While I like to operate a fair amount on PSK31 during non-contest time, that mode simply is not robust enough to cope with the flutter and Doppler smearing on the trans-polar paths. To be fair, the popularity of PSK31 and its performance on a light band make it a valuable asset in the mode arsenal. However, phase-shift keying detection requires a relatively pristine and stable carrier. Also, the lack of forward error correction in the data stream makes it completely vulnerable to corruption caused by the peculiarities of a turbulent ionosphere.
Various other weak signal modes on HF have captured my interest for a while, as you can see by reading earlier posts on playing with Olivia during a solar storm and impressive results on QRP via the Weak Signal Propagation Reporter Network. Modes based on multi-tone frequency-shift keying are inherently better suited than PSK for rough skywave conditions. Detection is easier, because the demodulator does not have to lock to an unsteady carrier. Also, if the data stream carries some redundancy and/or error-detection or error-correction coding, the impact of a lost tone here and there due to selective fading can be minimized.
Over a decade ago, Joe Taylor K1JT began to develop some rather robust modes for weak signal communications at VHF to bring exotic propagation modes like moonbounce and meteorscatter more down to earth for hams unable to erect giant arrays and to run legal limit power amplifiers. The project evolved into "Weak Signal Communicatons, by K1JT", or WSJT for short. The protocols now in use include FSK441 for meteorscatter and JT65, which has proven very popular for V/UHF moonbounce and is increasingly used by weak-signal enthusiasts for HF skywave propagation. JT65 has a number of sub-modes, but you will find JT65A in common use on the HF bands. If you have a soundcard interface for your radio and can keep the PC clock fairly well synchronized, you have all the fixin's for a JT65A operation once you add some software.
Several weeks ago I downloaded the WSJT 9.0 package, which is available for Windows and Linux. I must admit to have been a bit overwhelmed by reading through the user's guide, but eventually things started to make some sense. Getting everything set up and ready to use takes some deliberation, and it helps to have an experienced Elmer to guide you through getting on JT65A. (The Alaska-Yukon Weak Signal Group is a great resource for us northern hams.) Those a bit impatient who just want to cut to the chase with a single mode can download JT65-HF written by Joe Large W6CQZ. The software runs on Windows only. The interface is intuitive, and soon you can be pointing and clicking your way to DXCC.
Make no mistake. JT65A is not a ragchewing mode. Each transmission and decoding cycle takes about one minute, and messages are limited to 13 characters. Nevertheless, it is very effective getting basic information exchanged across an otherwise-impossible channel, even more effective than CW when the signal-to-noise ratio dips to the -10 to -20 dB range. Case in point: Last night on 40M, I monitored an exchange between ZS (South Africa) and FR (Reunion Island) in the SW corner of the Indian Ocean. I couldn't hear a thing, and I could barely make out the tones on the waterfall display.
JT65A is powerful stuff! Many operators use 10-25W into simple wires or vertical antennas for worldwide contacts. These are during conditions in which your grand pappy would have had sense enough to go fishing instead.