:)
Holger Waechtler wrote:
Ragnar Sundblad wrote:As for AGC-registers - I guess that should be in the SP8870?yes.I found one document that describes some registers, but far from everything you would like to know, and not even all that is needed to get it going on a TT-DVB-T (especially the mystery 0301 register). If anyone has more information on it than the sp8870c-ch-ds.pdf Rev 1.1/9903 I'd be happy if you could share it with me. I have some more problems with it, it hangs (or probably just screws up its settings so it has to be cleaned up) every now and then when retuning. I followed you advice and finally came to playing with cranking up the charge pump current, which did a huge improvement. I changed the charge from 0 to 3 (120uA to 1200uA), which gave me pretty a nice and clean reception for the first time! Setting it to 2 (555uA) wasn't enough. I am not sure this is the right thing to do, and would really like some advice here (and/or the datasheet :-). Q1: Could it be harmful or otherwise bad to change it?
no, you can just make the PLL loop more stable/unstable. Maybe you need different charge pump current settings for different frequency bands.
Sometimes you can improve things by increasing the current while setting up a new frequency divider ratio and decrease it again when the PLL is stabilized.
Q2: What does this do anyway - up the PLL feedback loop BW?
The PLL is a dynamic system, the charge pump setting affects stability to the VCO. You can imagine the charge pump current as the force to driver an oscillating or rotating mass-spring system into a new frequency and/or place. When you push or pull too fast you'll oversocillate, if the force is not powerful enough you'll never reach the destination.
You'll find many texts about PLLs on the web, here a pretty nice paper on the Philips homepage:
http://www-us.semiconductors.philips.com/acrobat/applicationnotes/AN177.pdf
I think one could also do some more work with the dividers than is currently done too have it retune faster and optimizing for being more stable, but since the SP8870 keeps screwing up on me it is hard to test it. Some advice from you tuner guys on how to proceed with this would be really good.
You could also try smaller/wider frequency steps. This changes the effective frequency offset the onchip AFC circuit has to compensate.
And remember to set all parameters the SP8870 can't find on its own, for low frequency bands (e.g. band 5) you'll have to use 7MHz bandwidth. IIRC the SP8870 can't probe this automatically.
PS. A note for anyone wanting to try this: You should also in sp5659_set_tv_freq change ...(pwr << 5)... into ...(pwr << 6)..., this must be a typo that newer has been a problem as pwr is 0. The same goes for alps_bsrv2.c.Could you please prepare + send a patch? thanks, Holger
-- Info: To unsubscribe send a mail to [EMAIL PROTECTED] with "unsubscribe linux-dvb" as subject.
