Dennis Noordsij wrote:
On Wednesday 16 October 2002 19:14, you wrote:Hi everybody, I appied some changes which make the open issues about old BSRV2 revisions and the strange modulation dependent offsets in DVB-C more robust, the zigzag scan catches now catches after some trials the right frequencies. For those of you who want to go on debugging the VES1820 code (and that would be pretty good since it could speed up tuning noticable) you might want to enable the printk() in dvb_frontend.c, line 332, dvb_frontend_set_parameters() to show the actual frequency offset. a side note: maybe it's better not to play with the reference AGC value but the AGC timing constants... - Holger
Tried again the latest CVS head last night, it will tune to QAM64 after a few tries (not stably though), but not to QAM128, although I think it is a different kind of problem than the ves1820 register values.
I introduced a bug, sorry! Could you please try again?
REL-0-9-2 tunes very fast and never fails, but with 2 cards in one PC I get occasional artifacts and glitches (never does that with 1 card), and only on one of the 2 cards. I even got a cable signal amplifier because I figured 2 splitters could cause problems. It helped a lot (without amplifier the glitches were much worse) but it's not perfect yet.in NEWSTRUCT/HEAD is some code to prevent interference between tuners, could you please try this?
Doesn't QAM128 allow denser packing of DVB data in a frequency, making it also more sensitive? (i.e. needs a better signal)
yes, QAM128 is more sensitive to noise.
Somehow I feel the new firmware with the HEAD drivers would be more solid, if only I could get them to tune :-)the firmware is the same, we just work around some bugs and catch some constellations that might crash the ARM.
Holger
--
Info:
To unsubscribe send a mail to [EMAIL PROTECTED] with "unsubscribe linux-dvb" as subject.
