CCS, ICAS, PEP and Duty Cycle: How RF Power Ratings Should Be Stated

CCS and ICAS are useful service labels, but neither is a complete engineering specification. PEP describes a transmitter waveform; it does not tell you the transformer’s mean loss or temperature. A reproducible HF balun, UNUN or transformer rating must define the electrical stress, thermal cycle and evidence behind the number.

Start With the Power Quantity

The current ITU Radio Regulations distinguish peak-envelope, mean and carrier power. PEP is the average power delivered to the antenna line during one RF cycle at the crest of the modulation envelope. Mean power is averaged over an interval long compared with the lowest modulation frequency. Carrier power applies with no modulation.

Those definitions describe the transmitter output. A passive component still needs its own operating and test conditions.

What CCS and ICAS Really Add

In amateur and legacy commercial literature, CCS is commonly expanded as “Continuous Commercial Service,” while ICAS is commonly expanded as “Intermittent Commercial and Amateur Service.” The words signal an intended service pattern. They do not, by themselves, define:

• carrier, PEP or waveform-mean power;
• maximum key-down time and minimum recovery time;
• frequency range and transformation ratio;
• ambient temperature, enclosure, airflow or mounting;
• load impedance, SWR magnitude and reflection phase;
• allowable temperature rise, insertion loss, distortion or permanent shift; or
• whether the test ended at a convenient time or reached thermal equilibrium.

That is why two products both labelled “5 kW ICAS” need not have comparable capability. The acronym is not a standardized waveform or a substitute for a duty-cycle table.

Duty Cycle Has Two Layers

First comes the envelope factor: speech, keying or data determines mean power while the transmitter is keyed. Second comes the operating factor: the station alternates between transmit and receive. For an illustrative stable system, the interval-average RF power may be written:

Component temperature does not follow that equation instantly. It follows the time integral of the component’s loss, filtered by thermal mass and thermal resistance. A short voice peak may be electrically severe yet thermally insignificant; a long digital transmission can reach a much higher temperature at the same PEP. Conversely, insulation flashover and connector arcing are peak-voltage problems, not average-temperature problems.

Published amateur exposure tables sometimes use representative duty factors—for example different values for unprocessed SSB, processed speech, conversational CW and a continuous carrier. Those are useful screening assumptions, not universal component derating multipliers. Measure the waveform when the rating matters.

CESSB does not create one universal conversion

Hershberger’s CESSB work demonstrated that controlling envelope overshoot can increase average speech power at the same PEP. Reported increases depended on the speech sample and reference system. That supports a simple conclusion: processed SSB can run hotter than unprocessed conversational SSB. It does not support a fixed “PEP = 2× ICAS” or “PEP = 3× ICAS” rule for every voice, processor and operating pattern.

The Magnetic Physics: Voltage and Current Set Different Limits

For a conventional transformer under sinusoidal excitation, a useful first-order relationship is:

Flux density B is driven primarily by winding voltage, frequency, turns and effective core area. Field strength H is driven by ampere-turns. Power enters only through the circuit relationship between voltage, current and load. A 4:1 impedance ratio does not by itself specify magnetizing current, core flux or loss.

Transmission-line transformers require extra care. The full 50 Ω line voltage is not necessarily applied as magnetizing voltage to every core. It depends on Guanella, Ruthroff or autotransformer connection, winding placement, parasitic capacitance, imbalance and which mode excites the core. The design must be analysed as the circuit it actually is.

Fair-Rite’s own flux and field calculators keep voltage-driven flux and current-driven field as separate calculations. Its power-handling calculator combines a chosen core-loss-density threshold with winding current density to estimate apparent power—and explicitly warns that those thresholds and available material data may be too conservative or too aggressive for a particular application.

Overheating can occur far below saturation

Saturation is not the only failure mechanism, and ferrite loss does not simply “rise exponentially” at one universal boundary. TDK’s ferrite definitions state that core loss depends on material, core shape, temperature, frequency and flux density; winding loss must then be added, including skin and proximity effects. Dielectric heating, connector loss, solder joints and circulating common-mode current add other paths.

A design can therefore exceed its allowed temperature rise without approaching the material’s quoted saturation flux density. It can also survive thermally yet fail electrically from insulation stress or arcing.

What Core Stacking Changes

Stacking identical toroids increases effective cross-sectional area. With the same turns and applied winding voltage, the average flux density in an ideal stack is reduced. The additional material also changes total loss and transient thermal capacity.

But power handling does not automatically scale one-for-one with core count. Stacking also changes winding length, leakage, capacitance, window fill and cooling. Contact surfaces between cores do not cool like exposed outer surfaces. The finished assembly—not a bare-core volume calculation—must pass the specified electrical and thermal test.

Matched-Port Numbers Show Why PEP Alone Is Incomplete

For a 50 Ω resistive load and a continuous sine wave, the port values are:

These are external matched-port values, not winding hot-spot values. A transformer may have higher internal voltage or current due to its ratio, topology, leakage inductance, capacitance and resonance. Under mismatch, SWR gives reflection magnitude but not reflection phase. Feedline length and load phase decide where voltage and current maxima occur and what impedance the transformer actually sees.

“Tested at 50 Ω” is not an illusion; it is a necessary reference condition. It is simply not the complete operating envelope.

The Minimum Reproducible Rating Statement

The measurement boundary matters. IEC 62044-3:2023 specifies high-excitation methods for measuring power loss and amplitude permeability of soft-magnetic cores. The newer IEC/IEEE 61007-389:2026 catalogues transformer tests including winding resistance, excitation apparent power, insertion and return loss, dielectric withstand, temperature measurement and related information to be stated. A ham product need not claim certification to those standards, but the measurement categories show how much information a bare watt number leaves out.

A Practical Qualification Sequence

1. Identify every part. Record core manufacturer, material, part number, turns, conductor, insulation, connector, enclosure and topology.
2. Measure cold behaviour. Record full relevant S-parameters or impedance, insertion loss, return loss, balance/common-mode behaviour and winding resistance at defined reference planes.
3. Define the stress. State frequency, waveform, PEP/mean/forward power, key-down time, transmit cycle, load impedance and ambient.
4. Instrument the hot spots. Measure core, winding, joint and connector temperatures. Treat infrared readings carefully because emissivity and hidden winding hot spots can mislead.
5. Test mismatch as impedance, not one SWR number. Exercise representative high-voltage and high-current reflection phases or a documented complex-load envelope.
6. Reach the stated endpoint. For a continuous rating, show thermal equilibrium. For intermittent service, repeat the declared cycle and include a hot restart.
7. Re-measure hot and after cooling. Check insertion loss, impedance, balance, insulation and permanent drift.
8. Publish the limit. State what failed first and retain margin for production tolerance, installation and ambient variation.

Engineering Takeaways

• PEP, mean power, carrier power and component dissipation are different quantities.
• CCS and ICAS do not define a universal test waveform or duty cycle.
• There is no universal CW-to-ICAS or SSB-PEP-to-ICAS multiplier.
• Voltage, frequency, turns and core area set flux; current and turns set magnetic field and winding stress.
• Core, copper, dielectric, connector and common-mode losses must all be considered.
• Saturation is only one limit; overheating, flashover and parasitic resonance may occur first.
• Stacking cores can reduce flux density, but power does not automatically scale with core count.
• A 50 Ω test is a reference point, not proof of the complete real-antenna envelope.
• Legal station power and component survival are separate questions.
• A complete rating is the watt figure plus its waveform, frequency, load, time, cooling, acceptance limits and evidence.

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