[linux-dvb] Re: Correction & Experimental patch against ves1x93

[a.share@t-online.de (Andreas Share)]

> On Thursday 29 July 2004 22:45, Andreas Share wrote:
> > Comments are welcome....
> here we go ;-)
>
> > Note: The frequence ranges dependent charge pump current value are based
on
> > an Philips comparison measurement between sp5659 and tsa5059, hopefully
> > this values are correct;)
> >
> > - raising up the dvb_delays in ves1x93_clr_bit() &
ves1x93_init_aquire().
> > With the old values the dvb_frontend threat doing (here) one or more
> > "re"tunings, because the ves1893 could not get a reliable lock in the
first
> > shot. Maybe someone can crosstest this to find the best value for the
> > delays.
>
> this is interesting .. you are writing, you expect that the unpatched
driver
> does not get an imdiate lock ... is it also possible that an lock is
later,
> possible for a very short moment, missing?
>
> maybe somebody can mail me the datasheet for the ves1893?
>
> hermann

The Alps specs talk about a lock up time from 100ms....

Maybe the lock will get lost for a moment, this could caused by the afc
funktion from the ves1893. So we must give it the nessasary time to get a
reliable lock.

Btw: Rev 1.3 cards have some problem with one of the 5V line in newer,
faster pc´s. this cause unrealiable lock an "jumping" str & snr values.
There exist an hardware mod to around this problem, using down regulated 5V
from the 12V line. This eliminates this problems.

Andreas