12.5 kHz Repeater Spacing: Coordination Beats Cavities

Put two elevated 70 cm repeaters on adjacent 12.5 kHz channels and the customary cure is “add better filters.” That is incomplete. Ordinary external UHF cavities can improve transmitter cleanliness and protect receiver front ends, but they cannot provide arbitrarily large rejection 12.5 kHz from the wanted carrier without also narrowing, delaying or attenuating the wanted analogue-FM channel. The scalable solution begins with coordination, path engineering and measured equipment performance.

Spacing Is Not Shift—and a Band Plan Is Not a Licence

Three different ideas are often mixed together:

• Channel spacing is the separation between assigned channel centres—12.5 kHz in this example.
• Duplex shift or offset is the megahertz-scale difference between one repeater’s input and output.
• Coordination separation is whatever combination of frequency, geography, antenna discrimination and equipment performance is needed to prevent unacceptable degradation.

The IARU Region 1 VHF Handbook 10.02 showed 12.5 kHz-spaced repeater structures in parts of 430–440 MHz. IARU announced Handbook 10.03 in February 2026. These are regional band-planning recommendations; actual authorisation, frequency availability and coordination remain national matters. Always check the current national band plan, licence conditions and coordinator data before changing a repeater.

A legitimate channel raster does not promise that two adjacent high-site systems will coexist at arbitrary locations. The engineering question is the wanted and unwanted signal level at each victim receiver and throughout each service area, including exceptional propagation.

Why 50 km Is a Warning, Not a Law

Terrain can isolate sites separated by 15 km, while two clear elevated paths may interact far beyond 50 km. Tropospheric enhancement can turn an ordinarily quiet path into a strong one. Distance alone therefore cannot be the acceptance criterion.

Free-space loss is still a useful transparent reference:

Those figures assume an unobstructed far-field free-space path. They exclude terrain diffraction, clutter, reflections, atmospheric refraction and anomalous propagation. For coordination above 100 MHz, the in-force ITU-R P.452 procedure is a more appropriate engineering framework because it treats several long-term and anomalous propagation mechanisms and a chosen time percentage.

A 25 W transmitter is approximately +44 dBm at its connector. After feeder loss and antenna gain, an EIRP of +48 to +52 dBm is plausible for some installations, but it is not universal. With a clear 50 km path and receive-antenna gain, the adjacent carrier might arrive around −65 to −60 dBm. A weak wanted user could be near −115 dBm under a defined bandwidth and SINAD or quieting criterion. The resulting 50 dB difference is credible, but every term must be measured or justified for the actual system.

What the ETSI Numbers Really Say

ETSI EN 300 086 V2.1.2 is the current published harmonised standard in this equipment family for analogue-speech land-mobile radios from 30 MHz to 1 GHz. For 12.5 kHz channel separation, its defined laboratory conditions include:

• maximum permissible FM deviation of ±2.5 kHz;
• controlled response above the 2.55 kHz modulation-frequency point;
• adjacent-channel power no more than 60 dB below conducted transmitter power, without needing to fall below −37 dBm;
• receiver adjacent-channel selectivity of at least 60 dB under normal conditions; and
• receiver adjacent-channel selectivity of at least 50 dB under extreme conditions.

The adjacent-channel-power test integrates modulation, hum and noise through the standard’s measuring receiver. The selectivity test applies a defined wanted signal and a modulated unwanted signal one channel away, then finds the unwanted-to-wanted ratio at a specified degradation. These are not interchangeable with a phase-noise point in dBc/Hz, a wide-offset blocking figure or an anecdotal on-air test.

For a 25 W transmitter, a 60 dB-below-carrier adjacent-channel-power limit corresponds to as much as −16 dBm at the transmitter reference plane before the −37 dBm floor becomes relevant. Path loss then attenuates that unwanted adjacent-channel energy. Conversely, the victim receiver’s selectivity attenuates the neighbouring carrier only inside the receiver. Compliance on both sides does not automatically create comfortable coordination margin.

The Combined Adjacent-Channel Limit

It is useful to think of transmitter leakage and receiver selectivity as two finite contributors. In linear power ratios, an effective adjacent-channel interference ratio can be represented as:

If both quantities were exactly 60 dB, their combination would be approximately 57 dB—not 60 dB. This simplified relation does not replace the ETSI waveforms or a full receiver test, but it shows why quoting only the cleaner transmitter or only the sharper receiver overstates system protection.

Why a Duplexer Does Not Separate Adjacent Channels

A repeater duplexer lets one system transmit and receive through one antenna by separating its output and input across the duplex shift. Its resonators are designed for a megahertz-scale difference, acceptable transmit loss and acceptable receive loss.

An adjacent 12.5 kHz carrier is a different problem. If it lies inside the receive branch’s passband, a conventional duplexer may provide little rejection. Adding cavities can help with far-out signals, transmitter broadband noise, harmonics and front-end protection, but an external cavity system cannot become an infinitely steep wall between two adjacent narrowband-FM emissions.

Receiver IF, ceramic, crystal, SAW or digital channel filters can create much sharper channel selectivity because they operate after frequency conversion and at lower power. They are essential. They still have finite amplitude response, group delay, frequency tolerance and strong-signal headroom, and they cannot remove damage that occurred in an overloaded RF amplifier or mixer ahead of them.

What the Resonator-Q Example Does—and Does Not Prove

For one simple resonator, loaded Q and 3 dB bandwidth are related by:

A first-order resonator response can be approximated near resonance by:

In that deliberately simple model, 30 dB attenuation at 12.5 kHz needs Q_L of roughly 550,000, giving a 3 dB bandwidth below 1 kHz. Such a single RF resonator would badly constrain the wanted analogue-FM channel.

This is an illustration, not a universal filter-design limit. Multiple coupled resonators, transmission zeros and lower-frequency channel filters can produce much steeper skirts. Their response must be synthesized for the wanted passband, stopband, insertion loss and group delay; individual resonator Q is not obtained merely by dividing 435 MHz by the channel spacing. At UHF transmitter power, practical cavities also face size, conductor loss, tuning drift and voltage or heating limits.

A Notch Is Not Free

A notch placed on the neighbouring carrier may add useful rejection, especially at a receiver input. At a 12.5 kHz offset it must remain deep, narrow and stable without damaging wanted sidebands. A carrier notch also does not automatically remove:

• the neighbouring transmitter’s modulation energy inside the victim channel;
• close-in phase noise;
• receiver reciprocal mixing;
• front-end compression before the notch;
• intermodulation created in an earlier nonlinear stage; or
• over-deviated user transmitters on the repeater input channels.

A bench notch depth measured with a quiet synthesizer is therefore not the same as coexistence between two complete repeater systems.

Four Equipment Paths Must Work

The interference need not be reciprocal. Transmit powers, antenna patterns, feed losses, site elevations, receiver architectures and terrain may differ. Each direction therefore needs its own measurements. Mobile and handheld users also matter: an over-deviated user near one site can interfere on the input side even when both repeater transmitters are clean.

Why Brute-Force “Fixes” Can Shrink Both Systems

Compatibility can sometimes be forced by sacrificing performance. That may be a valid coordinated design choice, but it must be described honestly.

• Lower deviation and audio bandwidth reduce occupied spectrum but may harm intelligibility and user compatibility.
• Narrower IF filtering improves selectivity but increases sensitivity to frequency error and may add amplitude or group-delay distortion.
• Input attenuation or reduced RF gain can improve strong-signal behavior while reducing weak-user coverage.
• Lower ERP, downtilt or directional patterns reduce unwanted coverage and may intentionally reduce wanted coverage too.
• Lossy preselectors can improve adjacent rejection but add their insertion loss ahead of the receiver, degrading system noise figure by the same loss when other terms are unchanged.

If two regional repeaters survive only after both surrender deviation, receiver sensitivity, ERP and intended coverage, the frequency assignment has consumed the value of both installations. On the other hand, a smaller deliberately shaped service area can be excellent engineering when it meets the actual communication need.

CTCSS and DCS Are Access Controls, Not RF Filters

Different CTCSS tones or DCS codes can stop the wrong signal from opening a squelch or keying a repeater. They do not remove RF energy at the antenna connector. The unwanted carrier can still reduce gain, create reciprocal-mixing noise, drive intermodulation or mask a weak wanted user before the tone decoder makes any decision.

A Coordination Workflow That Produces Evidence

1. Define both service areas. State the wanted field strength or receiver-input level and the percentage of locations and time to be served.
2. Choose the propagation model. Use terrain, antenna height and pattern data; include a suitable low-time-percentage case for enhanced propagation. Keep free-space loss as a reference, not the sole prediction.
3. Build both directional input budgets. Calculate the wanted user level and adjacent repeater carrier level at each receiver without counting path or antenna isolation twice.
4. Measure transmitter behavior. Check frequency error, deviation, audio response, adjacent-channel power and close-in noise at normal temperature, power and modulation.
5. Measure receiver behavior. Establish the wanted criterion—such as a stated SINAD or quieting—then apply a modulated adjacent interferer and record degradation. Also test blocking and intermodulation separately.
6. Account for installed filtering. Measure duplexer, preselector, combiner, connector and feeder transfer at the relevant frequencies and reference planes.
7. Test users as well as repeaters. Include representative mobiles and handhelds, frequency error and deviation at both input-channel populations.
8. Run simultaneous field tests. Use weak wanted signals near planned coverage boundaries while the neighbouring repeater transmits at normal power and duty cycle.
9. Retain implementation margin. Allow for temperature, component aging, cavity drift, site changes and propagation variability. A zero-margin paper result is not a robust field design.

Keep the Accounting Clear

At each victim input, record these values separately:

• wanted signal level and criterion;
• adjacent carrier level;
• transmitter energy measured in the victim channel;
• receiver adjacent-channel selectivity at the relevant level;
• receiver blocking and intermodulation margins;
• installed filter loss and rejection; and
• environmental and implementation margin.

This prevents a common mistake: adding a site-path loss to an unwanted-to-wanted level difference that already includes that same path, or treating a dBc transmitter figure as if it were receiver selectivity.

What Filters Are Genuinely Good At

This is not an argument against cavities or selective networks. They are indispensable when the task provides a usable transition band. Proper filters can:

• separate one repeater’s transmit and receive paths across its duplex shift;
• suppress harmonics and far-out products;
• reduce transmitter broadband noise at the victim frequency;
• reject out-of-band broadcast, paging, mobile or link transmitters;
• protect receiver front ends at shared sites; and
• support multicoupler and combiner systems.

The key is to specify rejection at exact frequencies, insertion loss in the wanted path, group delay where it matters, temperature stability and power handling. “Six cavities” is a parts count, not a performance specification.

The Practical Verdict

12.5 kHz spacing on 70 cm is valid narrowband engineering. Fifty kilometres is neither automatically safe nor automatically unsafe. Two adjacent repeaters can coexist when geography, antenna patterns, controlled ERP, clean transmitters and strong receiver selectivity provide enough measured margin.

What ordinary external cavities cannot do is rescue an arbitrary overlapping assignment by creating unlimited rejection inside the tiny transition between adjacent analogue-FM channels. The channel filter, RF front end, transmitter spectrum and propagation path must operate as one coordinated system.

Questions, antenna-factor records or height trials to share? Contact RF.Guru.