And again (I know this problem has been mentioned several 
times, but I found no Solution.
What I want to do is to capture the lyics section of the 
www.basc.de homepage. It works for the first hours, but 
then (after ca. 3 hours), the following Lines appear in my 
log:
...
15:30:03	Warning: 	File not parsed, looks 
like binary:
www.basc.de/lyriks/display.php?id=05cface2f2286f939a5aa72315841ac7
...
It's definitely no binary file, the File i get downloaded 
is best shown in a hex-editor, but it has nothing in 
common with a html file.
What happened? 
I limited the Number of connections to 2, the DL-Speed to 
10k, and so on, nothing works.