Phase cancellation to eliminate interference
by Fred W Daniel

Phase cancellation can be used to eliminate on-channel noise and interference, that is impossible any other way. It can also be used to reduce or eliminate noise or receiver overload, caused by a nearby transmitter, that notch filters are ineffective in controlling. There are many repeater sites that are common to high-power FM and TV broadcasters. Because of this close proximity, enormous amounts of broadband noise is created by the high-power transmitters, that radiates beyond the assigned frequency of the broadcaster. This is common to all transmitters, but when the transmitter operates at 5,000 or 50,000 watts, off-channel noise can be as high as 1 to 10 milliwatts. This does not seem like much, but because this noise source is very close to you, it can impair your receiver because it is broadband and actually on your channel.
In the Los Angeles area, Mt. Wilson is where many FM stations and almost all TV stations are located. This site for years has been difficult, at best, to operate a repeater from. With the proper selection of antennas, and other hardware, a near perfect receiver can operate, with the aid of phase cancellation. This technique cannot solve impossible problems, where more than 50 dB of noise reduction is required. However, this technique has been used for decades by cable TV head-ends, to eliminate interference from co-channel TV stations. Microwave Filter Co. has build the model 2903 for many years for the cable industry, but hardly anyone in the Land Mobile market is familiar with this technique.
There are some issues to remember, while designing, testing, and maintaining phase cancellation systems.

- Being able to directly sample a small signal from the offending transmitter is more reliable than off-the-air sampling of the signal. The null can be 10 to 20 dB deeper, and can be maintained longer, without alignment.
- Matching the phase angle is not as critical as the matching the amplitude. For a notch of more than 25 dB, the amplitude must be within 0.1 dB, so stable components are a necessity.
- The null is fairly broadband, but there is several dB roll-off, on either side of the center frequency, for several MHz.
- For alignment, an older analog spectrum analyzer may work better than a new $50,000 digital spectrum analyzer. The new digital analyzers don't show noise very well.

Be sure to test, and learn, on your bench FIRST. Use one or two signal generators and a analog FM receiver, with some cables, an variable attenuator, line stretcher, and spectrum analyzer. RF carriers are easy to learn with, but noise is a bit harder to see on a receiver or spectrum analyzer, and maintain alignment.
In the Los Angeles area, there are many repeaters on T-Band channels 14 and 21. A few years ago, a Tijuana, Mexico channel 21 TV station went on the air, about 78 miles from the nearest LA channel 21 repeater. This killed most channel 21 repeaters on hilltop sites, as the HDTV signal was almost 6 MHz wide, and covered all of the repeater receive band. Many repeater sites eventually installed phase cancellation systems to reduce the impact, We simply used a 8 element beam as the sample antenna, and coupled it into our main receive antenna with a small variable attenuator. We did not have a line stretcher at the time, so we used a variety of coax lengths to get the phase in the general range. We were forced to use about 6 right angle adapters in series to get the phase just right. We got a reasonably stable 32 dB rejection.  That saved us a second trip to the mountain.  Eventually, the Mexican station moved to another channel and we removed all the excess hardware.

To get started, you will need something to vary the phase of the null signal at least 180 degrees. The preferred device is called a "line stretcher," that is nothing more than a 50 ohm piece of transmission line that you can vary the length of. See photos for examples. Depending upon the frequency, other passive devices will work, provided the amplitude does not change while adjusting the phase. If you have a phase shifter that will not shift 180 degrees at your frequency, then you can create short Male-Female coax jumpers that are 1/4, 1/2 and 3/4 wavelength, with the cable velocity factor calculated into the length.  For example, a half-wavelength at 460 MHz. is about 12 inches. If you are using RG-213/U, the Velocity Factor [VF] is 0.66 so you would multiply the length by the VF, or 12 x 0.66 =  7.9 inches.  The overall measurement varies slightly, depending upon the connectors used. For example, Type-N connectors add slightly more to the overall length than UHF connectors like a PL-259, due to the air dialectric inside the connector.  Typically, UHF connectors work poorly above 150 MHz because the impedance is about 35 Ohms, instead of 50 Ohms. 
Next, you will need a variety of fixed attenuators such as 3-6-10-20-30 dB, plus a 3 to 10 dB multi-turn variable attenuator, to be able to adjust down to 0.1 dB. Examples are  shown in nearby photos. Some technicians prefer to use a expensive step/variable attenuator to test with, then assemble the required attenuator with mostly fixed elements. Last you will need an assortment of 10, 15, and 20 dB couplers, to use to inject the null into the receiver system, with the lowest main-line loss. Remember, replacing the attenuators or any element could change the phase, due to the length of the coax changing.  Sometimes an amplifier is necessary in the sample line to achieve a great enough signal to null properly.  Avoid using an amplifier if possible because it is hard to maintain the gain within 0.1 dB, even with a regulated supply.  

Last, you may experience difficulty measuring the level of noise affecting your receiver.  Sometimes it works best using the RSSI output on a standard FM receiver. Other people finish off their testing using the standard Motorola Desense Test setup with a IsoTee or coupler on the test receiver and signal generator.      

Off-Site Links of Similar interest [which may be copyrighted]:

50 MHz Receiver Noise Canceling Front-end
Phase Shift Cancellation of Unwanted VHF RF

Revised April 25, 2018