[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

RE: [linrad] SSB

Hi, Roger, Kohjin, Conrad, et al!

> Linrad can be used for SSB, although I don't think Leif has optimized it
> for SSB yet.  If you go to mode D "SSB" it will remember the filter
> settings, screen and other parameter choices you make and store them
> separately from those for mode A "narrow cw".  Then you can just flip
> back and forth between the modes with minimal effort.  I find that
> Linrad does quite well on SSB.  It takes a bit of fiddling with the
> vertical red bars in the baseband to get the bfo right and the station
> centered within the filter passband, but once its done its set till you
> want to change sidebands.  Then you must fiddle again.  Remember to turn
> off the AFC when you go to SSB ;)  Of course you can even listen to
> shortwave broadcasters this way.
The fiddling is not too complicated. First set the bandwidth you like.
2300Hz for example. For short delay times through the filter, use the
buttons in the upper left corner of the baseband graph to display the
filter function with few points. (The delay is proportional to the
fft bin bandwidth)

Then place the BFO line 250 or 300 Hz or whatever you like outside the
filter edge.

To receive an SSB signal, just click on the center of it in the 
waterfall or main spectrum. For SSB you should have the second FFT enabled,
then you can fine tune in the high resolution graph. The second FFT also
gives some overload protection because of the AGC even though it does not
keep the amplitude constant. It only limits it to fit in 16 bit numbers.

The way the audio limitation is done preservs the phase at the passband
center so the ssb does not loose too much in quality when the audio saturates.
Bringing the volume down to avoid saturation will bring excellent quality
to the received SSB if the signal is weak enough to not activate the AGC
associated with the second FFT.

Some day (?) when the hardware is no longer on my table I will write a 
proper routine for SSB with the appropriate AFC and AGC for SSB voice signals.
You can all hear easily if the BFO is not centered correctly. I am sure it will
be possible to teach the art to the PC;-)
> If you are calibrating your receiver, you should know that it expects to
> be calibrated for each mode that you are going to use.  This is not a
> big deal to do.  When I do it for A, I always just go ahead and do it
> for D as well.
The reason is that some fft implementations are approximate and they
contribute to the total filter response that the calibration is intended
to cancel out.
Some day I might move the "first FFT version number" into the U menu to
make calibration independent of the mode.

Another thing is that the filters you use on HF bands may be narrow enough to
affect calibration. If you use mediocre hardware, a narrow roofing filter
will protect the mixer from the sum of too many signals. If you need very
good noise blanking you will have to recalibrate when changing the center 
frequency. With present Linrad versions I suggest several directories for
different settings of the hardware if your hardware frequency response is
changed when you move around in frequency.

Running with inaccurate calibration should usually not be any problem at all.
Only when you need a very flat noise floor to look for extremely weak signals
(VHF, UHF, noise floor is flat over 100kHz) or if you need the very best 
performance for the noise blanker (very high repetition rate for QRN) there
is a reason to worry about calibration accuracy. 

Leif  /  SM5BSZ