Notes on Duplexer Cavities for HAM bands

6m 1MHz 2% spacing, 2m 600KHz or 0.4% spacing, 10m 100KHz 0.34% !

Note, none of these commercial units use the couple a bit of energy from one connector to the other approach. Note the gold trimmer. 70cm, 1% freq spacing.

http://www.mixw.co.uk/CAVITY/cavity.htm

Now I know why. 1/4 wave lines need a low low impedance at the far end to have a high impedance at the near end, at the antenna "T".

Other cavity types (two loop ones) present a high VSWR but not necessarily low impedance.


A $2000AU TeleWave TPRD-1556 2m 5" dia. set in Italy for sale on eBay.

This one has a lot of connectors and "T" pieces, top one is better.


A picture speaks a THOUSAND words. eg. cm-1008.pdf Q series. Files:

The Telewave specs here:
http://www.unixservice.com.au/hamradio/repeaters/duplexers/TWDS-6025.pdf

http://www.vk6uu.id.au/cavity-filters.html
http://www.repeater-builder.com/antenna/pdf/ve2azx-duplexerinfo.pdf
http://www.unixservice.com.au/hamradio/repeaters/duplexers/cm-1008.pdf
http://www.unixservice.com.au/hamradio/repeaters/duplexers/cm-1009.pdf
http://www.unixservice.com.au/hamradio/repeaters/duplexers/cm-1010-mr-series-mobile-duplexer-tuning.pdf
http://www.unixservice.com.au/hamradio/repeaters/duplexers/db-4048-2.jpg
http://www.unixservice.com.au/hamradio/repeaters/duplexers/design-of-high-isolation-duplexers.pdf
http://www.unixservice.com.au/hamradio/repeaters/duplexers/duplexer-loop-tests-ve2azx.pdf
http://www.unixservice.com.au/hamradio/repeaters/duplexers/emr-corp-antenna-duplexers.pdf
http://www.unixservice.com.au/hamradio/repeaters/duplexers/fp-fr-fq-series-tuning.pdf
http://www.unixservice.com.au/hamradio/repeaters/duplexers/q202-208-218-tuning.pdf
http://www.unixservice.com.au/hamradio/repeaters/duplexers/q2330e-duplexer-tuning.pdf
http://www.unixservice.com.au/hamradio/repeaters/duplexers/seits-duplexer-theory-and-tuning.pdf
http://www.unixservice.com.au/hamradio/repeaters/duplexers/seits-duplexer-theory-and-tuning.pdf.1
http://www.unixservice.com.au/hamradio/repeaters/duplexers/telewave-tprd-1554-and-1556.pdf
http://www.unixservice.com.au/hamradio/repeaters/duplexers/tx-rx-instruction-manual-vari-notch-duplexers-with-6-inch-cavities.pdf
http://www.unixservice.com.au/hamradio/repeaters/duplexers/tx-rx-technical-notes-on-duplexer-problems-and-remedies.pdf
http://www.unixservice.com.au/hamradio/repeaters/duplexers/tx-rx-t-pass-expandable-cavity-multicoupler-cavity-system.pdf
http://www.unixservice.com.au/hamradio/repeaters/duplexers/w1gan-duplexer.pdf
http://www.unixservice.com.au/hamradio/repeaters/duplexers/w2eup-hybrid-ring-duplexer.pdf
telewave-tprd-1554-and-1556.pdf

A little box with two connectors and the tuning cap, with a loop protruding into the cavity is ideal. Make the box twist +- 30 deg. for coupling adjustment. This changes the pass-notch spacing.
Trimmers must be GOOD eg. piston trimmers. These series resonate your loop, on 6m or 2m.
Make all this in brass or copper so all metal parts outside the cavity have NO corrosion issues. Stainless screws.

2m trials on 5 inch cavities, August 2014

BIG Pictures. 35db notch, but only 8db down at +- 5MHz

Feed

Feed inside

600KHz pass to notch

Single connector, note 2nd notch, heliax as capacitor

IMG_3126

IMG_3120

600KHz notch to pass

IMG_3121

Has anyone a comparison on these two methods?

Notes: The two loops method has a simpler pattern, but cannot be swapped as the bypass (inductor in this case) needs to be changed to a capacitor.
My use of heliax, needs to be adjustable. Commercial cavities use expensive piston trimmers, or mechanically difficult, di-electric adjustable coaxial capacitors.
Two loops method, If I swap the phase of the 2nd loop, what happens?

Fantastic, the notch and pass swapped and the impedance of the inductor is MUCH increased and higher Q see IMG_3134 here

IMG_3134

Now for a capacitor trimmer...

IMG_3135

Capacitor trimmer, poor notch depth.

2015 we now have a Wiltron 640

Using copper strip, for the final product

Back to inductors, easy to adjust.

IMG_3222

Haipin 1/8" wire 52mm long 17mm wide.

IMG_3223

Undoing the screws, lifting the assy, I saw the pass - notch spacing reduce.
50mm hairpin is too long, even 35mm is too long, squeezed the hairpin to 13mm wide.

Now for the capacitor case again with shorter haipin

IMG_3224

Hairpin 32mm x 13mm wide

IMG_3225

Pass 147.050 MHz, Notch 147.650 MHz. IMG_3224 IMG_3225

Now to make the back covers.
Some people love die-cast boxes. If I could solder to them, I would too.

Complete

IMG_3228

The T coax splitter from the antenna to the first Rx and Tx cavities.

Consider the signal going past the T junction, say Rx. If the Tx cavity input impedance is 0 ohm at Rx frequency, then after a 1/4 wave line, should be infinite impedance at the T junction .:. no power lost down the Tx line.
So, what do MY cavities measure?
Tests with a Field Fox showed, no good.
I made up a test lead, since I don't trust "T" pieces, I made one with two RG58 cables going into an N plug. Thus making the shortest distance to the tested "N" sockets on the cavities.
Results: Abysmal. None of my 3 Rx cavities show anywhere near 0 ohms at the Tx frequency. Now the Tx cavities. No better.

So the theory of 1/4 wave lines from the "T" to Rx and Tx chains is in question. Further testing needed of more cavities.
Never have I seen in the literature, other than 1/4 wave lines. See the two pixtures, top of this page.
Well, I have, in 'actual designs', cutting lengths, and was puzzled.
As it happens, I have an adjustable length coax cable by GR (General Radio) with, of course, GR connectors!! (and adapters to "N").

Well, it WORKS. I can match, spot on, the line length needed for best SWR.
Then, I have to work out it's equivalent in 'normal' coax.

Too hard
Now to re-work two cavities

DSC00001.JPG 35mm at 45deg hairpin 1/8" copper wire

DSC00003.JPG Postage stamp trimmer plate, 5 micas under it. Held down with a 3mm perspex block.

Now at my antenna "T" I have better than 25db return loss at the notch frequency.
This means very little of my Rx signal is lost down the Tx cavity chain, and/or, the Tx signal is not presented with a bad SWR at the antenna "T" due to the Rx cavity chain.

Magnetic coupling, two loop design

Was this cavity made under patent protection, and therefore, others used inferior techniques such as the "discrete component (L or C) coupling outside the cavity".

Motorola UHF IMG_3231 Pass 434MHz, Notch 449MHz 60db deep!!

Brainwave Place the loops vertically down the cavity side therefore swapping
the phase of the coupling magnetic field between loops and hence swapping the notch & pass.
I drilled a new set of connector holes and it WORKS. 55db Notch 433MHz, pass 438.5MHz.(loops now closer, bent a bit and 5.0MHz split)

Motorola UHF IMG_3232 ignore the label
From The Motorola doc, they didn't try this. http://www.repeater-builder.com/motorola/t1500/t1500-cavity-app-notes.pdf
http://www.cactus-intertie.org/LA/tec_not6.pdf shows their adjustable magnetic coupling.
I tried their pictured capacitance method, could barely get 50MHz split with 3mm and 2mm wire 15mm off the side-wall.

Doesn't work.

Superb 5MHz split etc.. 10mm strip 10mm up, 15mm along, 5mm down, 10mm to connector pin. Strip bent over at 90 deg.
The pass is quite broad.

Nothing is easy

One last method just tried, at 2m a parallel resonant circuit inside the cavity between two connectors on the cavity wall. This has an obvious advantage, at the notch frequency it is a predictable high Z, making the line to the antenna "T" a 1/2 wave. See tx-rx-instruction-manual-vari-notch-duplexers-with-6-inch-cavities.pdf

5MHz markers, then one adjusts the peak towards either notch. Same as for series resonant. Bonux, can be an Rx or Tx cavity.


2x5C13-3 set from Polar, antenna "T", 433.5 & 438.5 MHz

A bit of work.

The BMARC 2m repeater cavities

BMARC_2014