Receiver performance · 3D

The Sherwood list, in three dimensions

Rob Sherwood has been bench-testing HF receivers for decades, and the resulting table is the reference every serious operator checks before buying. It is also, in its native form, a wall of 163 rows and 14 columns. I scraped it into a database, added USD price research, and plotted it seven different ways — because the interesting things in this data are relationships between three variables at once, and a flat table cannot show those.

Source: sherweng.com/table.html  ·  page revision “Updated 01 August 2026”  ·  imported 24 Aug 2026

161receivers tested
1,933individual measurements
1955 – 2025years of designs
110 dBbest narrow-spaced DR3
$150 – $9,500used-market span

Seven ways to read it

Each view below is a live 3D plot in the panel further down — drag to rotate, scroll to zoom, hover any point for the radio behind it. These are the things worth looking for.

VIEW 01

💵 Does money buy dynamic range?

Price × era × DR3

The cloud does not slope upward with price. A 2022 FT-710 at $875 used scores 107 dB — better than the $10,000 IC-7800 (80 dB) and the $14,000 IC-R9500 (85 dB). Dynamic range tracks the calendar, not the invoice.

VIEW 02

📉 The phase-noise ceiling

RMDR × narrow DR3 × wide DR3

The hard physical limit behind the whole list. Not one of the 147 receivers with both figures beats its own reciprocal mixing dynamic range at close spacing. Points sitting well right of the diagonal are held back by something else — mixer or roofing filter — and still have RMDR in reserve.

VIEW 03

🏭 Who led, and when

Best DR3 by manufacturer and decade

The lead changes hands. Read left to right and the story of the last five decades of receiver design falls out of the bar heights.

VIEW 04

🧬 Performance fingerprints

Top 30 receivers × every metric, normalised

If a radio were uniformly excellent it would be a flat high ridge. Almost none are. A top-ten finish on the Sherwood list happily coexists with a mediocre noise floor or a weak filter.

VIEW 05

🎚️ What the preamp actually buys you

Bare noise floor × preamp gain × DR3

The classic front-end trade, laid out. Receivers with the worst bare noise floor are the ones with the most preamp gain on tap — the preamp is compensating for the front end, not adding headroom.

VIEW 06

📦 What holds its value

List price × age × residual

Age alone does not set residual value. The only receivers still above 100% of their original list are Collins S-line — the 75S-3C, 75S-3B and winged 75S-3 fetch 160-170% of a 1960s sticker — while solid-state flagships from the 1990s and 2000s shed most of theirs.

VIEW 07

📈 Seventy years of progress

Average score per decade, per metric

The specs did not move together. Phase noise and narrow-spaced DR3 roughly double from the pre-1990 era to the 2010s. Sensitivity and AGC threshold run the other way — 1970s sets score better — because modern front ends trade raw sensitivity for headroom. Blocking barely moves at all.

The plots

Seven WebGL panels served from my own Grafana. They need a moment to draw on first load.

Drag to rotate · scroll to zoom · hover a point for the receiver open full screen ↗

Data & method

The measurements

Scraped from Sherwood Engineering’s published table and normalised into a tidy schema: one row per measured value, so a cell like -127 / -136 b / -141 6 becomes three rows carrying the preamp state, the band, the test-tone spacing and the footnote that qualifies it. 161 receivers, 1,933 measurements, 47 footnotes. Nothing is rounded or reinterpreted — the verbatim token is kept alongside every parsed number.

Reading the terms

DR3 third-order dynamic range, the list’s headline number, measured at 2 kHz (narrow) and 20 kHz (wide) spacing. RMDR reciprocal mixing dynamic range — the published LO noise figure minus 27 dB. MDS noise floor in a 500 Hz bandwidth. Front-end classes: A tracking preselector, B bandpass or half-octave filtering, C broad or no preselection.

About the prices — read this before trusting them

Prices are mine, not Sherwood’s. Observed sale data comes from RigReference’s crowd-sourced price history, converted to USD where needed, and it is thin: only 65 of 161 receivers have enough real data points to source a used value from, and single outliers were rejected where they were obviously wrong. The other 96 are my own market estimates, carried in the database with confidence = low. Treat them as a sense of scale, not an appraisal. Current new prices exist only for the 29 radios still in production.