Weak-signal digital troubleshooting: how do you sort out time sync, audio levels and CAT before blaming the band?
- Replies 2
- Views 10
- Created
- Last Reply
A better way to browse. Learn more.
A full-screen app on your home screen with push notifications, badges and more.
A lot of "the band is dead" or "my decodes are poor" threads turn out to be station problems rather than propagation. This thread is for sorting out the usual suspects in a sensible order, across FT8, FT4, WSPR, JS8Call, Q65 and MSK144, and the keyboard modes like PSK31, RTTY, Olivia and MFSK in FLDIGI.
Start with the clock. FT8, FT4, WSPR, Q65 and MSK144 all work in fixed time slots. FT8 uses 15 second periods, FT4 uses 7.5 seconds, and WSPR uses 2 minute periods starting on even minutes. WSJT-X shows the time offset (DT) of each decoded signal. If most stations show a consistent DT of, say, +0.7 or -0.8 seconds, the offset is probably in your PC clock rather than in all of them. A DT close to zero for the average station is the goal. Windows does not always sync often enough by default, so a dedicated NTP or similar sync tool, or a GPS-based source, is a common fix. On a field laptop with no internet, a GPS puck or a phone hotspot for sync is worth setting up before you leave.
Next is audio level. The common mistake is driving the rig too hard. Digital modes are generally constant-envelope or close to it, but an overdriven audio chain produces splatter and a dirty signal, and it often shows up as ALC action on the radio. Set the transmit audio so the ALC barely moves, then raise or lower output power at the radio rather than pushing audio. On receive, WSJT-X suggests keeping the input level somewhere around the middle of its meter range, not pinned. Also check that the PC is not applying audio enhancements, system sounds, or sample-rate conversion to the interface. A 48 kHz, 16-bit setup is the usual baseline.
For CAT control, WSJT-X can use Hamlib, and the choices that matter are the rig model, serial port, baud rate, PTT method (CAT, RTS/DTR or VOX), and split operation. Using "Rig" for split is generally cleaner than "Fake It" because the radio holds the audio tones in a better part of its passband, which reduces harmonic problems. If the radio has a built-in USB codec, you may get both CAT and audio on one cable, but check that the PC is not sharing the serial lines with other programs such as logging or a keyer.
Common questions to start with: what is your exact interface (built-in USB codec, external box, or isolated sound card interface), and what shows up on the waterfall when you transmit into a dummy load? A clean tone set on the waterfall with no extra images or wide sidebands is a good sign. Also useful to know: does the decode list improve if you narrow the receive filter, and what is your reported SNR on PSKReporter or WSPRnet compared with nearby stations?
For comparison, the modes differ a lot in what they can do. FT8 has a decode threshold of around -21 dB SNR (in a 2500 Hz reference bandwidth) and a bandwidth of about 50 Hz. FT4 is faster but needs a stronger signal, with a threshold in the neighborhood of -17 dB. WSPR can decode much weaker signals, around -31 dB, but it is beacon-only and not designed for QSOs. JS8Call is built on FT8-style modulation but adds free text and a message relay, at the cost of slower exchanges and variable speeds. Q65 is aimed at difficult paths like EME, scatter, and some terrestrial work with longer periods, and MSK144 is designed for meteor scatter, usually on 6 m and 2 m. PSK31 and RTTY are far less sensitive than those weak-signal modes, but they are simple, widely supported, and good for casual chats and contests.
What is your first check when decodes look wrong: clock, levels, or CAT?
Link to comment
https://www.hamradiobase.com/forums/topic/11782-weak-signal-digital-troubleshooting-how-do-you-sort-out-time-sync-audio-levels-and-cat-before-blaming-the-band/Share on other sites