On Saturday 21 October 2006 00:40, Eric Paynter wrote:
> On Thu, October 19, 2006 2:35 pm, Eric Paynter said:
> > Right now, on channel 19 (151.25MHz) tuner 1 has loads of static -
> > picture is barely discernable, but on the same channel, tuner 2 is
> > near perfect. Nearby channels (17, 18, 20, 21), are all near
> > perfect on both tuners. On the channels in the 40's (above 300MHz),
> > tuner 1 has mild crosshatching and tuner 2 has strong
> > crosshatching, so in that range, I'd say tuner 1 is better.
> >
> > I'm going to play with various combinations of the "+4" in the
> > different frequency ranges to see what effect that has.
>
> OK, I did some poking around and have some questions:
>
> - Is there any way to detect the current signal strength and
> determine automatically if the Low Noise Amplifier needs to be
> enabled for a particular channel?
For PAL: yes, for NTSC: possibly (see below)
> - Is there any way to detect the current signal strength and vary the
> strength of the Low Noise Amplifier to achieve the optimum signal?
No.
>
> I bought an external amplifier and did some tests, which yielded some
> interesting results. I chose three channels that tend to give me
> trouble and I tested with no amp, with the internal amp, with the
> external amp, and with both amps. Results were:
>
> Channel 19: the external amp alone gave the best picture.
> Channel 45: the internal amp alone gave the best picture.
> Channel 9: using no amp gave the best picture on Tuner 2, but on
> Tuner 1, it was best with the internal amp.
>
> Both amps combined always yielded a very bad picture, which makes me
> think that too much amplification is overdriving the tuner. If it was
> possible to know when amplicification was needed per channel, perhaps
> the internal would be sufficient. If the internal could be varied,
> then even better.
90% of all tuners are similar to the Philips tuners (FM1236/FM1216
series). Generally the only differences are the precise frequency
boundaries between the three frequency bands and how to select the
correct band. The Samsung TCPN2221P30A is no exception and the only
special thing about the NTSC tuner is that bit 2 of the Band Switch
Byte controls the Low Noise Amplifier.
Nothing else is mentioned about the LNA in the datasheet.
The datasheet for the Samsung TCPG6221P30A (the PAL version) is a bit
more verbose. Again, it follows the standard tuner model, except for
the LNA bit. However, in the appendix there is a LNA On/Off flow chart
which shows when to turn the bit on or off. It uses the ADC value which
are bits 0-2 of the tuner status byte.
The algorithm is as follows:
LNA=0
loop {
read ADC
if (ADC == 4) {
LNA=1
}
else if (LNA == 1 && ADC <= 2) {
LNA=0
}
}
I've tested with this in the past, but with confusing results. However,
those tests were done by by someone else who had this PAL tuner and not
by me personally, so something may have gone wrong there. But it is
something that you can try and see what happens.
I've no idea if it is valid for the NTSC tuner as well (I expect it is,
but I'm not sure). And even if it does work for the NTSC tuner I've no
idea how to implement it in the kernel: no other tuner needs such
monitoring.
BTW, I find the 'ADC == 4' test suspicious. I'd expect ADC >= 4 there.
Something to test as well.
Anyway, that's all the Samsung-specific stuff from the datasheets.
Everything else is standard tuner programming.
Hope this helps,
Hans
_______________________________________________
ivtv-devel mailing list
ivtv-devel@ivtvdriver.org
http://ivtvdriver.org/mailman/listinfo/ivtv-devel
