Using a Scope to Trace RF
A multimeter tells you a number. An oscilloscope tells you a story. For DC bias points and simple AF signal checks, a number is usually enough — but once a fault lives in the RF or IF stages of a transmitter or receiver, you need to see the actual shape of the waveform to understand what is wrong: is it clipped, is it oscillating when it shouldn't, is the envelope the right shape for the modulation in use? This lesson extends the general signal-tracing principles from M21D specifically into the RF domain, where the tools and pitfalls are different enough to deserve their own treatment.
This is not a general oscilloscope tutorial — basic scope operation, triggering, and probe compensation are covered in Module 12. This lesson assumes you can already operate a scope and focuses specifically on the practices that matter when the signal on the screen is RF: protecting the scope from high power levels, reading envelope shape for distortion, and recognizing parasitic oscillation.
A clean two-tone envelope (left) is smoothly rounded; an overdriven amplifier (right) flattens the peaks, indicating distortion.
View LargerProtecting the Scope from RF Power
Many antenna analyzers, dummy loads, and dedicated wattmeters include a low-level sample output specifically intended for connecting a scope safely — check your equipment's documentation. If no such sample point exists, a simple home-built RF sampler (a small capacitive or inductive pickup with a fixed, known attenuation ratio, calibrated against a known signal source before use) is a standard homebrew project and a worthwhile one for any ham doing serious bench troubleshooting work.
Scope Bandwidth and Probe Loading at RF
A scope can only faithfully display a signal whose frequency is comfortably within its rated bandwidth. A 20 MHz bandwidth-limited scope will significantly distort and underrepresent the amplitude of a signal at 14 MHz, and will show essentially nothing useful at 50 MHz or above. Before relying on amplitude readings from a scope at RF, confirm the scope (and the probe) is rated for at least several times the highest frequency of interest — a common rule of thumb is to choose a scope bandwidth at least three to five times the highest signal frequency you intend to measure precisely.
Probe loading is the second RF-specific concern. A standard 1x probe (or worse, a bare wire used as an improvised probe) presents significant capacitance to the circuit under test — often 100 pF or more including cable capacitance — which can detune a sensitive tuned circuit or load down a high-impedance RF node enough to change the very behavior you are trying to observe. A proper 10x (10:1) scope probe presents a much lower capacitance (commonly 10-15 pF) at the cost of a 10x reduction in displayed amplitude (which the scope automatically compensates for if set correctly), and is the appropriate default choice for probing RF and IF stages. For very sensitive, high-impedance tuned circuits (an oscillator tank, for example), even a 10x probe's capacitance can shift the resonant frequency measurably — in these cases a loosely coupled pickup loop placed near, but not touching, the inductor is sometimes preferred specifically because it disturbs the circuit far less than any direct-contact probe.
Reading Envelope Shape: The Two-Tone Test
For SSB and AM transmitters, the single most useful scope-based RF diagnostic is the two-tone test, widely used both for routine linearity checking and for fault diagnosis. Two audio tones of slightly different frequency (commonly around 700 Hz and 1900 Hz) are fed into the microphone input simultaneously, at a level that would represent full voice modulation. A linear SSB transmitter responding correctly produces an RF envelope, viewed at the PA output (through a safe sampler, per the warning above), that looks like a smooth, evenly spaced series of rounded peaks — often described as resembling a string of pearls or a gently oscillating envelope, because the two tones beat against each other at the difference frequency (around 1200 Hz in this example) while the envelope amplitude rises and falls smoothly between zero and the peak power level.
| Envelope Appearance | Meaning |
|---|---|
| Smooth, evenly rounded peaks, each touching down near zero between peaks | Linear, correctly operating amplifier — healthy |
| Flat-topped peaks ("squared off" tops) | Amplifier compression or clipping — overdriven by excess mic gain, ALC misadjustment, or insufficient PA headroom |
| Peaks that do not return fully to zero between tones | Carrier leak or unbalanced modulator, or excessive ALC/compression keeping the amplifier from fully unkeying between peaks |
| Ragged, uneven, or asymmetric peak heights | Nonlinearity or instability in the amplifier chain — often traceable to a specific stage via the divide-and-conquer technique from M21B |
The diagnostic power of this test comes from the fact that flat-topping is directly visible on the envelope and corresponds exactly to the intermodulation distortion (covered in Module 19) that causes "splatter" complaints from other operators — a flat-topped envelope and a report of "you're splattering across the band" are two views of the exact same underlying problem.
Spotting Parasitic Oscillation
A parasitic oscillation is unwanted, uncontrolled oscillation at a frequency the circuit was never designed to operate at — usually caused by unintended feedback through stray capacitance, inductance in long leads, or inadequate decoupling (see Module 19 and Module 20's RF layout lessons). On a scope, parasitic oscillation typically appears as a burst of high-frequency "fuzz" or ringing riding on top of the desired waveform, often most visible right at the edges of a pulse or right at the peaks of an envelope, where the active device transitions rapidly between conduction states. Because true parasitic oscillation frequencies are often well above the desired signal frequency (sometimes into the VHF or UHF range even in an HF amplifier), you may need a wider scope bandwidth than the desired signal alone would require, specifically to catch this kind of fault.
Parasitic oscillation is a serious fault to find and fix correctly: it can cause spurious emissions outside your licensed band, damage output transistors through excessive dissipation at the parasitic frequency, and intermittently appear or disappear depending on temperature, supply voltage, or even hand proximity to the chassis — making it one of the more frustrating faults to chase without a scope, and one of the more straightforward ones to spot with one.
Worked Example: Reports of "Splattering" on Transmit
Setup: A two-tone audio signal generator is connected to the microphone input, set to a level intended to represent full normal speech modulation. The transmitter is keyed into a dummy load, and the scope is connected through the load's calibrated RF sample output (never directly to the unattenuated PA output).
Observation: The envelope on the scope shows clearly flat-topped peaks rather than the expected smooth, rounded "string of pearls" pattern, with the flattening becoming more pronounced as the input tone level is increased — consistent with amplifier compression rather than a fixed circuit fault.
Follow-up test: Reducing the microphone gain control on the transmitter and repeating the two-tone test produces a clean, properly rounded envelope at the same RF output power level.
Conclusion: The fault was not a failed component at all, but excessive microphone gain driving the speech processor and PA into compression — confirmed directly and unambiguously by the envelope shape on the scope, in a way that a simple RF power reading (which might have shown an unremarkable "normal" peak power number) would never have revealed.
Result: This example illustrates an important general lesson: not every fault is a broken component. The scope's envelope display revealed an adjustment/setup problem that a voltmeter, wattmeter, or simple presence/absence check would have completely missed.
⚖ Experiment: Observe Clipping on an Audio Envelope
This experiment reproduces the same envelope-clipping concept from the worked example using only audio-frequency equipment, so you can see the diagnostic principle directly without needing a transmitter, dummy load, or RF sampler.
- A small audio amplifier circuit (an LM386-based kit or breadboard build works well)
- Two audio signal sources at slightly different frequencies (two function generators, or a smartphone app capable of generating two simultaneous tones, such as 700 Hz and 1000 Hz)
- An oscilloscope
- A way to combine the two tones before the amplifier input (a simple passive mixing network using two resistors works for this purpose)
- Combine the two tones at a moderate level and feed them into the amplifier input.
- View the amplifier's output on the scope. You should see a beating envelope pattern — the amplitude rising and falling at the difference frequency between the two tones (300 Hz in this example) — with smooth, rounded peaks.
- Gradually increase the input level (or, if your amplifier has a gain control, increase that instead) until the output begins to distort.
- Observe the envelope shape change as distortion sets in — the peaks should visibly flatten.
- Reduce the level back to normal and confirm the smooth envelope returns.
At a moderate, correctly matched input level, the envelope should show smooth, evenly rounded peaks. As you overdrive the amplifier, the peaks should visibly flatten into a "squared-off" shape — the exact same visual signature described for an overdriven SSB transmitter in this lesson, just produced safely on the bench at audio frequencies. This directly demonstrates why envelope shape, not just amplitude, is such a powerful diagnostic tool.
Frequently Asked Questions
Can I just connect my scope directly to a low-power (5 W) transmitter's output?
Even 5 W of RF into a 50 Ω load corresponds to roughly 16 V peak — likely within many scope input ratings, but still risky and not a recommended habit, since input ratings vary and a momentary mismatch or higher-power test later could easily exceed the limit. Using a calibrated sampler or attenuator every time, regardless of power level, builds a safe habit and avoids needing to recalculate a safety margin for every test.
Why does a flat-topped two-tone envelope correspond to "splatter" complaints?
Flat-topping is the time-domain appearance of amplifier compression and clipping, which in the frequency domain generates intermodulation products and harmonics that spread energy outside the intended channel bandwidth — exactly what other operators hear as "splatter" on adjacent frequencies. The two views (envelope shape on a scope, and spectral splatter heard or measured on a spectrum analyzer) describe the same underlying nonlinearity.
How do I know if what I'm seeing is parasitic oscillation versus normal high-frequency content in the signal?
Parasitic oscillation is usually visible as fuzz, ringing, or bursts that are not harmonically related to the intended signal and often appear inconsistently — changing with hand proximity, temperature, or supply voltage. Normal signal content is stable and repeatable under the same test conditions. If you can make the suspicious content appear or disappear by lightly touching a component lead or moving a wire (without actually completing or breaking a circuit), that is a strong indicator of parasitic oscillation rather than legitimate signal content.
Do I need an expensive, high-bandwidth oscilloscope to do any of this?
For most HF-band envelope and audio-equivalent diagnostics shown in this lesson, an inexpensive entry-level scope with at least 20-50 MHz of bandwidth is generally sufficient, since you are usually observing the envelope (a relatively low-frequency pattern) rather than resolving individual RF cycles at the carrier frequency itself. Direct observation of RF carrier cycles, or chasing VHF/UHF parasitic oscillation, benefits from significantly more bandwidth, but is not required for the bulk of practical HF troubleshooting covered here.
Test Your Knowledge
Answer the questions below to check your understanding. Every answer can be found in the lesson above.