Chris:
I have used the QSD/QEC approach in my SDR. I do QEC by tweaking
the coefficients of my Hilbert Transform.
I can readily get 60 to 65 db of 'opposite sideband' rejection.
I've no doubt things can be made even better.
BUT....
Obtaining these performance levels and MAINTAINING them are quite
different matters. Temperature variations
in the two analog channels cause degradation of the balance. Further,
as one scans over the bands the balance
can change especially when the antenna is narrow band. Switching
between bands also changes the 'correction'.
Remember, 60 db means the amplitude AND phase match to something
like 0.1%. For the 20m band with a
period of 71 ns. the delay through the QSD must track to .7ns over the
temperature and voltage range.
I'm sure there are techniques for dealing with controlling the
tracking of the two channels BUT I am not
convinced that all this work is truly worth while. I have not used
the QS1R approach myself but having dealt
with the issues involved with QSD/QEC I would take a very hard look at
the QS1R approach.
ward
On May 13, 2010, at 7:40 PM, Chris wrote:
Certainly an interesting thread. But so far nothing about the QSD/QSE
approach. Which performs better a Flex 5000 or a QS1R?
Chris
--- In soft_radio@yahoogroups.com, "dharmaBum" <dharma...@...> wrote:
>
> Marco
>
> Thanks for a very understandable reply. I would also like to thank
other people for adding their knowledge and opinions.
>
> This is certainly an interesting thread...
>
> Wayne
>
> --- In soft_radio@yahoogroups.com, Marco IK1ODO -2 <ik1odo@> wrote:
> >
> > At 10:31 13-05-10, you wrote:
> > >
> > >
> > >Alberto
> > >
> > >Thanks for a great reply. I'm familiar with the analysis that you
> > >detailed. I should have mentioned that I'm interested in SSB and
PSK.
> > >
> > >You mentioned that 86 db might not be enough for HF. What drives
the
> > >quest for dynamic range? Is it the ability to dig out a weak
station
> > >next to a powerful one?
> >
> > Hello Waine,
> >
> > your receiver ADC gets all the signals that are present at his
input.
> > That may be the antenna, or the output of a preselection filter if
> > there is one between antenna and ADC.
> > You want to have a receiver sensitive enough to hear the band
noise;
> > that requires a noise figure in the order of 10 dB in HF. That
means
> > approx. -129 dBm of input power in 3 kHz BW.
> > On the other side, the ADC has to handle the strongest signals that
> > reach the input. If you listen to 40m by the evening in EU, you get
> > up to +10 dBm peak (1 V pk, measured...) due to BC signals close to
> > the amateur bands.
> > So the required dynamic range is of the order of 139 dB in 3 kHz.
> > Since it is currently unreachable, you need a preselector to limit
> > the signals that the ADC sees, and an attenuator to work on low
bands
> > (where the band noise is higher and a loss of sensitivity is
acceptable).
> > To have 139 dB of dynamic range one needs a 100 MS/s ADC with 16
bits
> > of ENOB, or 94 dB dynamic range ... current 16 bit ADCs have far
> > less, at the point that there is a very little difference in
dynamic
> > range between 14 and 16 bit converters.
> >
> > 73 - Marco IK1ODO / AI4YF
> >
>
