How To Tune A Repeater Duplexer
Ham Radio Home Homebrew 440Mhz Duplexer

This document was created by me back a while back. At the time I was very into repeaters and had a need to tune the duplexers I was building. I didn't have the money to buy the "right" equipment so I made up my own methods. I have used this method on Wacom, Decibel Products, TX/RX and my Home Brew Duplexers. It works on 2m, 70cm and even 23cm duplexers. I do not assume reponsibility for any damage to anything. If you have the "right" equipment and know how to use it - then do so. (Why are you here?) Most manufacturers will tune up your duplexers for a fee. However, the tuning could change on the way back due to vibrations and temperature extremes during shipping.

Duplexers normally do not need to be retuned. All other things should be checked first. Things do change though - vibration, corrosion etc. Regardless, testing performance is ok as long as you don't disturb anything. If you measure more than about 2.5 DB loss, or have excessive desense then a retune is recommended.

You will need a 5 watt transmitter capable of transmitting on the frequencies involved - with low spurious emmissions (>60db). A 50 ohm pad/6db attenuator ( or a tee with a 50ohm load on it and proper length cables - flawed but works). 50 ohm diode detector - (buy one or check schematic below). An analog voltmeter with a .6v full scale and other higher ranges. A digital 10 mega ohm voltmeter for accurate attenuation readings (not required). An oscilloscope works well for both if you have one. Cables with the proper connectors - 100% shielding (RG142 is good). Standard calculator for loss/att calculations. For tuning I use an analog meter, then switch to digital for the final reading. It take too long for the digital meter to settle down for each adjustment. For final loss/att calcs, I use a wattmeter for the loss and the digital meter for attenuation measurements.

This method will not work on some types of cavities. Some require that you set the notch spacing first and then set the pass frequency. Others are notch only or have the notch fixed internally.

Label all connections if you don't understand how they are connected. Especially the tee harness on the output of the duplexer set. The cable lengths may be different or possibly different type cables. Do not change the lengths or types of cables. If the cables are not 100% shielded or double shielded, get the proper cables. 9913,RG-214,RG-142BU and others are suitable. This goes for the entire repeater system. Always use 100% shielded cables, nothing else is acceptable. Also the antenna system S.W.R. should be less than 1.5:1 for the duplexer to work properly. Of course the S.W.R. at the feedpoint should also be low to assure minimum coax cable losses.

Hook up the test equipment as shown below - use as few adapters as possible. Establish a reference point without a cavity in line. (jump around cavity) Set the transmitter to 5 watt level, measure woltmeter reading and make sure the SWR bridge reads and SWR of 1:1 (if not check your cables, detector or dummy load). This will be your 0 DB reference. You will need about 1 watt into the SWR bridge - after the attenuator. The purpose of the 50 ohm padd/att is to protect the transmitter during the notch phase which can produce an SWR of more than 10:1. It also serves to keep the proper impedance to the cavity filter. Check that your tranmitter will produce the same output under different loads, especially during the notch phase - if not you will need to use more attenuation - however, this method will still work but the actuall loss/attenuation readings will be inaccurate. Always set one cavity at a time using this method. Do no key the transmitter any longer than needed to get a reading during any of the tests.

Loosen the locking mechanism on the tuning plunger. Key the transmitter on the pass frequency of the filter. While observing the SWR bridge, turn the tuning plunger to minimum SWR and maximum reading of the detector. You should be able to achieve an SWR of less than 1.5:1. This should correspond to the minimum loss position. Set the transmitter to the notch frequency of the same filter. Don't look at the SWR bridge, it will probably be off the scale. While watching the volt meter, adjust the notch adjustment on the cavity for minimum reading on the voltmeter. Recheck readings to be sure the pass readings are still OK. If you move the tuning plunger again, you must retune the notch. Always tune the notch last. Lock down the tuning plunger and notch adjustments. Check again after the lock down. Record the voltmeter reading while in the pass and notch modes for pass/reject calculations. Calculate the loss on the pass frequency and attenuation on the notch frequency. If the loss is too high you will have to rotate the connector on the cavity or somehow get the coupling loop closer to the center rod (depends on design). Some cavities have the connector base marked with DB loss markings. These figure are on a per cavity basis. Not all cavities have adjustable losses, some are fixed internally. If the notch isn't deep enough you will need to increase the loss to get greater attenuation on the notch frequency. No free ride here. You will have to check the manufacturers specifications to know what is correct or know from experience what is "normal". A 4 cavity filter usualy has a single cavity loss of .75 to 1.1 db, and 37 to 47 db notch depending on the design, quality and size of the cavity. This would give a total loss per side around 1.5 to 2.2 db and total attenuation of 74 to 94. You do not have to calc the loss/attenutation, but it is nice to know where you stand.

To calculate the loss or attenuation using a voltmeter, divide the pass or reject reading by the 0db Ref recorded earlier. Multiply by 100 to arrive at a percentage and look it up on the voltage charte. You may have to make a guestimate.

The loss can be checked with the filters all attached together. Attenuation muct be checked on a single cavity basis as the signal should be too low to get an accurate reading. Output a signal on the pass frequency on both sides of the duplexer. Note the watts going into the duplexer and out the tee harness. Also note the SWR going into the duplexer when it is fully assembled, it should be below 1.5:1

To calculate the loss by wattmeter, divide the output power coming out of the tee harness by the watts going into the duplexer. Multiply by 100 to arrive at a percentage and look up onm the power chart. Again a guestimate may be needed.

You can still use the proper log formulas, but the chart avoids you having to take a scientific calculator to the site.

After all the cavities are tuned/checked, hook it all back up and perform a desense test.

You should also check your feedline by measuring power into one end and out the other end into a dummy load. Then do the same power loss calculation. If it's over manufacturers specifications by more than 5-10%, replace it.

Single cavity tuning hookup:

                                            | C |
                                            | a |    ---------
----------      -------      ----------     | v |    | 50    |
| 5 watt |      | 50  |      | SWR    |     | i |    | ohm   |
| xmittr |------| ohm |------| Bridge |-----| t |----| rf    |
|        |      | pad |      |        |     | y |    | dect. |
----------      -------      ----------     -----    --------- 

VHF/UHF RF Diode Detector

  • D1,D2 - 1N34A geranium diodes - must be germanium type
  • C1 - 10 pf disc
  • C2 - .5 uf disc
Terminate probe with a 5 watt 50 ohm load. I solder the parts to one side of a bnc tee connector and put a dummy load on the other side.
            c1         |\|
coax  -----| |---|-----| |----|----- + to voltmeter
or tee           |     |/|    |
center          ---          ---  
            d1  / \          --- c2
                ---           |
                 |            |
                --- 	     ---
                |||          ||| 

Log formula DB=20 log(x1/x2)
Voltage Chart (vLow/vHigh*100)
Log formula DB=10 log(x1/x2)
Power Chart (pLow/pHigh*100)

Note: these charts are approximates

Desense test - NEVER hook signal generator to antenna or TX side of duplexer

-------------           -----------
|  rptr tx  |           | rptr rx |
|           |           |         |
-------------           -----------
          |                |
          |        --------|tee
          |        |       |              --------------
          |        |       |--------------| signal gen |
         ------------                     --------------
         | duplexer |
         |          |
              dummy load or antenna

With transmitter turned off or fed into dummy load if it can't be turned off. Set signal generator to rx frequency. Listen to the receiver and lower the signal generator until it sound noisy, but readable. Note the position of the generator output setting. Now turn on the transmitter or hook back to duplexer. If you are real lucky, there won't be any change in the signal. If there isn't any change, you may be able to add a preamplier and/or increase transmitter power. If there is a change, increase the signal generator level until you hear about the same noisy signal you heard before. The amount of increase is the amount of desense. If there is more than a couple of db desense, either retune the duplexer, get a duplexer with more attenuation, reduce transmitter power, reduce receiver sensitivity or get a better receiver. Also check for bad cables. If your generator has a db scale, great, if not you will have to know from experience with that generator what is acceptable. If you only experience desense with an antenna and not the dummy load, you may have an external interference problem. You will then be forced to find a spectrum analyzer and use it to find the source and possibly filter it out. You could also try an isolator to keep other strong transmitters from getting into your transmitter and being rebroadcast into your receiver. There is also the possibility that your repeater transmitter is producing excessive noise off the carrier frequency. Also while doing this test, you can check for bad connections by listening to the receiver for crackles when you wiggle connections (transmitter keyed).

Happy tunning!!! 73