are connected in parallel and single coils can be bypassed by the respective relay.
• Capacitors are organized into banks where each bank can have 16 capacitors - four parallel capacitors four times in series to make the bank withstand higher voltages (I have assortment of 500V NP0 capacitors).
• All relays are bipolar (latching) Fujitsu FTR-B4 4.5V with 17mA latching current. I am only using QRP power so the current carrying capability (1A per section and I wire both sections in parallel). I am using the same trick to drive them as in my manual matchbox from the previous article.

I am planning to use balanced-Pi configuration and haven't decided on the capacitor values yet so I planned for the worst case. I won't probably be populating most of the capacitors. All of them will be the NP0 kind though (zero temperature coefficient).

Panel with controller boards (tuner and trx sides) + one unrelated micro-controller board

Two capacitor bank boards with relays

The boards were made by Seeedstudio and it took about a month to get them. I was busy during that time so it was not really holding me back.. but I am glad I have them at home. You might have also noticed that there is no board for the coil part. That is because I am planning on reusing the board I already have in the manual matchbox.

That I all I can describe today as I haven't started soldering yet. Once I have a prototype ready, I will post the results.

Multiband dipole, ladder line and homebrew manual symmetric matchbox

I recently found out that my dipole was a bit shorter than it should be. Since I use 75 ohm HDTV coax (Nordix CM407 Cu PE) as my feedline, I was losing power in the impedance mismatch. I also started to be a bit annoyed by it being single band only. So I started looking for a possible solution.

The only reasonable change I could do regarding antenna was to use a bit longer wire (about 44 feet), ladder line and a matchbox to match it to the 75 ohm feedline. This is the basic concept of non-resonant multiband dipole as described in ARRL Handbook and Antenna book.

I built the feedline out of 0.5mm^2 insulated stranded wire (because I have plenty..) and cut 6mm diameter hard irrigation tubing (black, 15m for 99Kč - 4€ in a local hobby store) to pieces to make the separators. I then cut slots to the separators, inserted wire and secured using black plastic cable ties. I tried many different ways of making the ladder line, but this way was the fastest and cheapest.

Then I started looking at the available ATUs and quickly figured out that the prices are very high. I won't describe the details as DJ0IP already dedicated many pages to this topic at his website.

Since I am only interested in low power (10W max) it was quite easy to come up with home-brew solution. It is a switchable L network with 16 coils (2 times 8 to make it symmetric) and a variable capacitor I got at ebay wired as split stator (the two sections in series). I also added one additional fixed 200pF/1000V (2x4x100pF/500V NP0) capacitor that can be switched in and out of the circuit.

There is also an 1:1 current balun (7 bifilar turns on FT50-43) at the input side of the tuner.

Components before wiring. There is the 200pF/1000V capacitor at the top (the PCB)
and the relay controller board in the middle.

Front panel (I added one more switch later).
You can also see the current balun at the back side right from the BNC connector.

Coil switching is accomplished by FTR-B4 4.5V locking relays (or bipolar as we call them in CZ) that are directly controlled by front panel switches. I decided to do it this way to minimize wiring length and parasitic effects. The circuit is powered by single cell LiPol battery (3.7V nominal) that is used only when relays change state and so will stay charged for a long time.

Coils, relays and the control cable on a copper clad board.

Simulation circuit of the relay controller. R1 is the relay coil and V1 is the switch.

All the coils were wound on toroids from the Ferroxcube 4c65 material. I used three different sizes of cores (9mm, 14mm and 23mm OD) and I wound the inductances so they are about 0.125uH, 0.25uH, 0.5uH, 1uH, 2uH, 4uH, 8uH and 16uH. Since there are two coils of the same inductance in symmetric series it gives me an inductance range of about 0.25uH to about 64uH.

The finished matchbox

The result was an "ugly" box that is able to match my antenna at least on 20m and higher. Those are the bands I wanted to use. I used my miniVNA with vna/J in Smith chart mode to find the settings for different frequencies. The settings for 20m and my dipole were 3.75uH and minimal possible capacitance connected to the antenna side. The fact that I tried to learn how a Smith chart works helped me a lot when figuring out the values :).