RadCom April 2024, Vol. 100, No. 4

April 2024 61 Regulars requirement is less than 1mW, so the input combiner is less critical as the gains will compress. I decided to dig in to more detail at the effect of input power and phase imbalance at the output combiner, so did a simple AC analysis of the output Wilkinson using QUCS ( Figure 2 ). This allowed me a ‘quick-and- dirty’ way to investigate the effect of different power levels from each amplifier and, by adding a transmission line on one leg of the hybrid, to adjust the phase of one of the inputs. I ran this simulation for a number of combinations, looking at how much phase (length) error between the SMA leads from the PAs to the hybrid could be tolerated and what the effect of different output powers was. Now, a warning. Phase and power errors lead to reduction in output, this lost power being dissipated in the hybrid’s balancing or load resistor. In the extreme of one amplifier failing completely, half of the output power of the remaining amplifier (in my case 7.5W!) will be dissipated in the hybrid’s balance resistor. One of the issues with using a packaged and connectorised three-port Wilkinson is that the load resistor is internal and will usually be nothing more than a half- watt or less SMD resistor. On the other hand, with a four-port quadrature combiner you can connect an external high-power load to deal with this. When using packaged Wilkinsons, electronic measures must be taken to shut down the drive on amplifier failure or your expensive output combiner will be ‘toast’ in seconds! My simple modelling showed that at 10GHz with equal input powers, even a 1mm difference in physical length of the connecting cables will cause 350mW to be dissipated in the output combiner. So, you need to carefully phase match these cables with a VNA. The same, of course, applied to the input combiner leads. Small input power differences to the hybrid seem far more forgiving and, with perfect phase matching, and one of the 15W amplifiers turned down to 10W (nearly -2dB), only a quarter of a watt or so is dissipated. My final combined amplifier pair with one amplifier at 16W, and the other at 14W with hand-cut SMA leads, produced 27W out with no ill effects to the hybrid – but I’m still working on the electronics for the power fail circuits! Repeater and beacon news The Ofcombeaconmonitoring studymentioned in this column late last year is having positive effects, with two more GHz beacons recently completed and running. The new Bedford 24GHz beacon GB3BED on 24048.830MHz is now at its full height of 30m AGL in IO92SD. It doesn’t seem to have made much of a difference over the obstructed path to my QTH but, as reported earlier, it pops up now and again given the ‘right sort of rain’. It now has a 10GHz output. It’s on 10368.7783MHz when I checked it last and sends a Morse ID. On 8 February it was heard by G1EHF in IO91LH08 for an ODX of 101km. Finally, a new 1296MHz beacon, GB3IDT, is now operational on 1296.840MHz from IO92SA, just to the east of Ampthill. It runs 3W to a panel antenna facing east. G3XDY reports that the rebuilt GB3MHZ 10 and 2.3GHz beacons are in his shack on test and his plans to reinstall them are well underway. Hopefully by the time this magazine hits your doormat they will be back on the Martlesham tower. Finally, some rather more disappointing news is that the 10 and 24GHz GB3CAM beacons will have to move site before the end of March. The contract offered by the new commercial owners of the water tower at Wyton was just not acceptable or affordable for the beacon keeper Bernie, G4HJW. Bernie and I fund and maintain the beacons ourselves, so the decision has been made to remove them. We have had a good run of some 15 years on the site at low cost while it was owned by G4AJE. Thanks again to Paul. Other site possibilities are in the pipeline so I will write more about any possible move when it happens. Finally I’m seeing the #nohumans hashtag appearing more and more in relation to GHz bands activity. Let’s buck this worrying trend by all trying to do better (me included!) to make some QSOs instead of just spotting beacons. An area where I’d like to see much more activity is on digimodes above 1296MHz. FT8 is a non-starter, but there are lots of other far more suitable modes available. A full list of GHz bands events can be found by following the link at [5] to the UKuG web pages. If you have any comments on this, or have anything else to add to the GHz discussion, email me or message me via X, the application formerly known as Twitter, to @g4bao and @ukghz using the hashtag #GHz_bands. References [1] F5LEN tropo: http://tropo.f5len.org/forecasts-for-europe/ [2] Power combining: https://www.microwaves101.com/encyclopedias/ power-combining [3] Wilkinson Hybrid: https://www.microwaves101.com/encyclopedias/ wilkinson-power-splitters [4] Quadrature couplers: https://www.microwaves101.com/encyclopedias/ quadrature-couplers [5] Events 2024: https://www.microwavers.org/events.htm John C Worsnop PhD CEng MIET, G4BAO john@g4bao.com FIGURE 2: Simple QUCS model for Wilkinson hybrid.