QMX Windows WSJT-X FT8 Diagnostic Utilities

Windows 11 utilities for running WSJT-X Improved with the QRP Labs QMX+:

1. keep Windows time accurate enough for FT8; and
2. verify the QMX+ USB receive-audio path before launching WSJT-X.

The QMX+ LCD has an audio-level indicator at the top-left character, under the A VFO symbol. A single dot during transmit or tune means little or no transmit audio drive is reaching the QMX+ over USB. Typical causes include wrong WSJT-X output device, USB re-enumeration after a reconnect, FxSound audio routing, or exclusive-mode contention.

qmx_audio_check.py measures receive audio (QMX+ -> PC). It does not test transmit audio directly, but a missing or silent receive device is evidence that USB enumeration or device selection must be fixed before debugging WSJT-X transmit audio.

Operating procedure is in CHECKLIST.md. Debugging notes are in NOTES.md.

Scope

Tuned for one specific configuration:

Component	Version / detail
OS	Windows 11
Radio	QRP Labs QMX+
QMX+ firmware	1.03.002
Software	WSJT-X Improved 3.0.0
Antenna	AlexLoop magnetic loop (2nd floor, urban courtyard)
Band / mode	30 m FT8, 10.136 MHz
Other audio software present	FxSound

Files

File	Purpose
Ensure-FT8-TimeSync.ps1	Configures and tests the Windows Time service so FT8 timing is usable after a rebuild, reboot, or failed w32tm /resync.
qmx_audio_check.py	Lists Windows audio devices, identifies likely QMX+/USB audio devices, captures a short sample, and reports where its peak level falls on a measurement scale from SILENT through CLIPPING.
CHECKLIST.md	Operating checklist for routine WSJT-X + QMX+ operation, including what not to change.
NOTES.md	Debugging notes and post-commit validation.

Requirements

Windows time utility

• Windows 11
• PowerShell
• Administrator PowerShell session
• Network access to public NTP servers

QMX+ audio diagnostic

• Windows 11
• Python 3
• QRP Labs QMX+ connected by USB
• Python packages:

pip install sounddevice numpy

Python virtual environment setup (the .venv/ directory is gitignored; create it yourself, do not commit it):

python -m venv .venv
.\.venv\Scripts\Activate.ps1
python -m pip install --upgrade pip
pip install sounddevice numpy

All commands below assume PowerShell is open at the repository root.

Usage

1. Synchronize Windows time for FT8

Open PowerShell as Administrator at the repository root, then run:

Set-ExecutionPolicy -Scope Process -ExecutionPolicy Bypass
.\Ensure-FT8-TimeSync.ps1

The script:

• requires Administrator privileges;
• sets the Windows Time service to automatic startup;
• starts w32time if it is not already running;
• configures manual NTP peers:
  • time.windows.com
  • time.nist.gov
  • pool.ntp.org
• restarts the Windows Time service;
• forces a resync;
• prints w32tm /query /status; and
• runs a short stripchart check against time.windows.com.

For FT8, the final offset should be well under +/-1 second.

2. Check the QMX+ receive-audio path

Run from a normal PowerShell session at the repository root:

.\.venv\Scripts\Activate.ps1
python .\qmx_audio_check.py

Default behavior:

• searches for audio devices whose name matches known radio hints (Digital Audio Interface, USB Audio Codec, QMX, QDX) or a user-supplied --pattern;
• lists input and output devices;
• marks likely input candidates;
• prints the Windows default input and output;
• warns if the default Windows output appears to be a radio device; and
• captures 3 seconds of mono audio at 48 kHz from the detected input device.

Useful variants:

python .\qmx_audio_check.py --pattern QMX
python .\qmx_audio_check.py --pattern USB
python .\qmx_audio_check.py --device N
python .\qmx_audio_check.py --duration 5
python .\qmx_audio_check.py --samplerate 48000

Use --device N when several input candidates are listed. Windows can enumerate the same physical device through multiple audio APIs (MME, WDM-KS, WASAPI shared/exclusive). Prefer a 48 kHz QMX+ input when available.

Audio interpretation

qmx_audio_check.py reports peak level, RMS level, and peak dBFS, and classifies the result by peak dBFS.

Label	Peak band	What it tends to indicate
SILENT	peak < -100 dBFS	No useful audio reached the capture device. Check USB cable, device selection, Windows privacy settings, and disabled devices.
VERY QUIET	peak < -60 dBFS	Usually wrong device selected or exclusive-mode contention. Check device selection first.
QUIET	peak < -40 dBFS	Low in dBFS terms, but may still decode when the Windows clock is synchronized.
NORMAL	-40 to -15 dBFS	Operating range observed on this station. WSJT-X typically shows 3-4 green dots.
HOT	> -15 dBFS	High but not clipping. Consider reducing for headroom.
CLIPPING	peak > -0.3 dBFS	Near full scale. Reduce QMX+ AF output or Windows input level.

Level is not the usual culprit

On this station, FT8 decoded normally at about -40 dBFS peak and -55 dBFS RMS when the Windows clock was synchronized. A 4-second clock offset produced zero decodes while the audio path still looked plausible.

If WSJT-X decodes nothing while the band is active, check time before adjusting audio.

On this setup, USB RX audio from the QMX+ behaved as effectively fixed-gain. No USB RX gain control was found in the firmware 1.03.002 menu during this session; SSB -> USB Gain affected TX only. No adjustment was needed for this station.

WSJT-X workflow

See CHECKLIST.md for the full operating procedure. In brief:

1. Connect and power the QMX+.
2. Run Ensure-FT8-TimeSync.ps1 after rebuilds, reboots, or failed syncs.
3. Run qmx_audio_check.py before launching WSJT-X.
4. In WSJT-X, explicitly select the QMX+ audio devices rather than relying on Windows defaults.
5. If QMX+ audio devices disappear from WSJT-X, power-cycle or reconnect the QMX+ and restart WSJT-X.

Typical WSJT-X audio-device names on Windows look like:

Digital Audio Interface (N- QMX Transceiver)

The N can change after USB reconnects. Recheck WSJT-X audio settings when Windows renumbers the device.

Windows default audio warning

Do not make the QMX+ the Windows default output device. System audio processors such as FxSound target the default output; Windows system sounds can be routed to the QMX+ audio device. Set speakers or headphones as the default; select the QMX+ explicitly inside WSJT-X.

Troubleshooting

PowerShell says the time-sync script cannot run

Use a process-local execution-policy bypass:

Set-ExecutionPolicy -Scope Process -ExecutionPolicy Bypass

Then run the script again from the same Administrator PowerShell session.

The time-sync script says Administrator rights are required

Open PowerShell as Administrator. The script exits if not elevated.

w32tm /resync reports no time data

Run Ensure-FT8-TimeSync.ps1 from Administrator PowerShell. It configures manual NTP peers, restarts w32time, and forces a resync.

The audio script lists multiple possible QMX+ devices

Run again with the specific input index:

python .\qmx_audio_check.py --device N

QMX+ LCD shows one dot during transmit/tune

QMX+ LCD top-left character, under the A VFO symbol, shows a single dot. On this station, that indicated little or no transmit audio drive reaching the QMX+ over USB.

Likely causes:

• WSJT-X output is not set to the QMX+ device;
• Windows default output is FxSound or another non-QMX+ device, and WSJT-X is configured to Default;
• QMX+ re-enumerated to a new device number since WSJT-X last bound to it;
• another application is holding the device in exclusive mode;
• QMX+ USB audio routing changed after a FxSound or Windows update.

Diagnostic sequence:

1. In WSJT-X, File -> Settings -> Audio, confirm Output is set explicitly to Digital Audio Interface (... QMX Transceiver), not Default.
2. Run python .\qmx_audio_check.py to check the RX side. If the script reports VERY QUIET or SILENT, fix USB enumeration, device selection, and routing before returning to WSJT-X.
3. If the script reports QUIET or NORMAL but the QMX+ still shows one dot, treat the WSJT-X output binding or Windows output routing as the live suspect.

A QUIET result on its own is not a problem when the clock is synchronized.

The QMX+ device is absent from WSJT-X

Power-cycle or reconnect the QMX+, then restart WSJT-X. Windows may assign a new device number.

Windows Settings -> Sound freezes or shows no devices

Known failure mode when the FxSound APO has been unloaded but not cleanly, or when the audio stack is in an inconsistent state. Use the legacy control panel, which bypasses the Settings app entirely:

mmsys.cpl

The Playback and Recording tabs in that dialog respond normally when the Settings app won't.

If mmsys.cpl also fails to commit changes, restart the audio service from an elevated PowerShell session:

Restart-Service -Name Audiosrv -Force

This forces the audio stack to re-initialize and drops stale APO hooks. Any app currently using audio will briefly lose its connection and reconnect.

Safety

Ensure-FT8-TimeSync.ps1 changes only Windows Time service configuration and queries time status.

qmx_audio_check.py is read-only. It captures a short in-memory audio sample to compute level statistics and changes nothing.

License

MIT License. See LICENSE.