[linux-dvb] Re: Cinergy 1200 DVB-T TDA10046H / auto mode causes problems

[Andrew de Quincey <adq_dvb@lidskialf.net>]

On Friday 19 Nov 2004 20:23, Wolfgang Wegner wrote:
> Hi,
>
> I do not know this card, but in general the TDA10046 works quite
> well (even) in auto mode. Lock time below 1s is what I would call
> normal.
>
> The BER values that can be read from the demodulator are quite
> "jumpy", but if you take an average of some samples, it works quite
> well. The SNR seems to be quite reliable, in a range of around
> +/- 2dB from the value "offered" by a measurement transmitter (R&S SFQ)
> with a sufficiently strong signal.
> (Maybe even better, did not document that measurement)
>
> However, depending on how the board is designed and which frontend
> module is used, the power supply can be a problem. The Philips TU1216
> seems to be quite immune against noisy power supply, while some Alps
> module (do not remember the exact type) with the TDA10046 had to get
> a _real_ stable and filtered +5V for the RF part.
> I think there was such a problem with the old DVB-S cards, wasn't it?
> But then there should be also problems with the windows software,
> such as bad sensitivity or low SNR.
>
> In case this really is a problem it also depends on the algorithm
> used for locking; is the lock algorithm currently started only once,
> or periodically? Give the demod enough time before retrying, or it
> will never lock...

It keeps going until it locks. But the thing is - it DOES lock very quickly - 
just that the data is very corrupt. The interesting thing is that the BER is 
always 0 post-lock.... assuming the BER measurement is OK, that means the 
data must be being lost after the demodulator stage.