You know, I have to confess: CW isn’t my favorite operating mode. It’s ok; sometimes I tune down to the CW portions of the ham bands and copy it for a while just to keep my hand in, but not too often. Consequently, another confession: I’ve not looked really hard into SdrDx’s CW handling until today. When I did, I found there were improvements that could be made, and, because I’m somewhat compulsive, I went ahead and made them, and you’ll find them in the 2.12r OSX beta that’s available for download as of right now. But don’t stop reading yet, please.

My old J-38 straight key

The first thing I did was some work on the high and low adjustment ranges of the CW mode demodulator envelope; CWL now allows from -10 to -1500 Hz, CWU allows from 10 to 1500 Hz. These are sane ranges; previously, it was a little weird, and I apologize for that.

The second thing was change the demodulator envelope adjustment from 100 Hz/step to 50 Hz/step. This allows you to create a properly centered demodulator envelope of an odd width, such as 300 Hz, or to cut the bandwidth all the way down to 50 Hz — a little extreme perhaps, but amazingly usable if the received station is stable.

Lastly, I experimented a bit with the audio notches to see how to best tailor the audio, and I found that with the Q of a notch set to about the third hashmark from the right of the Q scale, the notch was narrow enough to sit right on the upper edge of the demodulator envelope. I’ll give a detailed example:

I prefer to listen to CW at 1000 Hz. So to create a 100 Hz wide CW demodulator, I first select CWL (this is better than CWU because the receive frequency is “right side up”, or goes up when you tune up.)

Then I set the audio scope with SPE, FIX, and 2kc. This lets me see the CW range I’m interested in for precise tuning purposes, and puts a vertical line right at 1000 Hz, dead center where you want the signal.

Next, I adjust the lower demod envelope edge to -1050 Hz, and the upper demod envelope edge to -950 Hz. This produces a 100 Hz wide window precisely around 1000 Hz.

Next, I press SDR and set the SDR Bandwidth / Sample Rate to 50 kHz so CW tuning is as precise as possible.

Optional, but I may adjust S+ and S- even narrower, and set the DSP FFT to a higher value like 8192 so the signals are finely resolved on the display. Follow this by using GS to adjust the gridding to taste.

Next, I turn on a notch and set it for 1050 Hz, with the Q slider set to the third index marker from the right. This reduces a bit of the ringing that such a sharp filter produces; you need to keep the notch pretty narrow if your demodulator is narrow, though, or it’ll impact the actual signal, which you don’t want.

CWO is left at zero; adjust AGC to taste. I like it longer, around 5000 ms if the signals are stable and similar in strength to one another in a conversation. Single source signals like W1AW broadcasts, I always set to as long as I can get away with. Keeps the S/N as high as possible.

Finally, I press DSP and set the Mouse Click Resolution to 10 Hz. This affects not only point and click, but also, critically, the resolution at which the mouse roller steps the tuning.

Now we’re ready to go. Find the CW signal using the RF spectrum display and the waterfall, drag the demodulator envelope over to it, and use the mouse roller (while positioned over the spectrum or waterfall) to fine tune the CW signal until it is dead center on the audio scope. Time to copy!

I think you’ll be most pleasantly surprised by what you can do with this if you’re a CW aficionado.

One question for you CW fans: Given that CWU tunes backwards – that is, when you move the frequency up, the CW tone goes down – does anyone think that CWU needs to be kept? I’m considering removing it, as space is at a premium, and I can’t think of a use for it that isn’t better served by CWL. Please answer in the comments for this post. Barring any contrary feedback, I’ll remove it.