[Elecraft] SSB DSP passband

[wayne burdick]

Rob Sherwood wrote:

> Subject:  SSB passband.
> When one uses the PBT function, does the DSP work in such a way as to 
> keep the filtering centered in the roofing filter window, or do you 
> start to run up against the edge of the roofing filter?

In SSB modes, the DSP passband is always "virtually" centered in the 
roofing filter. That is, if you SHIFT the DSP passband in either 
direction, the center of the roofing filter and DSP IF are kept 
aligned. (In CW mode it's more complicated; we had to choose between 
putting the DSP passband in the center of the roofing filter, or 
sliding down until it hits the lower edge of the roofing filter. We 
chose the latter, with intent to maximize opposite sideband rejection.)

When shifting the DSP passband downward, you reach a point where the 
following criteria is no longer met:

      DSP center >= crystal filter bandwidth / 2 + margin

In this case, the low edge of the requested DSP passband would pass 
through F = zero (at AF). Since this makes no sense functionally, we 
start truncating the DSP's IF and AF passbands on the low end. But 
because we're still shifting the crystal filter and DSP passbands 
downward, the desired goal -- removing high-pitched QRM -- is achieved.

In the above equation, "margin" is at present 0.2 kHz. It could be made 
user-programmable by mode, and this is on the firmware list. For SSB, 
you might prefer to have the margin be zero; this would in effect shift 
the default passband 200 Hz lower. (Everything, including this, is a 
Small Matter of Programming; but we have to carefully re-test the 
effect of any changes, so they won't happen overnight.)

73,
Wayne
N6KR