2.6 Transmission Quality: Frequency Drift, Parasitic Oscillation and Over-Modulation

2.6 Transmission Quality: Frequency Drift, Parasitic Oscillation and Over-Modulation

ACMA Exam Focus: Syllabus items 5.9, 5.11 and 5.12 - describe the causes and effects of frequency drift, explain what a parasitic oscillation is and how it is prevented, and describe over-modulation and the interference it causes to other spectrum users.

A transmitter can be on the correct frequency, at legal power, and still be an unacceptable signal. Three faults account for most on-air complaints: the signal wanders, it produces energy on frequencies unrelated to the wanted one, or it is driven too hard and splatters across its neighbours. Each has recognisable causes, effects and cures.

Frequency drift

Frequency drift is a slow, unintended change in the operating frequency of a transmitter or receiver after it has been set. It is distinct from chirp, which is a fast frequency shift during keying.

Causes

• Thermal. Every component has a temperature coefficient. As the PA and power supply warm the cabinet, the VFO's coil, capacitors and semiconductor junctions change value and the frequency moves. This is by far the most common cause, and it is worst in the first fifteen to thirty minutes after switch-on.
• Supply voltage variation. An unregulated or sagging rail changes the bias and junction capacitance in the oscillator. A battery falling from 13.8 V towards 11 V during a portable or SOTA activation will pull an inadequately regulated VFO.
• Mechanical. Vibration, a flexing chassis, thermal expansion of the coil former, or movement of the oscillator wiring alters inductance and capacitance. In mobile operation, road vibration is a real source of drift.
• Component ageing. Quartz crystals age, typically by a few parts per million in the first year, then more slowly.
• Load pulling. Without an adequate buffer stage, the changing load presented by a keyed or modulated PA reflects back and pulls the oscillator.

Effects

Drift is normally quoted in parts per million (ppm), and its practical impact scales with operating frequency:

A 142 Hz wander is barely noticeable on an SSB voice contact but is fatal for a PSK31 signal only 31 Hz wide, and will spoil an FT8 decode. On 2 m and 70 cm the same specification produces errors of kilohertz, enough to miss a repeater input or to fall outside a narrow CW filter mid-contact. In the worst case drift can carry a signal outside an authorised band edge, which breaches the amateur licence conditions.

Reducing drift

1. Allow the equipment to warm up for 15-30 minutes before critical work such as a VHF contest or a satellite pass.
2. Fit a TCXO (temperature-compensated crystal oscillator), which senses temperature and applies a correcting voltage, giving roughly 0.5-2 ppm.
3. Use an OCXO (oven-controlled crystal oscillator), which holds the crystal in a small thermostatically controlled oven at a fixed elevated temperature and can reach 0.001 ppm.
4. Lock the rig to a GPSDO (GPS-disciplined oscillator) for microwave, EME or beacon work, where the reference is steered by satellite timing.
5. Regulate the supply rail, keep the oscillator thermally isolated from the PA, build it rigidly, and always buffer it.

Parasitic oscillation

A parasitic oscillation is an unwanted self-oscillation occurring in an amplifier at a frequency unrelated to the wanted signal. It arises when stray lead inductance and stray capacitance accidentally form a resonant feedback path around a stage that has plenty of gain at that frequency.

The classic case is a VHF parasitic in an HF power amplifier: the inductance of the anode or drain lead resonates with device capacitance somewhere around 50-200 MHz, and the amplifier oscillates there while apparently working normally on 14 MHz. Low-frequency parasitics also occur, usually through shared power-supply impedance, producing a slow "motorboating" instability.

Symptoms to recognise:

• Unexplained gate, grid or drain current, or a PA that runs hot for no output.
• Output that does not track drive smoothly, or that jumps as the loading control is moved.
• Spurious emissions reported on frequencies with no harmonic relationship to your transmit frequency.
• Arcing, blown devices, or television interference appearing only on transmit.

Prevention and cure:

• Fit a parasitic suppressor - a low-value non-inductive resistor of about 50-100 ohms in parallel with a few turns of wire - in the anode or drain lead, so the circuit is damped at VHF but unaffected at HF.
• Keep RF leads short and direct, and use good grounding and layout so stray inductance is minimised.
• Neutralise the stage by feeding back a small out-of-phase signal that cancels the device's internal feedback capacitance.
• Bypass and decouple every supply lead properly, and screen or compartmentalise input from output.
• Ferrite beads on device leads damp VHF resonances cheaply.

Over-modulation

Over-modulation occurs when a transmitter is driven with more audio than it can handle linearly. For AM the modulation percentage is:

• m (%) = (E_max - E_min) / (E_max + E_min) x 100

With an envelope reaching 150 V peak and falling to 50 V, m = (150 - 50) / (150 + 50) x 100 = 50 per cent. At 100 per cent the envelope just touches zero. Beyond that the carrier is cut off for part of each cycle, the envelope is clipped, and the clipping generates harmonics of the audio.

The consequences are the same for AM and SSB:

• Splatter. An AM transmitter modulated with 3 kHz audio should occupy 6 kHz. Clipping creates audio harmonics at 6 kHz and 9 kHz, so the occupied bandwidth can exceed 18 kHz - covering several adjacent QSOs.
• Intermodulation distortion (IMD). Driving an SSB linear beyond its linear range makes the speech components mix with each other, producing odd-order products that spread symmetrically either side of the wanted signal.
• Spurious emissions. Both effects put energy outside the necessary bandwidth, breaching the unwanted-emission conditions of the amateur licence and attracting interference complaints.
• Poor readability. An over-driven signal is distorted, not louder; average talk power actually falls.

Prevention:

• ALC (automatic level control) samples the PA output and reduces the drive when it exceeds the set level. Set the microphone gain so the ALC meter just enters its marked zone on voice peaks, and no further.
• Use speech processing sparingly; heavy compression plus high mic gain is the usual recipe for splatter.
• Monitor your own transmission with a monitor scope (an oscilloscope trapezoid or envelope display) or a two-tone test.
• Ask for an on-air report from a station a few kilohertz away, and never advance a control simply because the output meter reads higher.