Repeater Desense at a Shared 70 cm Site

When two 70 cm repeaters share a site and one transmitter operates only 125–500 kHz from the other repeater’s receive frequency, receiver desensitisation is a serious risk. Antenna separation helps, but it does not replace proper receive filtering and transmitter clean-up.

Receiver desensitisation—usually shortened to desense—is a loss of receiver sensitivity caused by a strong nearby signal. The unwanted signal does not need to open the squelch or fall exactly on the receive frequency. It can overload the receiver front end, compress its gain, raise the effective noise floor or create intermodulation products.

At a shared site, the practical question is not whether both repeaters work individually. It is whether each receiver can still hear a weak user while the neighbouring 50 W transmitter is keyed.

Scope: the numerical example uses 435 MHz, a 50 W transmitter and two vertically polarised 3 dBd omnidirectional antennas. The isolation figures are planning estimates. The real result depends on the complete antenna patterns, antenna dimensions, mast, feeder routing, shielding, reflections and surrounding structures.

What Desense Actually Is

A receiver may have excellent sensitivity when tested on its own and still perform badly at a busy radio site. A strong off-frequency signal can drive the receiver’s first amplifier or mixer towards compression. The receiver then loses gain, produces additional noise or creates mixing products, making a weak wanted signal difficult or impossible to hear.

The symptoms can be misleading:

• the repeater works normally until the neighbouring transmitter keys;
• weak users disappear while strong local users still get through;
• the receiver noise floor rises without an obvious signal opening the squelch;
• coverage appears to shrink only during simultaneous operation; or
• interference appears on frequencies where no transmitter is intentionally operating.

A conventional sensitivity test with every other transmitter switched off will not reveal this problem. Desense must be tested while the potential interfering transmitter is operating.

Why Two Duplexers Can Still Overlap

A repeater duplexer is designed primarily to separate that repeater’s own transmit and receive frequencies while allowing both paths to share one antenna. Its receive branch passes the wanted receive frequency and rejects the repeater’s associated transmit frequency.

That does not make the receive branch a narrow, brick-wall channel filter. When another repeater transmitter is only 125–500 kHz from the receive frequency, it may fall inside or close to the receive branch’s passband and receive little useful attenuation.

In other words, both duplexers can be correctly tuned for their own repeaters and the combined site can still suffer desense. Each duplexer is solving its own TX-to-RX offset; neither is automatically protecting the receiver against every neighbouring transmitter.

Why Vertical Separation Is So Effective

Vertically polarised omnidirectional antennas radiate mainly towards the horizon. Directly above and below the antenna are low-response regions in the vertical radiation pattern. Placing one antenna directly above the other can therefore combine free-space loss with pattern discrimination.

At 435 MHz, the wavelength is approximately:

Using the commonly applied vertical-isolation estimate for vertically aligned antennas:

The antennas should be installed genuinely one above the other, not merely on opposite sides of the mast with a small height difference. Their active elements should not overlap. With physically long collinear antennas, the distance between the nearest active sections matters as well as the nominal distance between antenna centres.

The calculated 62 dB is an idealised estimate. Tower steel, mounting brackets, poorly routed feeders, nearby antennas and surrounding structures can scatter RF into the theoretical vertical null. Actual antenna-to-antenna isolation must therefore be measured after installation.

What 62 dB Means with a 50 W Transmitter

A 50 W transmitter produces approximately +47 dBm at its output connector. If the antenna system provides 62 dB of isolation, the approximate coupled level is:

Feeder loss, duplexer loss and other passive losses will change the final number, but −15 dBm is still an extremely strong signal at a receiver input. A weak wanted mobile may arrive around −110 to −120 dBm. The receiver may therefore be asked to recover a weak signal while surviving an unwanted signal roughly 95–105 dB stronger.

This is why five metres of vertical separation should be viewed as valuable isolation—not as proof that no additional filtering is required.

Why Horizontal Separation Needs Hundreds of Metres

With horizontal separation, two omnidirectional antennas see each other through their strongest azimuth directions. The antenna gain therefore works against the required isolation.

The free-space estimate is:

A gain of 3 dBd is approximately 5.15 dBi. Using 5.15 dBi for each antenna and solving for the same theoretical 62 dB provided by 5 m of vertical separation gives:

The answer is not five, ten or twenty metres. Under the stated assumptions, obtaining comparable isolation horizontally requires roughly 230–240 m of clear separation.

This comparison explains why vertical stacking is normally the practical choice on a shared mast. The vertical pattern provides discrimination that horizontal distance alone must replace with a very long propagation path.

Arrangement	Approximate separation	Planning isolation	Main limitation
Vertical stack	5 m	About 62 dB	Depends on a real vertical-pattern null; reflections and the mast can reduce it.
Horizontal, two 3 dBd antennas	About 235 m	About 62 dB	Both antennas point their high-gain azimuth patterns towards each other.

Why Separating Two Receive Antennas Is Not the Fix

Receive antennas do not desensitise each other merely because they are receiving. Moving two receive antennas apart does not create narrow frequency selectivity. Unless site geometry or directional patterns provide significant discrimination, both antennas will still collect the nearby 50 W signal.

The useful separation is between the interfering transmitter antenna and the victim receiver antenna. Even then, separation supplies only part of the total isolation budget. It does not replace the filtering needed ahead of the receiver’s active circuitry.

Protection against a transmitter only 125–500 kHz away normally requires a serious high-Q preselector, designed around measured signal levels and actual equipment performance.

What a Serious Receive Preselector Looks Like

The preselector must attenuate the neighbouring transmitter before that signal reaches the receiver’s first amplifier, preamplifier or mixer. Depending on the required rejection and acceptable insertion loss, the system may need:

• one or more large high-Q cavity band-pass filters;
• a dedicated notch cavity tuned to the neighbouring transmitter;
• a multi-cavity receive window filter;
• low-loss, well-shielded interconnecting cables;
• careful physical separation between transmitter and receiver equipment;
• proper bonding and elimination of loose or corroded metal junctions; and
• alignment with a network analyser or suitable service monitor.

The preselector belongs ahead of the first active stage. Placing a selective filter after an overloaded preamplifier or mixer is too late; the gain compression and intermodulation have already occurred.

At 125 kHz separation, the loss-versus-rejection trade-off becomes severe. A small helical filter or ordinary mobile duplexer may not provide enough close-in selectivity. Several large cavities or a specialised filter may be required, and every decibel of receive insertion loss directly reduces sensitivity unless the complete system is redesigned around it.

The Transmitting Side Must Also Be Clean

Not every desense problem can be fixed at the victim receiver. The interfering transmitter may produce broadband noise at the victim receive frequency, close-in phase noise, harmonics or intermodulation products.

A receive filter cannot distinguish unwanted transmitter noise from a weak user when both occupy the same frequency. That energy must be reduced at the transmitter with appropriate cavity filtering, isolators or circulators, a properly engineered combiner, clean power amplifiers and good shielding.

Passive intermodulation must also be considered. Strong RF currents flowing through rusty brackets, loose connectors, poor bonds or corroded tower hardware can create new signals. No amount of receiver selectivity will remove an intermodulation product that lands directly on the wanted channel.

Measure the Complete Site

A shared-site design should not be accepted from calculations alone. At minimum, verify:

1. Antenna-to-antenna isolation. Measure the installed paths in every relevant direction.
2. Receiver desense. Record sensitivity with the neighbouring transmitter off and again while it transmits at full power.
3. Transmitter noise. Measure unwanted energy at each victim receive frequency.
4. Filter response. Measure insertion loss and rejection with the actual cables and terminations.
5. Intermodulation. Test combinations of every transmitter that can operate simultaneously.
6. Long-term stability. Allow margin for temperature, ageing, vibration and accidental mistuning.

In Summary

Two 70 cm repeaters with only 125–500 kHz between one transmitter and the other receiver can interfere even when both duplexers are correctly tuned.

For vertically polarised antennas around 435 MHz, a 5 m vertical stack can theoretically provide about 62 dB isolation because each antenna is placed near the other’s vertical pattern null. With two 3 dBd antennas, obtaining comparable isolation horizontally requires roughly 235 m.

Even 62 dB leaves the coupled level from a 50 W transmitter near −15 dBm before other losses—still enormous compared with a weak wanted signal. Antenna spacing is therefore only one part of the solution.

The receive side needs a properly designed high-Q preselector ahead of its first active stage. The transmitting side may also need clean-up filtering, isolation and intermodulation control. Finally, the complete system must be measured with all transmitters operating.

Good repeater performance comes from an isolation budget, not from one duplexer specification or one mast-spacing rule.

Want more technical RF content? Subscribe for new deep dives and lab notes.

Have a question or field observation? Contact RF.Guru.