Home » Radio » Radio Frequency Interference » X-Phase Performance on Shortwave

X-Phase Performance on Shortwave

x-phase performance

So, what’s the verdict on X-Phase performance for reducing RFI during shortwave listening? Pretty good, actually, for something so cheap and cheerful.

Total cost of my X-Phase QRM Eliminator comes to around $30. This includes the RA0SMS kit, shielded plastic box, switches and connectors. I spent about a day putting it all together. Let’s compare this to an ANC-4, which comes in at around $200, or MFJ-1026 at $240.

My main use case is reducing RFI from noisy neighbors, 100 to 500 feet distant. My main offender is poorly wired halogen lights in one nearby kitchen. Other sources include router birdies, VDSL emissions and many crappy switch-mode power supplies that permeate my environment. My antenna setup includes either noise probes with my main antenna or a pair of wideband active loops 100 feet apart.

My verdict on X-Phase performance is 70 to 80% compared to the ANC-4 at 15% the cost. It does really well on eliminating the halogen lights RFI, pretty well on the router birdies and mixed on the SMPS emissions in the HF range.

So, if you want to give the RA0SMS X-Phase a try, you should not be disappointed on a price-performance basis.

You will find various assembled units online at prices ranging from $70 to $120. (All amounts U.S. dollars.) My advice would be if you are going to spend more than $100 for an assembled X-Phase, consider just going for the ANC-4 if you can find one.

X-Phase Performance Nitty Gritty

As with most devices of this type, there is quite a bit of interaction between the three controls, particularly the AUX antenna gain and the Phase control. Generally I set the Main antenna gain first, then adjust the AUX and Phase to get a QRM null. Also, I find that the Phase control tends to function mostly near the CCW position regardless of how the Gain controls are set.

Switched off, insertion loss is about the same as the ANC-4 at 6 dB or so. Switched on and adjusted, RFI reduction can easily exceed 20 dB.

When using the X-Phase with a wideband loop, it’s easy to get intermodulation products from local AM broadcast stations if you set the Main Gain too high. You can eliminate this problem with an AM bandstop filter inserted ahead of the QRM eliminator.

You can easily test X-Phase performance by feeding a signal generator into both the Main and AUX inputs, and watching OUT performance on a scope. On my unit, I find performance under 3 MHz drops off quickly.



  1. Zaba says:

    Hello John!

    Thanks for the overview. The “nitty gritty” addresses well the points I mentioned in my reply on your RA0SMS-assembly report a few days back… Generally you can optimize only one interference source at one instant, but one can be lucky with the phase /amplitude of other sources at that moment… I will drill my DK9NL board and put it together. In addition I will try a couple of delays-lines with a switch deck, using thin 75 ohms coax…. Cheer/73, Zaba OH1ZAA

    • Zaba says:

      It’s been over 20 years that I have been drilling holes in a PCB board. Needed 0.8 mm for the components and 1.5 mm for the pots. I did not have 5 k pots, so for the 1 k version I will scale the phasing caps to 470 pF and the AUX transformer ratio 1 : 3 (50 : 450 ohms)… I checked the
      ANC-4 and MFJ-1406 schematics, and it seems that the independence of phasing and amplitude is a bit restricted, while full the 360 deg scan is missing. In this one the two pots at the AUX-input can make a 180 deg reverse of phase, but phase adjustment range is still limited.

      The phase range can easily be doubled by the insertion at the AUX-port of a 0/90-degree hybrid combiner/splitter like the MiniCircuits JSPQ-65W+ (one port selected at one time, while the other terminated in 50 ohms)… But to make the X-phase a gadget in its own class, one should insert a 360 deg phase shifter like the MiniCircuits SCPHS-13.6+ (this one will operate beyond its specified range together with the existing circuit). The modifications to the original would be simple and minimal. Just one steering voltage for the phase-shift needed. This would make alignment much simpler with better independence of amplitude and phase settings … Cheers/73, “Zaba” OH1ZAA

  2. Zaba says:

    My DK9NL-version is close to testing. I doubled the in/out 1N4148 diodes as well. Initially I could not find the J310’s so I put a pair of selected BF245B’s. Those are pulling about 15 mA each. Potentiometers are all 1k and the phasing caps tiny 470 pF chips. To test my phase-expander idea, I took an Amidon T106-2 (red) toroid, with 2 x 15 turns of wire to construct a 90 deg hybrid coupler. It turned out a bit low in frequency (generator + oscilloscope test), as I demolished a smaller core while cleaning the wire ends (too much force). I will take some other small core. These hybrids are wide-band. It does not need to be precise with phase, as long as the offset is some 70 degrees at least on the other port. The MiniCircuits neatly packed couplers are quite expensive. I am still missing the AUX input-transformer, but then I will be ready to test…. famous last words … Cheers/73, Zaba OH1ZAA

  3. Zaba says:

    Using the oscilloscope as an indicator for amplitude I get the following initial results (without using any of the proposed enhancements, like coax delay line, 90 degrees hybrid coupler or an voltage controlled analog 360 deg phase shifter). First of all, without AUX signal connected the
    DK9NL board can be used as a preamplifier. There is quite a bit of gain with the input potentiometer maxed. The AUX-transformer is far from ideal (the ferrite is from a PC-motherboard CPU SPS-circuit), but when I jump the main antenna with a 75 ohms resistor to the AUX-input, then I can zero-out the input signal fully on the output with multiple iteration. Well, if this would be an interfering signal, then the principle works, at least for this in-phase feed on AUX. The real world can present any phase to AUX, so now I am going to test it with real antennas. I am sure I will see something usable, but I am pretty confident that the function
    can be made prefect with the suggestions from above. Mainly I trust the solution with the
    electronic delay line for the best versatility and ease of use … Cheers/73, Zaba OH1ZAA

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.