[linux-dvb] Re: DVB-T and zigzag scans question

[John Dalgliesh <johnd@syd.microforte.com.au>]

On Sat, 21 Jun 2003, Andrew de Quincey wrote:

> With DVB-T, the zigzag scanning code chooses a step size as follows:
> 
> stepsize = fe->info->frequency_stepsize * 2;
> 
> frequency_stepsize for OFDM frontends seems to be set to 166667 usually,
> which means it zigzags using a stepsize of 333334.
> 
> Thing is, in my experience, frequency bending doesn't work for DVB-T either. 
> Using a stepsize of 166667 for trying to lock on to a slightly displaced 
> signal is way too large, let alone 333334.
> 
> My guess for the problems with the tda100045h is that it loses the lock,
> kicks off a zigzag scan, then never manages to tune again because the
> steps are so large.
> 
> Does zigzag scanning/frequency bending work for anyone with DVB-T cards?
> Or is it some quirk here in the UK that means it achieves nothing
> useful?

I'm not using standard tools, but I can tell you that in order to receive
Channel 9 in Sydney which is supposedly on 191.5MHz, I have to go up to
the next step size, e.g. 191.6MHz does it fine. The two (quantised)
frequencies on either side give me nothing. So going up double the 
step size would have been a bad idea (if I understand how these tools are 
supposed to work...)

In my scanning I've been going up _half_ the step size to make sure I
catch any weird rounding of the frequency at the frontend level. Of course
that was before I discovered that my building antenna was completely
broken for digital TV :)

> I notice for DVB-C cards, zigzag scanning is disabled: stepsize is hardcoded 
> to 0. 

{P^/



-- 
Info:
To unsubscribe send a mail to ecartis@linuxtv.org with "unsubscribe linux-dvb" as subject.