On VHF and UHF it is not power that decides, but the right mode. Everyone knows FT8 from HF — but on 2 m, Q65 still pulls signals out of the noise long after FT8 has fallen silent, and meteor scatter only really works with MSK144. This article sorts the WSJT-X modes by propagation type, using the official figures from the WSJT-X User Guide and the Q65 Quick-Start Guide. And it corrects a mistake we made ourselves: FT2 is not a WSJT-X mode — more on that below.
The overview: eleven modes, four jobs
The current release WSJT-X 3.0.2 (June 2026) ships eleven protocols: FST4, FT4, FT8, JT4, JT9, JT65, Q65, MSK144, WSPR, FST4W and Echo. Only four of them really matter on VHF/UHF — the rest is either for HF (FT8, FT4, JT9), for LF and MF (FST4, FST4W) or a special tool (Echo measures your own moon echoes).
The sensitivity figures below come from the WSJT-X User Guide and denote the threshold for a 50 % probability of decode, referenced to 2500 Hz noise bandwidth. For comparison: a signal becomes visible on the waterfall at around −26 dB; it only becomes audible at roughly −15 dB.
| Mode | Period | Bandwidth | Threshold | What it is for on VHF/UHF |
|---|---|---|---|---|
| FT8 | 15 s | 50 Hz | −21 dB | Everyday work: tropo, Sporadic-E, aurora, aircraft scatter |
| FT4 | 7.5 s | 83 Hz | −17.5 dB | Contesting, when signals are strong |
| Q65 | 15–300 s | 19–1733 Hz | −22.2 to −33.8 dB | Ionoscatter, troposcatter, rain scatter, TEP, EME |
| MSK144 | 5–30 s | 2400 Hz | ping dependent | Meteor scatter, aircraft scatter |
| JT65 | 60 s | 178 Hz | −25 dB | The EME classic, largely superseded by Q65 |
| JT4 | 60 s | 17.5 Hz (JT4A) | −23 dB | Microwave EME, wide submodes up to 24 GHz |
Q65: the actual VHF weak-signal mode
Q65 was introduced in 2021 with WSJT-X 2.4.0 (Joe Taylor K1JT, Bill Somerville G4WJS, Steve Franke K9AN and Nico Palermo IV3NWV). It uses 65-tone FSK and is explicitly built for paths with heavy Doppler spread — exactly what characterises troposcatter, rain scatter, ionoscatter, TEP and EME. Where Doppler spread exceeds a few hertz, Q65 is the most sensitive mode in WSJT-X, full stop.
The price is complexity: Q65 has five period lengths (15, 30, 60, 120, 300 s) and five submodes A to E with 1×, 2×, 4×, 8× and 16× tone spacing. A designation such as Q65-30A therefore means a 30-second period with the narrowest tone spacing. The base figures from the Quick-Start Guide:
| Period | Tone spacing (A) | Bandwidth (A) | Threshold | with AP decoding |
|---|---|---|---|---|
| 15 s | 6.67 Hz | 433 Hz | −22.2 dB | −23.7 dB |
| 30 s | 3.33 Hz | 217 Hz | −24.8 dB | −26.6 dB |
| 60 s | 1.67 Hz | 108 Hz | −27.6 dB | −30.2 dB |
| 120 s | 0.75 Hz | 49 Hz | −30.8 dB | −32.5 dB |
| 300 s | 0.29 Hz | 19 Hz | −33.8 dB | −37.4 dB |
Which submode for what? The developers give concrete recommendations in the Quick-Start Guide, so there is no need to guess:
- Ionoscatter on 50 MHz: Q65-30A (QRP: Q65-120E)
- Ionoscatter on 144 MHz: Q65-60C
- TEP on 50 MHz: Q65-15C or 30C
- Troposcatter and rain scatter at 10 GHz: Q65-60D
- EME: 50 and 144 MHz Q65-60A, 432 MHz Q65-60B, 1296 MHz Q65-60C, 10 GHz Q65-60D
- Small-dish EME, 10/24 GHz: Q65-120E
The ionoscatter case is remarkable: K1JT and K9AN demonstrated on a 1150 km path that with 300 W practically every Q65-30A transmission is copied correctly. With a modest Yagi and 100 W, contacts out to around 2000 km are possible on 6 m on most days of the year — with a dead band, no Sporadic-E involved. It works best around midday and in summer.
MSK144: meteor scatter in fractions of a second
MSK144 is the only "fast" mode that really counts on VHF. It is built to exploit extremely brief signal enhancements: meteor trails, aircraft reflections, other scatter. Where the "slow" modes send one message per period, MSK144 repeats the frame continuously — a standard frame lasts only 72 milliseconds, the shortened Sh variant with hashed callsigns just 20 ms. When a ping hits, the message is in there several times over.
Technically: OQPSK at 2000 baud, the signal occupies the full SSB bandwidth (about 2400 Hz) and is always centred on 1500 Hz audio. Transmit and receive filters should be as flat as possible from 300 to 2700 Hz, and the frequency offset between you and your partner must stay within ±200 Hz. Period lengths of 5, 10, 15 and 30 s are selectable — 15 s is the recommendation for most purposes.
Best time: the morning hours between roughly 04 and 08 UTC, when the Earth turns into its orbital direction and sweeps up more meteors. Add the showers: Perseids in mid-August, Geminids in mid-December, Quadrantids in early January. If you want to practise, use the April Lyrids as a dry run before things get serious in August.
Frequencies: IARU Region 1
This matters, because many guides written in the US will send you to the wrong place: Region 1 uses different meteor scatter frequencies from Region 2. On 2 m, MSK144 here lives on 144.360 MHz — 144.370 MHz is the old FSK441 frequency, not MSK144.
| Band | FT8 | Q65 terrestrial | Meteor scatter (MSK144) | EME |
|---|---|---|---|---|
| 6 m | 50.313 MHz (DX: 50.323) | — | 50.280 MHz | Q65-60A |
| 2 m | 144.174 MHz (alt. 144.178) | 144.180 MHz | 144.360 MHz | Q65-60A |
| 70 cm | 432.174 MHz | 432.180 MHz | 432.370 MHz | 432.060–432.150 MHz (Q65-60B) |
| 23 cm | 1296.174 MHz | 1296.180 MHz | — | 1296.060–1296.150 MHz |
Terrestrial Q65 is normally run as Q65-15B with 15-second periods, whereas EME uses 60-second periods. JT65B still sits on 144.170 MHz for terrestrial work — a leftover from the days before Q65.
And FT2? It belongs in this article only as a contrast
This page previously claimed FT2 had been introduced into WSJT-X by K1JT in 2026. That was wrong, and we are correcting it. The facts:
- FT2 comes from Martino Merola, IU8LMC, together with the ARI Caserta team — not from Joe Taylor.
- It does not run in WSJT-X but in Decodium, a separate fork.
- The first on-air demonstration dates from 16 February 2026; since 22 March 2026 FT2 has been registered as
SUBMODE=FT2in ADIF 3.1.7. - Key figures: 3.8-second period, around 150 Hz bandwidth, decode threshold measured so far at −12 to −14 dB, clock accuracy ±50 ms.
And that threshold is exactly why FT2 has no place in a VHF weak-signal article: it is roughly 9 dB less sensitive than FT8 and more than 15 dB less sensitive than Q65-60A. FT2 is a speed mode for strong signals on HF — contests, pile-ups, high QSO rates. For meteor scatter, ionoscatter or EME it is simply the wrong tool. If you want to try FT2, we have devoted a separate article to it.
Setup: what matters on VHF
The software is free from wsjtx.github.io for Windows, macOS and Linux. Three points where VHF differs from HF:
- Time synchronisation is not optional. On HF, FT8 forgives a lot; with MSK144 and the short Q65 periods a skewed clock kills every decode. Set up NTP and check it — not "Windows time will do".
- The right submode beats more power. There are almost 3 dB between Q65-30A and Q65-60A — the equivalent of doubling your transmit power, but it only costs patience.
- Check split operation and filters. MSK144 needs the full SSB bandwidth; a filter set too narrow clips the signal and prevents decodes.
If you are setting up WSJT-X for the first time — sound card, CAT control, levels — this covers the basics:
In short: which mode when
- Band open, signals usable (Sporadic-E, tropo, mountain reflections): FT8, or FT4 in a contest.
- Band dead, but ionoscatter always works: Q65-30A on 6 m, Q65-60C on 2 m.
- Meteor shower or morning hours: MSK144, 15 s.
- Moon is up: Q65-60A/B/C depending on band, JT65B only where still needed.
- Maximum QSO count on HF: FT4 — or FT2 in Decodium, if the signals allow it.
73 – your oeradio.at editorial team
Sources
- WSJT-X User Guide (K1JT et al.) — mode parameters, sensitivities, MSK144 details
- Quick-Start Guide to Q65 (Taylor, Somerville, Franke, Palermo, 2021) — submode tables and recommendations
- WSJT-X project page — current version 3.0.2
- VERON VHF/UHF frequency list — IARU Region 1 practice
- ft2.it and DXZone — FT2/Decodium
Transparency notice
This article was researched and written with the support of AI (Claude, Anthropic). The technical figures come from the primary sources linked above. The article was revised on 2 August 2026: an earlier version wrongly attributed FT2 to Joe Taylor K1JT and treated it as a WSJT-X mode — both were incorrect and have been corrected. Spotted an error? Write to us at [email protected].