Homebrew VHF and UHF Diplexer

Glen Ross, G8MWR

Diplexing

There are now several good dual- band (144/430MHz) antennas available. Unfortunately, many of the dualband rigs which are available have separate antenna input sockets for each band. How do you cope with the problem of getting one plug into two sockets? The answer is a simple bit of circuitry called a diplexer. This is a device which sorts out the various frequencies and routes them to the appropriate rig. They are available commercially at a rather nasty price but those that I have measured, whilst safe to use, do not show up any too well on separation and also tend to have an unacceptable loss when placed in circuit.


Figure 1

The Circuit

The circuit of a home-made diplexer which is well within the construction capabilities of the newcomer to homebrewing is shown in Fig. 1. It consists of three coaxial sockets and four series- resonant circuits. Hopefully you will remember that a serics-resonant circuit has a very low impedance at resonance and a high impedance off resonance. How does the circuit work?

Consider a 144MHz (2m) signal coming in on the antenna socket SK2. The tuned circuit L2/C2 is resonant at 144MHz and. having a low impedance, passes the signal to the 2m output SK1. The tuned circuit L3/C3. being resonant at 433MHz, exhibits a high impedance at 2m and so stops the 144MHz signal from reaching the 70cm output socket SK3. On 433MHz the opposite action takes place.


More Protection

The action already described will do a fair job but it can be improved upon. The tuned circuit L1/C1 which is connected from the 2m output to earth is series resonant at 433MHz and so any signal at that frequency which manages to find its way through L2/C2 is shorted to earth. As it has a high impedance off resonance. L1/C1 has no effect on the 144MHz signals. The tuned circuit L4/C4 is series resonant at 144MHz and removes any leakage at that frequency which reaches the 70cm output socket.


Specification

How well does the circuit do its job? Looking first of all at the insertion or through loss. this was measured at less than 0.1dB on 144MHz, and was slightly higher at 0.17dB on 433MHz. When you consider that you need a loss of 3dB to lose one S-point of signal strength, these losses can be disregarded. The blocking of 144MHz at the 70cm output, and of 433MHz at the 2m output was greater than 60dB. This means an unwanted output of 1 microwatt for every 1 watt of power applied, which is more than satisfactory.


Construction

The unit can be built in a small diecast box, and a suitable layout is shown in Fig. 2. Trimmer capacitor types required will depend on the transmitter powers to be used. Ceramic piston and compression types are suitable for low powers, for higher powers airspaced trimmers (e.g. Jackson C804 series) will be necessary.-Ed.

Tuning the unit is simple. First connect the rigs to the correct output sockets. Until all the following steps are completed DO NOT TRANSMIT.

Tune the 144MHz rig to a strong signal and adjust C2 for the highest S- meter reading. Tune the 433MHz rig to a strong signal and adjust C3 for the best S-meter reading.

Now connect the 144MHz rig to the 70cm output on the diplexer and the 433MHz rig to the 2m output. Tune to a strong 144MHz signal and adjust C4 for minimum S-meter reading. Tune to a strong 433MHz signal and adjust C1 for minimum S-meter reading. For safety, run through all the above steps a second time then reconnect the rigs to the correct outputs and the job is completed.

Adjustments can be made using a Micro VNA.


Reference:
Practical Wireless, October 1988, page 22
Ham Radio, September 1989, pages 30 and 31

A 2m/70 cm diplexer with a difference

Frank, ZS6TMV / PA3GMP
Based upon a design by Henk, PA0HVA, published in Elektron, June, 1994


Note that this design would need to be scaled to the US 2m and 70cm bands.

Diplexers - filters intended to split and combine signals on different frequency bands - generally use conventional L/C networks: a low pass filter for the 2 m band and a high pass filter for the 70 cm band, each consisting of several coils and capacitors with a cut-off frequency around 250 MHz or so. The maximum RF power that the diplexer can handle depends mainly on the breakdown voltage of the capacitors. Your average ceramic capacitor is rated at 50 V, so your diplexer will start to burn out at about 50 W. High voltage (trimmer) capacitors can be hard to find, and can be bulky enough to interfere with optimal filter construction, thereby increasing the diplexer's insertion loss and reducing its suppression.

The diplexer described below takes a different approach to the task at hand. It is based upon two quarter wave coax stubs, each com- bined with a simple filter that only requires one trimmer capacitor each. At resonance, the filter grounds one end of the stub, this results in a high impedance for that frequency at the other end of the stub.



For example, a 145 MHz signal connected to the left hand side (2 m) terminal will traverse the left hand side coaxial line to the common terminal, virtually without experiencing any loss. The left hand side filter resonates at 434 MHz, and will therefore not affect the 145 MHz signal. The right hand side L/C circuit is resonant at 145 MHz, though, and the right hand side end of the right hand side coaxial stub will therefore be grounded. The right hand side coax stub will present a very high impedance to the common terminal as a result, thereby stopping the 145 MHz signal from continuing into the right hand side coaxial line. The same (in the opposite direction) applies to signals at 434 MHz.

This results in a diplexer that can be used to connect two antenna's to a dual band transceiver, or two single band transceivers to a dual band antenna.

Construction is simple, but must be done accurately, and it is important to use the proper components. The coaxial cable should ideally be RG-316, which is a very thin, 50 ohm coax with Teflon insulation and dielectric. Its velocity factor is 0.695, which means that a quarter wave stub for 145 MHz will be 359 mm, and 120 mm for 434 MHz. The main advantage of RG-316 over RG-174 (another thin but non--Teflon 50 ohms coax) is that good, short soldered connections can easily be made to the braiding, which is essential for the filter's performance. Other coax could be used as well, but thicker cable (such as RG-58) makes it more difficult to solder the cable and fit it in, while non-Teflon varieties are much more difficult to solder close to the dielectric. Some alternatives are RG-142 (more or less a Teflon variety of RG-58), RG-174 (thin, non-Teflon, with a higher insertion loss than RG-317) or, if you really have no other option, RG-58. If you use anything else than RG-316, though, YOU MUST OBTAIN THE EXACT VELOCITY FACTOR for the cable you use from the manufacturer's data sheet. Do not guess, do not use rule of thumb, do not use the specs for the same type of cable from another manufacturer. Then recalculate the length of the stubs to the millimeter.

The trimmer capacitors should be of the best quality you can get. Ceramic types are preferable because they can take higher voltages - (and therefore more RF power) while tubular capacitors are preferred from a construction standpoint because they can be mounted directly into the chassis of the box. (Hint: that old valve radio stuff that you passed up on, the last time you were at a ham radio flea market generally has the caps you need!) I used tubular trimmer capacitors of 6 and 12 pF for 70 cm and 2 m, respectively. It does not really matter if the caps you use are a few pF over the 'desired' capacity - that is why they call them "variable." The ones I used are rated at well over 200 V, which means that the diplexer allows for more power than my ham radio license.

The coils are made out of 1.5 mm solid copper wire. I used silver plated wire, but you can also use enamelled copper wire without any appreciable loss of filter quality. The coil for 2 m (200 nH) consists of 7 windings on the smooth end of a 9 mm drill bit, stretched until the length of the coil is 20 mm. The one for 70 cm (67 nH) has 5 windings on a 6 mm drill bit, stretched until the coil is 10 mm long.

The shield. of the coax is soldered directly to the lugs, while the core is clipped off at a length of 5 mm or so, and connected to the center pin of the terminals. See the photo for construction details.



Note that the length of the wire between coax and center pin is not counted as part of the length of the stub, i.e. the 70 cm stub should be 120 mm along the length of the shield, plus 5 mm of bare center lead at each end.

Alignment is simple. Apply a 145 MHz signal to the common terminal. Connect a power meter (SWR meter, directional watt meter or RF volt meter) to the 434 MHz terminal, and a 50 ohms dummy load to the 2 m terminal. Adjust the 145 MHz variable capacitor until the RF signal at the 434 MHz terminal dips to zero. Then move the power meter to the 2 m terminal and the dummy load to the 70 cm terminal, apply a 434 MHz signal to the common terminal, and adjust the 434 MHz variable capacitor until the RF signal at the 145 MHz terminal dips to zero. (Don't get confused here: you should connect the Wattmeter to the 2 m terminal when adjusting the 70 cm filter using a signal at 434 MHz, and vice versa!) I used two separate Wattmeters, but this is by no means necessary.

It is easier to use a Micro VNA to set the trimmers.


Measured performance

The suppression of the diplexer is especially important when you use the diplexer to connect one dual band antenna to two single band transceivers. A single band radio can be expected to have an input filter providing at least 20 to 30 dB of suppression of out- of- band signals. That brings the minimum total suppression to 45-55 dB or so, which means that if you use two single band transceivers and one dual band antenna, 100 W at 145 MHz will result in only a few mW into the first stage of your 70 cm transceiver, and vice versa. Both transceivers will be able to handle this with ease.

For the US Amateur Bands - the 2m stub for 146 MHz would be 357 mm, the 70cm stub for 446 MHz would be 117 mm.

Reference:
Radio ZS, May-June 2009, pages 15, 16, and 20