Dear $reader, about four month ago I put together a box[1] for VDR-usage. From the beginning on there were problems with reception quality (a/v artefacts, short gaps, VDR crashes) on vertically polarized transponders. These can be circumvented by using one special card and leaving the other one alone. Starting VDR with both cards (or the other, not so special one, alone) always results in distortion on the affected channels after minutes or hours. My first suspicion fell on the reception equipment[2], but the only (interesting?) thing I found after realigning the dish, swapping cables, etc. was that the failures reduced significantly when the multiswitch was disconnected from the power line. Sounds like there is a ground loop, but as the problems did not vanish, I guess this is the wrong path. Asking in various newsgroups about this point did not help either. Swapping cards in the PCI slots, testing in another system (also VIA) and modifying the PCI latency did not bring any improvements. Most of the time it is possible to reproduce the problem using szap/ dvbstream; I wrote some short scripts to ease the process (attached with results). However, the measurements do not really produce constant values, which drives me crazy. Sometimes everything seems fine, some time later one of the cards loses its lock when both are active, or the BER climbs up (UNC is always 0). Daytime (day/night, difference in signal strength?) has apparently no influence, just like an attenuator attached to one of the cards. The two Nova cards were ordered (and delivered) together, but obviously are from different batches: one has black tantalum e-caps, the other has beige ones. Please hit me with the clue stick! [1] HW: ASRock K7VM2 board (VIA chipset), Athlon 1700+, 256MB RAM, 2x Hauppauge Nova-S, Sigma Designs Hollywood Plus SW: Debian unstable, kernel 2.4.22-pre7 (started experiments with an older one, of course), DVB CVS-current, VDR 1.2.0 [2] Invacom LNB -> 20m cable -> Axing multiswitch -> 15m cable -> wall jack -> 10m cable -> receiver Regards, -- Steffen Beyer <sbeyer@reactor.de> GnuPG key fingerprint: 6C9B 2844 AF75 AC7A C38C 9FFD 06CB A788 398B D2D9 Public key available upon request or at http://wwwkeys.pgp.net
Attachment:
dvblyzer.pl
Description: Perl program
#!/bin/sh # NOTE: dvbstream seems to need a small patch to allow parallel invocation set -e PATH=.:$PATH echo "*** CARD0 - V" dvblyzer.pl 0 tp104 255 & sleep 45 echo "*** CARD1 - V" dvblyzer.pl 1 tp104 255 & sleep 45 echo "*** CARD0 - H" dvblyzer.pl 0 tp89 163 & sleep 45 echo "*** CARD1 - H" dvblyzer.pl 1 tp89 163 & sleep 45 echo "*** CARD0 - V, CARD1 - V" dvblyzer.pl 0 tp104 255 & dvblyzer.pl 1 tp104 255 & sleep 45 echo "*** CARD0 - H, CARD1 - H" dvblyzer.pl 0 tp89 163 & dvblyzer.pl 1 tp89 163 & sleep 45 echo "*** CARD0 - V, CARD1 - H" dvblyzer.pl 0 tp104 255 & dvblyzer.pl 1 tp89 163 & sleep 45 echo "*** CARD0 - H, CARD1 - V" dvblyzer.pl 0 tp89 163 & dvblyzer.pl 1 tp104 255 & sleep 45
Attachment:
results.txt.gz
Description: Binary data
Attachment:
pgp00002.pgp
Description: PGP signature