I understand that most single-band repeaters will use a tuned cavity duplexer to isolate the received signal from the transmitted signal, since the frequencies are very close and extremely sharp filtering is needed to separate them.

But while the frequencies are close, another difference between the incoming and outgoing signals is just that: they are traveling in opposite directions! Could a repeater use a directional coupler to separate the two signals?

  • $\begingroup$ friendly reminder: please accept an answer or clarify what is not satisfactory about the answers you've gotten. This site stops working if askers don't give feedback!! $\endgroup$ Commented Apr 15, 2017 at 10:24

4 Answers 4


Are you thinking something like this?


simulate this circuit – Schematic created using CircuitLab

Let's say you have a really good directional coupler with a directivity of 55dB. And the antenna is a good match with a VSWR of 1.1, or equivalently a return loss of 26 dB. And the transmitter power is 10 W, or 30 dBm. And we'll say we select a directional coupler with a coupling factor of 10 dB (though as you'll see, the coupling factor doesn't really matter).

30 dBm, after the 55 dB directivity of the coupler, and the 10 dB coupling factor, means the receiver will see 30-55-10 = -35 dBm at its input from the transmitter.

Let's say we want to hear a 1 W (20 dBm) station that's 1 km from the repeater. At 144 MHz that's a free space path loss of 76 dB, so the power received at the antenna will be in the neighborhood of 20-76 = -56 dBm. Maybe a little more after adding antenna gain.

That -56 dBm is then subject to the 10 dB coupling factor, so at the receiver it becomes -66 dBm.

Signal power: -66 dBm
Noise power (from the transmitter): -35 dBm
SNR: -12 dB

Obviously this is not going to work. And it's actually worse than this, because any reflected power from the antenna will also show up at the receiver. The 30 dBm of transmitter power, after the 26 dB return loss, is 4 dBm. Then subject to the 10 dB coupling factor is another -6 dBm of transmitter power finding its way into the receiver.


In addition to what Kevin said, if you can, you'd usually use frequency duplex.

That has technological reasons:

  • It's nearly impossible to build a broadband (i.e. $\frac{f_\text{max}}{f_\text{min}} < 0.75$) directional coupler – most designs depend on the wavelength being within close limits. Filter design is a lot more flexible.
  • for isolation purposes, and within reasonable size constraints, practically only circulator are viable; those are finely tuned, very expensive devices, compared to filters
  • the isolation of directional couplers that aren't extremely narrowband is usually much worse than what's realizable with filters. For example, you can reasonably buy 20dB-isolating circulators. Good ones sometimes reach ~43 dB. Compare that to the out-of-band suppression of your cavity filter

Among the problems is that a directional coupler only discriminates based on direction, and in a real system there will be reflections from many elements of the system beyond the coupler:

  • Imperfect transmission line
  • Imperfect antenna
  • Birds sitting on the antenna, or other environmental effects detuning it
  • Reflective objects in sight of the antenna

Any such reflection would end up in the receiver. A duplexer will instead discriminate based on frequency, so the reflection is re-reflected back out the antenna (very similarly to the function of an antenna tuner).

Directional couplers can be used when the reflections from any of these sources are actually what is of interest, usually in RF test equipment or some types of radar.

  • 2
    $\begingroup$ Yep, radar is the "classical" use case for circulators. Those things tend to get a little expensive if they're any good, and thus, they're most dominantly found in "high-cost" systems. In many radar systems, you can avoid using one either by having highly directional, separate RX and TX antennas, or by switching between RX and TX. Also, you need massive dynamic range,or the ability to work with heavily distorted/hidden RX signal,but since you know the TX, you can also subtract it from your receive signal – that allows people to do OFDM radar with communication devices. $\endgroup$ Commented Mar 10, 2017 at 10:50

One small possible quibble to W8II's answer; both transmit and receive should be affected by feedline losses; because returned signal has to go both ways down the feedline whereas the received signal only has to go one way, I think you'd need to subtract "x" from both sides as well as an extra x from the returned transmit power.

His basic point still stands -- you can't do duplex because the local signal will overwhelm the remote signal at the receive, even under near-perfect SWR conditions.

However I am looking into using a directional coupler with a QRPp SDR transceiver which has both TX and RX ports, so that it can use a single antenna in simplex mode (the alternative would be a coax switch or relay). Because the power is pretty low to begin with and I am not planning on using a big amplifier (will probably not amplify more than 2 watts), I think/hope that the attenuation from directivity, coupling, and feedline losses ought to be sufficient to keep the receiver from getting fried by the returned power on the transmit signal.

EDIT: I originally said that the amount to subtract is 3db for double-feedline-loss, but after a good night's rest realized how stupid that was. Yes double is 3dB but that is relative to x, not relative to our starting transmit power.


You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .