I've done some more testing, and I feel much more confident about my
theory on the cx24123 bands. From my testing, I believe that the tuner's
VCO #8 has a center frequency around 2300MHz, and is only useful when
using the /4 divider (so there are effectively only 9 band selects, not
10). I think the datasheet is essentially correct in terms of showing
individual band ranges, it just mixes up the assignments.
Also, I did confirm that I can tune low symbol rate signals using the
original spec'd VGA offset values from the spec if the demodulator
sample gain is set correctly.
I did also do some experimentation to determine the behavior of the
demodulator's frequency search capability, and it definitely does do
some of that, and can be controlled with the range and initial frequency
parameters, but I haven't quite nailed down the exact behavior (for one
thing, it looks more like +/-4MHz instead of the advertised +/-10MHz).
The final thing I want to figure out before I'm satisifed with the
performance of the driver is the interaction between the demodulator's
frequency search and the linux dvb software zigzag.
I haven't gotten any comments on any of this so far -- is anybody
actually interested in these improvements? I have no idea if this is a
common card or not.
_______________________________________________
linux-dvb@xxxxxxxxxxx
http://www.linuxtv.org/cgi-bin/mailman/listinfo/linux-dvb
