How to Read Transceiver “Lab Test Reports”

Without Getting Lost in the Numbers

A lot of ham “performance arguments” start the same way: someone posts a table of MDS, RMDR, DR3, NF, IP3, IMD3, SBN… and suddenly people talk as if those numbers predict how your next QSO will sound.

They don’t. They describe how a radio behaves in a controlled laboratory setup. That’s useful… but only if you know what’s being isolated, what’s being ignored, and what the test actually measures.

This article is a practical decoder ring for common lab metrics (the kind you see in ARRL-style tables and independent reports like the Sherwood list, Werner/DC4KU, ...).

RF Generator ≠ Antenna

an RF generator is not an antenna. But that’s intentional.

• An antenna brings in everything: wanted stations, atmospheric noise, man-made noise, fading, multipath, local junk, impedance mismatch, common-mode currents, and more.
• A lab signal generator gives a known, repeatable signal level into 50 Ω so you can isolate what the radio itself is doing.

Think of it like a car dyno: dyno numbers are not your commute. But they let you compare engines under the same conditions.

Noise Source ≠ “Noise Floor”, and Phase Noise ≠ “Noise Floor”

These terms get mixed up constantly. Here’s the clean separation.

Receiver Noise Floor and MDS

MDS is the input level where a weak signal becomes just detectable above the receiver’s own noise, referenced to a stated bandwidth and mode (often CW in ~500 Hz). It’s about the receiver’s internal noise performance under controlled conditions.

Noise Figure and a Calibrated Noise Source

A calibrated noise source injects a known noise power so you can compute Noise Figure (NF), which is how much the receiver degrades SNR compared to an ideal thermal-noise reference.

Important nuance: converting between MDS and NF can produce errors if you don’t know the receiver’s equivalent noise bandwidth (it is not always equal to the “filter bandwidth” written in a menu).

Phase Noise and Sideband Noise

Phase noise is the “skirts” around an oscillator/carrier. Close-in phase noise can mix strong nearby signals into your passband (reciprocal mixing), effectively raising the “noise floor” only when a strong neighbor is present. That’s why RMDR / SBN tests exist.

The Big Idea: Many HF Stations Are Externally Noise-Limited

On HF, the noise at the antenna terminals is often dominated by external sources (atmospheric + man-made + galactic), not the receiver’s internal noise… especially below ~20 MHz and especially in populated areas.

That’s why many operators never “use” the last few dB of sensitivity: the band noise is already well above the receiver’s own noise.

A Quick “What Does This Number Mean in Real Life?” Guide

MDS (Sensitivity)

What it is: the weakest signal the receiver can just pull out of its own noise, in a stated bandwidth.

What it is not: “how well you hear weak stations on 40 m with an antenna.”

If your antenna delivers external noise that is higher than the receiver noise floor (very common on HF), improving MDS further does not improve copy.

Noise Figure (NF)

What it is: receiver SNR degradation vs an ideal thermal-noise reference.

When it matters most: VHF/UHF weak-signal work, low-noise antennas, EME, satellites… any case where external noise is low enough that receiver noise becomes the limit.

On HF: often not the deciding factor unless you’re in an unusually quiet environment or using very low-noise receive antennas.

RMDR / SBN (Reciprocal Mixing)

What it is: how badly a strong nearby signal raises your in-band noise because of LO/clock phase noise.

This is the “gets deaf next to a big gun” mechanism. More sensitivity or tighter roofing filters won’t fix that if oscillator/clock noise is the limiting factor.

You notice RMDR differences when:

• the band is crowded (contests, pileups)
• a very strong signal sits 1–3 kHz away from what you want to copy
• multi-station environments, shared sites, strong local transmitters

Two-Tone IMD Dynamic Range (DR3)

What it is: how strong two equal off-channel signals can be before the receiver produces third-order intermod products that rise to the noise floor (under stated spacing, bandwidth, and settings).

Why it’s useful: a clean, repeatable way to compare strong-signal handling across radios.

Why it can mislead: real bands are not only two signals, and the “critical” spacing depends on what filters are actually in play (20 kHz, 5 kHz, 2 kHz… results can move a lot as signals move through different filter stages).

IP3 (Third-Order Intercept)

What it is: a derived intercept concept from analog nonlinearity models.

Practical reality: in direct-sampling SDR receivers, ADC overload and architecture-specific limits can dominate before a textbook IP3 model fits neatly. Treat IP3 as a comparative hint, not an absolute promise of behavior in every scenario.

Blocking Dynamic Range (BDR)

What it is: how strong a nearby signal can be before it causes gain compression (often defined as a 1 dB drop in the desired signal).

Blocking is different from intermod: blocking is “the receiver gets pushed out of its linear region,” not “it generates mixing products.”

A Reality Check Worth Repeating

Big table differences can be real… and still irrelevant for many stations.

If the intermod products and phase-noise-limited effects remain below your real band noise with an antenna connected, you won’t hear them. That’s why two radios can look far apart on a lab table, yet sound similar in everyday HF operating: the operating environment never reaches the stress conditions those tests are designed around.

Transmitter Tests: The Numbers That Matter to Your Neighbors

Receiver performance is only half the story. The other half is how clean your transmitted signal is.

• Output power (PEP) under realistic duty cycles
• Harmonic suppression (staying out of places you don’t belong)
• Two-tone IMD (how clean your SSB chain is)
• TX sideband noise / close-in phase noise (how wide your “shoulders” can become near the carrier)

Two-tone tests are useful, but broadband/noise-like excitation can be a more realistic “many tones at once” stress test for some TX chains. The goal is the same: avoid splatter and keep your signal clean where it matters.

How to Read Any Lab Test Report in 60 Seconds

• Start with test conditions. Bandwidth, mode, spacing, preamp/attenuator, AGC state, and any “IP+ / dither / randomization” features. Without these, comparisons are shaky.
• Ask what limits your station. If your HF noise is already S4–S7, you’re probably not using the last dB of MDS.
• If you operate in crowds, prioritize close-in behavior. RMDR at narrow offsets and DR3 at narrow spacing are often the real “pain points” in contest/pileup life.
• Don’t compare apples to oranges. Different spacing, bandwidth, and end criteria produce different numbers even for the same radio.
• Remember the system beats the radio. Location noise, antenna placement, common-mode suppression, and good band-pass filtering can dwarf small radio-to-radio lab differences.

Bottom Line

Lab tests are valuable… but they’re a microscope, not a crystal ball.

• They isolate one mechanism at a time (noise, phase noise, intermod, compression).
• They help compare designs under identical conditions.
• They do not include your antenna, your neighborhood noise, your grounding, your AGC setup, or your operating style.

These tests are meaningful, but most operators will never hit the edge cases where small differences matter. Treat them as lab benchmarks, not as a promise of on-air superiority.

Interested in more technical content? Subscribe to our updates for deep-dive RF articles and lab notes.

Questions or experiences to share? Feel free to contact RF.Guru for antenna and RF support.