I have been trying to design a cheap home-made SWR meter, for a Radio Rampage project. It is not working out, so this is a call for help to any readers who have knowledge of these things. I have tried several different iterations of the heart of the SWR meter, the directional coupler.
The problem I have been having is that the forward and reverse power signals are not accurately representing forward and reverse power. In particular, reverse-coupled energy is much higher than actual reverse power.
The project started with analysis of my Revex UHF SWR meter connected to my antenna, driven by 4 watts from an Icom IC-400PRO. Indicated SWR was 1.1:1. The meter's directional coupler output was 634mV from the forward port, and 53mV from the reverse port. Values were as expected.
The first iteration (no picture available) was an aluminium tube 100mm long with SO-239 connectors at each end and brass rod centre conductor. Impedance was correctly calculated as per D/d. Sense loops were 1mm copper wire spaced equidistant between centre conductor and inner wall, with air dielectric. Rectification was 1N5711 with 1nF, 50R resistor at the opposite end. Components transverse on the opposite side as per directional coupler convention.
Iteration 1 was connected to the same antenna and radio. Output from the forward port was around 900mV, output from the reverse port was unexpected at 350mV. After verifying mathematics and construction, I decided that the design was faulty. Leading me to create a different design.
I have seen many directional couplers incorporated into radios, in order to sense bad SWR, so my next iteration was to copy that design. The second iteration (no picture available) was a stripline directional coupler made on a piece of double-sided PCB, with SMA connectors at each end. Source of mathematics I no longer recall, but rectification was the same as iteration 1, but this time using 5GHz diodes. Results were nearly the same as iteration 1, output voltage from the forward port was 414mV (shorter coupler), and voltage from the reverse port was 198mV. Once again, reverse voltage was much higher than expected, almost half the forward voltage. |
 | The third iteration was a design from an amateur radio magazine. The design involved shaving the two sides of a piece of rigid coax to expose the dielectric, and soldering two more pieces of coax similarly shaved against the sides of the first.
Rectification was was with HSMP3892 SHF diodes with 1nF filtering, and 50R resistor at the other end. Transverse on the opposite side.
Once again, forward port output was a reasonable value at 864mV, and reverse port output was unexpected at 316mV. Again almost half the forward port output.
I decided that the design, or my construction, was faulty. In any event, this design was too hard for the home constructor. |
 | The fourth design came from an amateur radio website. The design involves threading a wire under the braid of a piece of RG-213 coax. It seemed simple enough for the home constructor. The first version of this iteration was 120mm long, with the sense loops about 80mm long.
It was also a failure in that the reverse port output was about 40% to 45% of forward port output.
I consulted an RF engineer who designs mobile telephone base stations for a living, but as directional couplers are not commonly used in this industry, he could not help me except to ask - 'aren't they supposed to be a quarter-wavelength long?'. I am sure that is not required, but I rebuilt it a quarter-wavelength long, as shown in this picture, just to see if it did matter. Unfortunately I was not thinking, and made it a physical quarter-wavelength long, instead of an electrcal quarter-wavelength, so the port output voltages are are transposed. It didn't matter anyway, the voltages were as per previous iterations. |
 | I then noticed that the website version of this iteration was enclosed in a metal box, so I again rebuilt it, into a cast aluminium box as shown in this picture.
A trimcap, 2-40pF, was added from the sense loop output to ground, before the rectifier, as I have seen in some commercial SWR meters. It did not make any significant difference to the results.
This iteration has a forward port output of 413mV, and a reverse port output of 181mV.
With the Revex meter in the transmission line with the experimental directional coupler, the Revex reads slightly higher at 1.15:1 as opposed to 1.1:1 without the experimental directional coupler in line. This is as expected, and indicates that the directional coupler does not have a serious impedance anomaly. |
I have no idea why I am getting such a high reverse power reading. If you know this stuff and you know what is wrong, please contact me via the contact page, it will be much appreciated.
Please, no 'have you tried this' suggestions, I will ignore them. Either you know what's wrong or you don't. Thanks. |
|
