EME and CW

THIS ARTICLE IS STILL BEING WRITTEN. IT MAY CONTAIN ERRORS AND IT MAY NOT CONTAIN ALL NECCASERT INFORMATION. USE AT YOUR OWN RISK!

CW was at one time (before the year 2001) the only option for small to medium EME stations. The alternative was SSB, and that takes a pretty big gun! Then Joe Taylor. K1JT, introduced WSJT, and the rest is history. Digital modes can be used to exchange information at far lower signal strengths than CW and so became the modes of choice for most stations. However CW still has features not shared by computer based digital modes. It does not depend on the use of a computer to send or decode the message. It does not requires an SSB transceiver. All it needs in an RF source that can be turned on and off, a receiver and a human at each end of the link who can copy CW. It's this simplicity (some might say "purity") which attracts amy operators to CW.

But how well does CW fare with operators who have grown up using digital modes and programs like WSJT and WSJTX? While WSJT10 retained some CW sending capability using timed message sequences and standard exchanges, WSJTX (up to at least 2.6.0) has no such modes. WSJT10 had no ability to track Doppler shift, while WSJTX does, What does all this mean?

Back in the dim and distant past (e.g. 1980), to the best of my recollection, a CW sked would tale place on a given Tx frequency, say 432.026 MHz. Both stations would transmit on that frequency and then calculate where to listen for the echo of the other station. You would not listen on your own echo frequency, you would listen on the DX station's echo frequency. This obviously required prior scheduling. At some point this changed over to the system used today for most QSOs, where the two stations transmit on different frequences, but with their echoes on the same frequency. This is also the common mode used for Digital contacts on 432 and lower frequencies. It's the typical "listening on my own echo" CQ call. The station replying to the CQ transmits on a frequency that puts their echo on the same frequency as the station calling CQ. WSJTX does all this for you when one station uses "Own echo" Doppler mode to call CQ and the other station uses "On Dx Echo" mode to reply. This requires knowledge of the other stations location in order to calculate the relevant Doppler shift.

"Old School" CW operators don't Doppler track..."

So how does a WSJTX station work a CW station on CW?

Well, the short answer is "with some difficulty" and with some station reconfiguration. WSJTX, as of version 2.6.0, has no facility for CW operation. It does have facility for setting and tracking Doppler and it can transmit a carrier, but it can't send CW. WSJT10 can send CW, but only in a fixed time sequence format which is likely to confuse the Dx station if they don't know ahead of time what you are doing. WSJT10 does not have any facility to calculate or track Doppler shift. Actually, in practice, you probably don't need to track Doppler shift changes during a QSO on the lower frequency EMEbands. The rate of change of Doppler even at 1296 is rarely higher than 10Hz/minute and since all CW decoding for EME is currently done by ear, it's pretty easy to follow small frequency shifts either in your head or periodic use of a change in RIT value. As long as you know the initial Doppler shift, you can do everything with RIT.

Your goal is to put your echo on the same frequency as the DX echo. You can do that in WSJTX by using the "On Echo" Doppler mode. You tune the DX CQ call so that it shows a trace at, say, 800Hz on the waterfall. You set your Tx audio frequency to 800Hz. Then if you transmit, your echo will appear on top of the Dx station's echo. You can send a carrier by just clicking on the "Tune" button, but how do you transmit CW?

Sending CW while using WSJTX for Doppler Control

This is messy because there are many ways to control a rig via WSJTX. You can send audio via analog or digital USB) ports, you can control TX/RX changeover via CAT lines, Vox or the RTS or DTR lines of a serial port and you can set split to "radio" or Fake It". Different ways of sending CW may require different configurations in WSJTX.

Using CWType to send

My default configuration is using an FT-897 radio with analog audio connections, split set to "radio" and PTT controlled by the RTS line of the serial port. This combination works well with a free CW sending program called CWType. I operate with WSJTX using the analog audio microphone input for receive and the speaker output for transmit. I use one of the serial port control lines, RTS, to control the PPT input on my radio. These same lines (Speaker out and RTS control of PTT) are also supported by CWType. It turns out that you can run CWType and WSJTX at the same time and WSJTX does not interfere with CWType operation (more on that later). So you set CWType to output on the speaker output and trigger PTT via the serial Port RTS line. Now WSJTX will continuously control Doppler shift and set the Radio, while CWType will control TX/Rx PTT and the timing and content of the CW message being sent.

NOTE #1 - while WSJTX does not interfere with CWType, CWType does interfere with WSJTX, so CWType must be closed before going back to full use of WSJTX for digital modes.

Note #2 - If you have audio issues, take a look at the Windows settings for audio output port. It's possible that WSJTX may be addressing a different default output port than CWType. You may need to change the default output port in Windows between WSJTX and CWType.

Using "Tune" and keying the analog audio connection

Another simple way to send CW using a Morse key is to put it in the analog audio line from the PC to the rig. When you want to send CW click on "Tune". This will send a continuous audio tone to the rig, which you can then manually key on and off. This required nothing external to WSJTX, but does require that you are using an analog audio connection to your rig,

Using WSJT10 and WSJTX at the same time

It may also be possible to run WSJT10 to send and receive CW (in strict timed sequences) while WSJTX runs simultaneously and controls Rig frequency and Doppler tracking. This is a bit messy and requires strictly timed messages, something rarely used at the moment for the vast majority of CW operation.

In terms of frequency control, once you find the DX station's signal on the WSJTX waterfall, you tune the rig so that the trace is at the audio frequency you want for CW (e.g. 800Hz). You can then copy the signal by ear. You select "On Echo" Doppler mode, and an audio frequency the same as the Rx tone (e.g. 800Hz). This will put your CW on the same frequency at which the Dx station is hearing their own echo.

What sort of CW capacity should WSJTX provide to give maximum compatibility with existing operation while making it easier to make CW contacts?

These are just ideas. I don't write the code for WSJTX so it's very easy to make suggestions which might be difficult to implement, but I think these suggestion would be useful if they could be put into practice.

1. A CW sending mode much like that found in WSJT10. It would use pre-formated messages just as digital modes do. These messages could be manually changed if desired. They would be standard exchanges like "AB1CD WX2YZ OOO OOO OOO" and "WX2YZ AB1CD RO RO RO". If desired the messages could be edited to send "RST 539" reports.
2. Unlike WSJT10, these messages could be sent on command, not just in a strict time sequence. A time sequence would be optional (e.g. 2 minute overs), but since much EME CW traffic uses random timing, it's important that the message could be started (and stopped if necessary) at any time by the operator. This would allow QSOs with stations under full manual control who were not using fixed time sequencing (i.e. almost all of them).
3. Speed would default to 12WPM with an option to set it between 10 and 20 WPM. 12 WPM is usually seem as about the optimum speed for weak EME CW and most operators can probably copy at 12 WPM.
4. Tone would default to 800Hz, with an option to change between 400Hz and 1200Hz.
5. CW reception and decoding would be by ear for two reasons. First, machine decoding of weak CW signals is not as good as even marginally competent CW operators (like me!). Second, machine decoding of CW would just turn it into a very poorly performing digital mode. If you are going to do that, use Q65 instead!
6. While CFOM Doppler control would work between two stations with full CAT control of frequency, and indeed would be an excellent mode not requiring QTH Grid knowledge, it would in incompatible with simple stations who almost always listen on their own echo. Therefore the default Tx mode for calling CQ would be "Own Echo" and the default Doppler mode for receiving would be "Call Dx". These are the same modes used of digital modes when you tune to an unknown "CQ" call. Other modes could be used if previously agreed on, but for random contacts, "Own Echo" and "Call Dx" mimic the Doppler settings of simple stations operating without CAT frequency control and Doppler tracking.

I think something like this would make CW operation much easier for many stations. It would not require external software or hardware, everything would be done from WSJTX. This, along with the ability to use strictly times sequences containing known messages, might encourage CW operation by stations who would otherwise not be inclined to try the mode.