Topband: PSK-31

[Jim Brown]

On Tue, 27 Apr 2004 19:09:39 -0400, Joe Subich, K4IK wrote:

>Unfortunately, most sound cards lack sufficient headroom ...
>particularly in the so called line out channel.

Your point is well taken. I just looked at the output of the computer
running full gain from the PSK software. The third harmonic (the
strongest spur) is 24 dB below the fundamental. Not hard clipping, but
not great either. What we used to call "communications quality."  I
dropped the output gain 3 dB and the third drops to -45. I dropped it
another 3 dB (that is, 6 dB below full gain) and the third drops to -60
re: the fundamental. 

Tom said:

>I wouldn't put any stock in the idea that a pad reduces
>common mode buzz or RFI.

Well, it does if it's in the right place. Let's say that the pad is
right at the input of the audio equipment (the radio's mic preamp, or a
power amp, or whatever) and there is 10 mv of noise induced into audio
cable. That noise will be added to the signal. If the signal is 100 mv,
the s/n is 20 dB. If the signal is 1 volt, the s/n is 40 dB. If the pad
lets me put 1 volt on the audio cable without clipping the input stage,
it has improved the s/n ratio by the loss of the pad. This is exactly
what I've done with my interface. 

Let's say that the computer's output stage generates 1 mv of noise. If
I use the 20 dB pad on the input of the radio and crank the computer to
use all of its headroom and make the signal be 1 volt, that gives me 60
dB s/n. If I turn down the output of the computer by 20 dB and run it
straight into the radio with no pad, the computer's output noise is
still 1 mv but the signal is only 100 mv. That's a s/n of only 40 dB. 

That's how a pad can reduce noise, hum, buzz, and RF. BTW, the hum/buzz
we're talking about is power-related noise between the ground of the
computer and the ground of the radio -- typically harmonics of 60 Hz,
but also other noise on the power line that might be coupled into the
ground and cause current flow on the shield.