23cm QRO coaxial hybrid combiner

As an exercise in milling and turning, I decided to have a bash at making a 0 degree six-quarter-wave 3dB ring hybrid to use with a pair of DF9IC 250W 1.3GHz PA modules. It was huge fun, although if I charged my time I think it would be scarily expensive! The finished thing is about 145mm diameter, with three N sockets for  the two PAs and a dump port, and a 7-16 DIN for the output port.

Starting from a piece of 150 x 25 aluminium bar, I turned a giant Polo mint, then used a rotary table on the mill to remove a 17.3mm deep by 19.4mm wide channel.  According to a simulation using ATLC2, that would give 70.7 ohm characteristic impedance (50*sqrt(2)) with a 6mm round centre conductor.

Made up a jig from an aluminium plate so the outside diameter matches the required ID of the ring.  Brass plug turned to match ID of tube and bolted to the edge of the disk.


The ring was soldered using a blowtorch, although I think the 75W iron would have done it OK.

During the build, I snapped off the 7mm pin from the 7-16 DIN socket, so had to drill the pin and the back of the socket and insert a 2mm copper tube, which was then soldered in place.

I milled flats at 60 degree intervals and drilled 4mm clearance holes for the centre PTFE sleeves of the N sockets, and tapped 3mm mounting holes. I made 4mm tapped holes and 16mm clearance plus a stepped indent for the 7-16 DIN socket.

Had to shorten the cap screws by screwing them into a tapped plate, bandsawing them off and touching them on the grinder.

To ensure the centre conductor was exactly in the correct position for drilling and soldering, I made up some spacers, which were pushed in to the groove.  I set up some springy wire to ensure the tube was held in tension and then clamped the spacers with some 2 inch Kant Twist clamps.

Spacer detail – 19.4 mm wide, 6.05 mm slot, 5mm thick, 18.5mm high to fit 17.3mm groove.

Spacer in place, ready to clamp

The 2mm tubes were soldered to the each of the sockets and cut to length.  The tube was drilled through the socket centre hole. The 2mm tubes were pushed through and cut to length, then soldered to the ring.

Should really have filed the solder flat.

I tapped 5mm mounting holes underneath at 135mm spacing.

Top view without the lid.

and with the lid bolted in place

Firs measurements show the balance is OK to 0.3dB from 1240 to 1370MHz. With 15dBm in, I got 12.15 from one port and 11.85 from the other, with -15.6dBm from the dump port

With all three Ns terminated in 50 ohms and +15dBm on the DIN, I saw -35.6dB at the dummy load port at 1296MHz and 26.4dB at 1306MHz.

Input return loss to the DIN port was worse than I expected, at 17.3dB from 1240 to 1370MHz. I need to investigate that more closely.

Removing one of the 50 ohm loads, the level at the dump port rose to +10.22dBm

I have not tested the input RL at the two 0-degree ports yet.

The calculated impedance of the coax cavity is around 70.8 ohms according to ATLC2

Leave a Reply

Your email address will not be published. Required fields are marked *