Why do my digital-mode signals decode on some stations but not others? Checking audio, timing, CAT and mode choice
- Replies 2
- Views 26
- 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 digital-mode problems look like propagation problems. A weak-signal mode can only reach its decode threshold if the signal is clean, correctly timed and sent at the right level. This thread is a place to compare notes on the chain from the antenna to the decoder and to decide which mode suits the path.
1. Decode thresholds are not the same as usable QSO thresholds. Published figures are signal-to-noise ratios referenced to a 2500 Hz noise bandwidth. They are approximate and depend on fading, QRM and the decoder. Rough orders of magnitude: FT8 decodes down to about -20 dB, FT4 needs a few dB more, and WSPR goes down to about -30 dB because it uses a 2-minute transmission and carries very little information. Q65 has several sub-modes, with T/R periods from 15 seconds to 5 minutes. Longer periods improve sensitivity but also add fading and Doppler considerations. MSK144 is built for meteor-scatter pings, which may last a fraction of a second. PSK31, RTTY and Olivia usually need a signal you can see on the waterfall, though Olivia and other MFSK modes can work in poorer conditions at the cost of speed.
2. Time sync. FT8, FT4, WSPR, Q65 and MSK144 all depend on accurate clocks. In WSJT-X, a consistent DT (time offset) in the decode window is a quick check. If most stations show similar DT values that are not near zero, your own clock is probably off. Check this before changing anything else. A Windows machine that has not synced for a while, or a laptop that has been asleep, can drift enough to hurt decodes. Use NTP, a GPS-fed time source or a dedicated sync utility, and check that the sync service is actually running. Mode timing tolerance is fairly tight, but it is not a precise-to-the-millisecond problem. A second or so of error is a problem, though.
3. Audio levels and the sound card path. For WSJT-X, avoid overdriving the transmitter. Keep the audio level low enough that ALC shows little or no movement, and treat the Pwr slider in WSJT-X as a power control only if it is the only control you are using. Overdrive produces splatter and spurious products, and the first signs are often other stations complaining about wide signals rather than your own decodes failing. On the receive side, the input level should sit around the middle of the meter range, not pinned. Make sure the OS is not applying audio enhancements, automatic gain or sample-rate conversion on the device you use for the radio. A good 48 kHz sample-rate setup avoids many odd problems. Isolation transformers and ferrites on the USB and audio leads can cut RF feedback and ground-loop hum.
4. CAT control. Rig control gives you frequency readback and, with the right settings, PTT and split operation. Use the correct baud rate and the same data and stop bits as the radio menu. If the rig changes the audio passband in data mode, the waterfall may look different from what you expect. In WSJT-X, use Split Operation (Rig or Fake It) so that the audio tones stay in a clean part of the transmit passband, usually around 1500 Hz or in the area your rig filters cleanly. A frequently reported problem is CAT polling clashing with another program using the same COM port. Only one program can open the port at a time unless a virtual port splitter is in use.
5. Grid squares. Maidenhead locators are used for logs, awards and weak-signal exchange. Four characters, such as FN31, identify a grid square. Six characters add a subsquare for finer position. Use at least the four-character locator in FT8 and similar modes and make sure your own locator is correct in the settings. A wrong grid makes distance calculations and spotting data misleading, and in contests it can cause scoring issues.
6. Picking the mode for the job. FT8 and FT4 are structured and automated and are good for fast contacts on crowded bands. JS8Call is based on the FT8 modulation family but supports free-text messaging, relays and heartbeat-style presence, at the cost of slower exchanges. PSK31 and RTTY are keyboard modes that let you watch a conversation in real time and are common on HF. Olivia, MFSK variants and Hellschreiber are less often used but each has niches: Olivia for robust text in poor conditions, Hellschreiber for decoding visually, even through noise. Winlink with VARA, ARDOP or Pactor-style modes is a message transport rather than a casual chat mode, and it has different configuration (gateway selection, CMS/RMS, bandwidth) from the WSJT-X family. VARA is proprietary software, so check the licence terms and the gateway's supported modes before relying on it.
Questions for discussion: What do you check first when decodes drop off? Do you watch the DT column, the ALC meter or the PSK Reporter map? And which mode do you pick for the same band at different times of day?
Link to comment
https://www.hamradiobase.com/forums/topic/11418-why-do-my-digital-mode-signals-decode-on-some-stations-but-not-others-checking-audio-timing-cat-and-mode-choice/Share on other sites