Having spent over 40 hours getting the D-star 2M duplexer ready has been a most interesting exercise.
The 6″ cavity filters we have had to be modified to allow them to work as notch filters and is bassed on the original design (4″ filters) by W1GAN JOHN J BILODEAU form an article in QST July 1972. What worked for the 4″ cavities had to be fiddled with on the 6″ ones.
The filters have been finished and now I’m fine tunning the coax lengths that connect the filters together. This part of the design was never covered well in the original or later articles. The lengths were given but not the reasons for the lengths. The lengths as given work but not as well as should be achievable.
The tracking spectrum analyser that WARG bought a while back has been fantastic in converting the 6″ cavities. Without it I could have tuned the duplexer up but would not have been able to do it easily and would not have been able to get results that can be achieved with the analyser.
The analyser shows you not just pass and notch but the type of curves. This has shown the interaction with the cable lengths that give various shapes, various peaks and dips that indicated the overall design is not working as well as it can and should.
With out the analyser the pass loss can be as high as 2.5 dB while with the analyster it can be as low as 1.5dB. Okay only a dB but it all adds up. The overall notch depth can not be seen on the analyser as it will not show 70dB plus notch depth. Each cavity can be seen okay but what happens when they are all connected together?
I can work this out using the analyser’s sig gen and a receiver and will do so next.
The attached photos show the PCB (painted) boxes on top and the difference with the pass frequency when you connect two cavities together with incorrect cable lengths. Each cavity has a good pandpass but when connected cause the bouble hump….Working on it.