Duplexer notes for the 2m Amateur band

Full duplex FM repeaters on Amateur bands such as 2m (146 to 148MHz) have very close Receiver-Transmitter frequency spacing.

Also, this is rarely done in VHF commercial systems so this requires special consideration.

Noting the very close frequency spacing of the 2m and 6m Amateur Band repeaters, 600KHz in particular for the 2m band (146 - 148MHz segment)

just bandpass cavities at perhaps 15db of attenuation will never achieve the 90db of attenuation needed. Three cavities in Transmit

and three in the Receive chain is the norm. So, we need to achieve a 30 to 35db notch with about 0.5db loss per cavity.

Cavity size is also important. A 4" set is inadequate in my experience, 6" is good.

There is a chart on http://Repeater Builder giving the cavity notch depths required for different transmitter power levels.

Here I'll discuss the BandPass / BandReject cavity types and their foibles.

1) The two coupling loops, couple a bit of energy from one to the other type with an L or C.

2) The two coupling loops, magnetically coupled.

3) A single coupling loop, series resonant to GND.

4) Two connectors with a parallel resonant circuit between them.

5) The capacitor divider method with lengths of large heliax.

Then I'll discuss a full duplexer system.

.

1) Two coupling loops, couple a bit of energy from one to the other type with an L or C.

See page 16 of reference (b). This produces a single notch and pass, spacing is adjusted by the trimmer.

One can put a coil in place of the trimmer and the pass and notch are swapped. Adjustment by bending

the coil

Two coupling loo ps, one with swapped phase

One can also swap the phase of one loop as in the RHS image and, to get the same Rx Tx offset use

a larger value L or smaller C.

Downside:

At the notch frequency, the Return Loss is bad or SWR looking into it though high. is not a known impedance, high or low.

UHF Two loop, magnetic coupling, 5MHz Rx Tx offset

This makes it difficult to proceed it with a length of coax to get a very high impedance at our antenna "T" where we split off

the "other", Rx or Tx signal.

.

2) Two loops magnetically coupled.

Again, this gives one pass and one notch. Swapping the pass and notch, I couldn't find a successful way.

Downside: As above, impedance presented at.notch frequency unknown.

.

3) Single coupling loop series resonant to GND

The most popular method.

If the resonance of the series L and C is close to the cavity frequency we get two notches with a pass at

the cavity frequency between these notches. By adjusting the capacitance with a trimmer or in my picture, adjusting the area of

the coil, one can adjust the resonance lower or higher than the cavity thus moving a notch closer to the pass

frequency. One can also adjust the angle of the assembly in the cavity hole and adjust the coupling of the

loop. Adjusting both of these, one can get good notches with low insertion loss and good SWR

Single loop, series resonant to GND, no N connector to maximise performance.

Downside: Requires good T pieces which is why I chose to link two coaxes inside the cavity to do away

with the T piece. Worked but is unwieldy.

Upside: Very low Z at notch frequency so use quarter wave lines to the antenna T.

Example for 2m[1]

.

4) Two connectors with a parallel resonant circuit between them.

Only mentioned in Reference (a) below.

Again, we get two notches with a pass between them.

Both Rx and Tx loops, parallel resonant

Better performance by about 3db. No N type T pieces

Note the loop design, this makes much easier pass to notch spacing adjustment.

Notice in reference (a), the low side loop had an extra two turn coil atop. This was not really

mentioned in the text. I copied it with wire and noticed much easier Tx Rx offset adjustment.

The coil design is not critical. These give 35db notches

Advantage: The impedance at notch frequency, High Z. Use a half wave line to the antenna T.

.

Things to try: Series loop design, with the coil like the parallel design.

Bottom right: My set of test cavities for testing. Note, no inter-connect cables came with them.

My test set of six VH F HighBabd cans

Often the case, as these were pulled from commercial service.

Why did I not use piston trimmers? To show what can be done, with junk-box bits.

.

The full duplexer system.

1) The antenna T here we separate the Tx and and Rx chains.

2) The cavity inter-connect cables.

References

a) Repeater-Builder on Coupling Loop Research: http://www.repeater-builder.com/projects/coupling-loop-research.html

b) Repeater-Builder article by VE2AZX: http://www.repeater-builder.com/antenna/pdf/ve2azx-duplexer-info.pdf