The WB5WPA 6M Duplexer (as Modified by KF6YB) General - The WB5WPA Six Meter Duplexer consists of six band-reject coaxial "stubs" fabricated from 1-5/8" Heliax. The stubs are inter-connected with 1/4 wavelengths of RG/58 coaxial cable. The duplexer exhibits over 70 dB of attenuation with 1.5 dB of insertion loss compare this with the DB PRODUCTS 30-50 MHz duplexer's insertion loss spec'd at 2 dB and 800+ bucks! If more attenuation is required, another section can be simply added to the transmit leg (to reduce transmitter noise at the rcv frequency) or added to the receive leg (to knock down the transmit signal more as seen by the receiver) at the cost of slightly more insertion loss. ------------------------------------------------------------------------------ Stub construction - Each of the band reject "stubs" is constructed using a length of Heliax cable with one end shorted. The length is about 9.8% short of the electrical 1/4 wavelength making the "stub" inductively reactive. Use the stub calculator provided on the main web page to determine the length required. I used some 1/4" strips of copper stock to short the end of the stub by soldering them from the outer conductor to the inner conductor in a cross fashion. Prepare the stub by filing the outer shield until you get shiney copper and tin the copper to make the strips assemble easier. Now for the business end of the stub: I started by tinning the BNC connector with a bit of rosin flux and solder Dont be afraid to glop some solder on the threads as this will come in handy later when you go to solder it to the shield of the Heliax. I prepared the Shield of the Heliax with a file to expose shiney copper then tinned it using a propane torch, flux and solder. I then used a pair of pliers to hold the BNC and soldered the BNC to the tinned shield. It didnt take much heat to make the solder fluid, touch the BNC to the shield and let them cool. This makes a surprisingly strong bond. Try to keep both BNC's 180 degrees from each other and in line. Series Capacitor "Gimick Capacitor" is made of a piece of RG-8/RG-213/214 center conductor, dielectric and braid, pushed down into the center conductor of the Heliax. This yields a high voltage and high Q capacitor with a value of between 5 and 8 pf. The starting length of coaxial cable required is 5". Remove the black jacket, fold the braid back onto itself on both ends about 1". Cut one end flush with the braid leaving a dielectric length of 4". Carefully remove 1" of dielectric from around the center conductor on the other end. This end will stick out of the Heliax when the braid, dielectric and center conductor are inserted *into* the Heliax's center conductor. The center conductor of the RG/8/213/214 piece is then soldered to a piece of .1" wide brass or copper stock which is connected between the two center conductors of two BNC connectors whose bodies have been soldered to the outer conductor of the stub. The BNCs form what we call the 'GOZINTA' and 'GOZSOTA' (it doesn't matter which is which). See the Picture on the main web page. At this point, you have a completed stub ready for rough tuning. Sweep the stub to find the notch frequency. If the notch is lower than the intended frequency by more than 100khz, then try chopping a bit off of the shorted end. Be careful while chopping, just go bit by bit since going much higher than the intended frequency will cause you to start all over again. It is easy to tune the notch higher in frequency but difficult to lower the notch. You should be able to put the notch almost dead on frequency by trimming the gimick capacitor. The gimick capacitor should still be close to 3" in length to keep a nice shape factor to the notch. try to not deviate from the 3" length by more than 3/8" of an inch. Keep a log of the length of the final stub and the gimick capacitor, this will allow you to make 2 more exactly the same as the first prototype. Keep a log of the notch attenuation and the pass frequency loss and compare it to the final 2 stubs, they should be pretty close to each other. Any deviation should suggest a problem in the construction. ------------------------------------------------------------------------------ Shunt "reactance" - A shunt reactance is placed between the BNC to BNC connection (called the thru-line RF) and ground. This shunt reactance reduces reduces the insertion loss .5 MHZ away from about 1.3 dB to .4 dB *without* affecting the notch frequency or the attenuation at the notch frequency. To reduce the insertion loss *above* the notch's frequency a capacitor of 50 - 70 pF is needed. To reduce the insertion loss *below* the notch's frequency a 70 - 120 uH inductor is needed. A 56 pF silvered mica can be used for the capacitor. this can be done when the notch is completely built and tuned. Notch Frequency Trim Tab - changes the frequency 60 - 80 KHz is used to 'fine tune' the notch's frequency. This little square of metal is soldered to the outer jacket of the Heliax (ground) and moves the notch freq LOWER when pushed closer to the Thruline center conductor. 7//8", 1 1/4" and 1 5/8" Heliax were tried as band-reject stubs. 1 1/4" and 1 5/8" are most suitable. Individual stub attenuation values are measured to be 18 to 19 dB; when three 1 1/4" 19 dB stubs are connected with 1/4 electrical wavelength coax (approx 36") the attentation measured 73 db, an "apparent" attenuation of 23 db per stub! Insertion loss with a shunt reactances in place (explained above) measured less than .4 dB per stub.