AntennaBench

Early preview · run from source

Better antenna comparisons. Evidence included.

Propagation changes. Receivers come and go. Switches happen late. AntennaBench turns a WSPR comparison into a guided, repeatable experiment—and keeps the limits visible beside the result.

Read the Summary See how it works
  • Local-first
  • No account required
  • Your bundle stays authoritative
Experiment record Plan ≠ action observations stay attributed
Report boundary Result + limits no invented winner

Illustrative interface motif; see the real generated report below.

The problem

A stronger signal is only useful when you know what produced it.

A quick A/B test can confuse propagation, timing, receiver coverage, and missing data with antenna performance. AntennaBench records the plan, what the operator actually did, and what each source observed as different facts.

A guided experiment

From a clear question to a report you can audit

The first supported workflow uses WSPR’s fixed two-minute cadence to alternate antennas while changing conditions are distributed across the run.

  1. 01

    Plan the comparison

    Choose the question, station, antennas, band, direction, and repetition count before the run begins.

  2. 02

    Switch when ready

    Follow one operator-paced instruction at a time. The recorded ready action—not the schedule—says what was connected.

  3. 03

    Collect attributed evidence

    Keep public WSPR spots, local decodes, imports, notes, missed cycles, and corrections distinct and traceable.

  4. 04

    Review results and limits

    See matched-path comparisons, reach, run quality, missing evidence, and the complete audit trail—without an invented winner.

Walk through the complete workflow

Why WSPR?

The default is about experiment design, not mode loyalty.

In the measured 40, 20, and 15 meter snapshot, WSPR exposed more reporting paths and a broader occupied-grid footprint, while its fixed cadence made repeated A/B cycles easier to control. RBN remains the better answer when the question is how your actual CW or RTTY signal is being copied.

See WSPR and RBN side by side

A real generated report

Start with the answer. Keep the evidence close.

Summary is the ordinary reading and sharing path. It leads with what the run showed, keeps each evidence population separate, and puts support counts and the principal limitation beside the result.

  • Paired shared-path signal with explicit support
  • Common-opportunity detection kept separate from observed reach
  • Uncontrolled paths described without turning missing spots into zeroes
  • Full evidence available for the complete human-readable audit
Read the Summary Open Full evidence

Read the two-minute guide before going deeper. Both documents are generated from one real, sanitized station session. They are derived views; the lossless session bundle remains the durable record. This site does not accept uploads or host operator reports. You can also see an inconclusive example with no shared paths.

Designed around the evidence

Useful without becoming a black box

Local-first by default

Plan, run, recover, review, and export without an account or hosted service.

The bundle is the record

Reports are derived views. The portable session bundle keeps the durable evidence.

Uncertainty stays visible

Gaps, corrections, source boundaries, and insufficient evidence remain part of the result.

Current availability

An early preview for contributors and evaluators

AntennaBench currently runs from source on macOS 15 or later. The local workflow can conduct manual WSPR comparisons and export reports and verified bundles. There is not yet a signed end-user download, hosted report publishing, account, or automatic antenna conclusion.

View the project on GitHub