On Mon, Nov 28, 2005, Steve Toth wrote: > >>Hm, the data sheet of the FMD2116ME MK3 is - lets say - a problem. > >>One issue is the DVB-T IF frequency. I got told it is 36.125MHz but > >>this can't be right. The tuner can't tune to this frequency, it should > >>be 36.167MHz. Tuning step size definitely is 166.667kHz (4MHz / 24), > >>so this is wrong in your patch. Also all examples in the data sheet > >>have bit 6 of the bandswitch byte set... > >> > > > >Errors in data sheets seem to be quite common, I only trust > >tested code. Steve, if you can check Philips reference code > >or Hauppauge Windoze driver code, this would be a more > >reliable source for these numbers... > > > > > Thanks. > > The datasheet dated Sep 8 2005, in a number of places, says the DVB-T IF > is 36.13. In fact, various version of the datasheet say the same thing. > You're telling me it's just plain wrong? I'm not saying the the FMD2116ME MK3 data sheet is wrong. But actually I don't think there's any data sheet which is free of errors. It's just as unlikely to exist as bug-free software. > The 166.7 figure comes from the control bite matrix. I guess can be > accounted for by a rounding issue in the sheet. Fair enough. > > Bit 6, (P6) is not mentioned anywhere in the datasheet other than using > the Philips example tool. I chose not to set the bit because it wasn't > documented. Tuning clearly works with or without it. could be a GPIO which is not required for your board but for other boards with the same tuner? > Any comments on the upp/lower freqs of the bands? Everyone who has the hardware should test the patch and report if it improves tuning or makes it worse. It's all I can say about this. Johannes
