Cavity Duplexer needed?

Discussions related to schematic capture, PCB layout, signal integrity, and RF development

Moderator: robert.ghilduta

Post Reply
vonStierlitz
Posts: 7
Joined: Thu May 21, 2015 7:45 am
Location: UK

Cavity Duplexer needed?

Post by vonStierlitz » Sun May 24, 2015 6:33 am

If I need to increase the TX power do I need to purchase an appropriate Cavity Duplexer for the frequency range in question?

Can I just plug the RF Amplifier inline and use a separate TX/RX antenna?

Excuse the naive question.

User avatar
rtucker
Posts: 77
Joined: Sun Jan 25, 2015 10:38 am

Re: Cavity Duplexer needed?

Post by rtucker » Sun May 24, 2015 11:14 am

A cavity duplexer is, basically, a set of very sharp notch filters. The stereotypical application for a cavity duplexer is as part of a repeater system. Repeaters tend to have a few characteristics which make cavity duplexers ideal:

1) Full-duplex operation
2) No frequency changes
3) Receive and transmit frequencies are significantly different
4) Usually a shared RX/TX antenna, for cost, tower congestion, and link symmetry reasons
5) Hostile RF environment, with other nearby spectrum users on the same tower

It's important to remember that a cavity duplexer is a notch filter, and it's not a very agile piece of hardware. It's manually tuned for a specific frequency pair, and changing this requires, at best, hand tools and a bit of elbow grease. When you're only intending to use a single frequency pair, there are MUCH better options out there than a bladeRF (or any other frequency-agile SDR). So I'll assume a cavity duplexer is not for you. :-)

In your case, using separate RX and TX antennas is entirely doable, especially if the antennas are in each other's nulls (e.g. vertical antennas separated vertically) and they're operating far enough apart frequency-wise. Obviously this means you've got twice the feedline and twice the antennas to deal with, but it's full-duplex and it may very well be good enough. Filters are, as always, a good investment, but they don't need to be anywhere near as sharp as a cavity.

If you don't need to work full-duplex and can get away with muting RX while transmitting, triggering a transmit/receive switch via GPIO is a splendid option. This can be quite useful if your amplifier has a LNA pre-amplifier you'd like to use, too.

Or, if you're sticking within a particular range of frequencies, a circulator might do the trick for you. It's another kind of duplexer. Depending on how much amplification you're planning on doing, though, one might not provide adequate isolation...
Rey Tucker (she/her)
Systems Engineer, Nuand LLC
Rochester, NY, USA

#WontBeErased

vonStierlitz
Posts: 7
Joined: Thu May 21, 2015 7:45 am
Location: UK

Re: Cavity Duplexer needed?

Post by vonStierlitz » Sun May 24, 2015 2:06 pm

Thankyou for the most detailed reply - I did actually learn something :)

My chief worry was that by inserting inline on the TX side an RF amp (say 5W max on E-GSM 900) this could cause damage to the circuitry of the SDR (bladeRF x115).

It was the OpenBTS book that recommended using a cavity duplexer (although will be running yateBTS likely) I think needing 50db of isolation?

http://shop.sysmocom.de/products/dx900-kt30 seems good for the task but obviously don't want to buy if its not needed.

[This will be part of a POC design to be used in an area of the world where licensing will not be an issue so I will have the appropriate GSM licence]

User avatar
rtucker
Posts: 77
Joined: Sun Jan 25, 2015 10:38 am

Re: Cavity Duplexer needed?

Post by rtucker » Sun May 24, 2015 5:08 pm

Alright, you're doing full-duplex on constant frequencies, so you're on the right track with your cavity search!

Depending on what your tower situation is like, you might be good with separate RX/TX antennas, separated vertically by some distance. This is difficult to test on the ground, and there's a chance that the "leakage" from TX to RX will be below the damage threshold but still high enough to affect your sensitivity. And even if tower space is "free", you've still got the antenna itself, the feedline, lightning protection, etc, and that could cost as much as a duplexer. (Some things get cheaper and smaller at higher frequencies.)

On the other hand, a duplexer will work the same on your bench as it will at the base of the tower (assuming the RF gods are on your side).
Rey Tucker (she/her)
Systems Engineer, Nuand LLC
Rochester, NY, USA

#WontBeErased

Post Reply