SM 5 BSZ - Receiving Weak CW Signals
(Feb 11 2006)

Different opinions

Among weak signal enthusiasts there are different opinions on what is the best way to receive a marginal signal. The two main points of view are:

(1) Use a narrow filter to get better S/N.
(2) Use a fairly wide filter to avoid the ringing of a narrow filter.

The fact that both alternatives have many supporters is perhaps unexpected, since the S/N, the signal to noise ratio is inversely proportional to the bandwidth, and typical bandwidths for the two alternatives may be 1kHz and 25Hz - a difference in S/N of 16dB.

The explanation why the difference between the two methods is so small that there is an argument on which method to choose lies in the fact that the signal processing does not end at the ear phones !! The human brain is a good signal processor and under optimum conditions it acts as a very narrow filter. Nothing will be gained by putting a wider filter in front of it!!!

My own personal experience is as follows:

1. It is much easier to copy a weak signal if all external noise is suppressed. I am using head-phones intended to protect the ears in noisy environment while listening to something pleasant - radio or a CD... I can not even hear when the telephone rings!!

2. Using both ears simultaneously through head phones is more safe than using a loudspeaker in a quiet room. Multi path propagation of sound waves may suppress certain frequencies. It may be a good idea to switch the two wires to one side of the headphones now and then to produce a 180 degree phase shift between the ears - this can relieve fatigue when trying to hear a weak signal in noise.(Another way of relieving fatigue is to change the pitch of the signal.)

3. If the bandwidth is too large, above 1kHz, the "signal processor" between my ears gets overloaded and a significant loss of performance occurs.

4. Only at bandwidths below about 50 Hz I can get any improvement over what I get at 1kHz - but only at slow or moderate CW speeds, and with signals that are very frequency stable. This indicates that my "brain bandwidth" is about 50 Hz. I can work for much longer times without loss of performance when the bandwidth is low. I practically never use bandwidths above 100 Hz. Only for aurora, when a bandwidth of maybe 300 Hz together with a high pitch gives an improvement. (For aurora, using a matched filter one looks for small amplitude changes of the noise without any change in the character of the sound, so a high pitch is better. With a low pitch the ear automatically listens for variations in the signal frequency contents and the small amplitude variations become more difficult to distinguish.)

5. Noise and hum in the AF part of the receiver (and all other parts as well) has to be kept well below the noise floor originating in the pre amplifier (or antenna). Some hams may argue that this is not necessary, but I do not accept less than a 20dB increase of the voltage across my headphones when switching on my pre amplifier up in the antenna. I do count unnecessary losses by tenths of a dB.

6. When using bandwidths above about 30 Hz it is important to avoid any signal distortion for signals within the passband. The AF part has to be of HIFI quality. 5% intermodulation between two carriers (same peak amplitude as noise floor has in weak signal mode) means roughly that 5% of the signal energy is smeared out into the noise, while intermodulation between noise components raises the noise floor at the signal frequency by 5%. This kind of degradation can not be removed by a good filter after the non-linear part of the signal chain.

7. When using bandwidths below about 30 Hz it may be a good idea to introduce distortion to the signals within the passband. When the receiver has already filtered the signal to the extent that the ear - brain can not hear any difference between different frequencies within the passband, then the only task for the ear - brain is to analyse amplitude changes. The ear is not very good at this - it is logarithmic in nature and developed to watch for signal character, not amplitude. Therefore expansion of the dynamic range or symmetric clipping can make receiving much less tiring, and even improve the detect threshold for a weak signal.

Symmetric clipping at a suitable level will add odd overtones that rapidly change the characteristis of the sound as percieved by the ears when the amplitude grows above the clipping level.

To SM 5 BSZ Main Page