[opendtv] Re: Latest S/N tests: tuner architecture

Dale Kelly wrote:

> S.J. Birkill's following message is very much worth reading. I'm reposting 
> since the subject line used might cause one to pass it over.

- sorry, forgot to change the sunject line from my digest version.

Doug McDonald wrote:

> Do these filters use varactor tuning diodes? How do they
> avoid overload problems with intrinsically nonlinear
> devices (short of dual balanced designs to remove
> even order nonlinearities of course). Where I live
> one FM station drives about 0.3 volt peak of signal into 75 ohms.
> 
> But with varactor diodes doing the tuning, one should use
> feedback at assure that the RF tuning was centered on teh
> channel, thus assuring no problem in that regard.

and Dale wrote:

> It seems that you can construct a relatively good single conversion receiver
> by incorporating a tracking filter and an LNA. If you intend to build an
> inexpensive receiver, you eliminate those devices and forego their benefit,
> which appears to be the current U.S. DTV model.

- Here is Infineon's reference design for ATSC apps: 

http://www.infineon.com/cmc_upload/documents/084/467/AppNote_TUA6034_PartIII
_ATSC.pdf

Tuning varactors are in single-ended configuration, with the filter split
either side of the MOSFET AGC RF amp. Tracking is achieved by matching
between filter and VCO diodes. This is standard practice for analog tuners
and most digital tuners, everywhere AFAIK. Are things different in the US?
Don't their sets also use those same little square tuner cans from Philips,
Thomson, Salcomp, LG, Samsung, Hwa Lin, Siel, Microtune, Hitachi, Sharp or
Sony (have I forgotten anyone?) I don't think they'd survive, even in an
analog off-air world, without that tracking filter.

Donald Koeleman wrote:

> Well, Steven J. has been developing tuners/demods, as long as I've been
> alive, and has quite a reputation for it, wich means most of the stuff in
> his message went over my head. Just like I quickly realised I wasn't going
> to mess with the 'Mitsubishi module';-).
> 
> Steven, will your early satellite books ever become available as pdf's?

- Thanks Donald. I have no plans to issue the old satellite stuff in
electronic format. Maybe when I get a few hours to rub together! But it's
all pretty much out of date now, isn't it? Except the pure theory, I suppose.

Bert Manfredi wrote:

> Also, with all the incredibly bad press 8-VSB
> receiver manufaturers have been getting over the
> past 7 or 8 years, you'd think they'd at least get
> the most basic of basics right, wouldn't you? All
> the equalizer prowess in the world isn't going to
> fix such basic problems as IM distortion, right?
> 
> Perhaps it's the search for that -85 dBm
> sensitivity that is driving this strange behavior?
> Inquiring minds would love to know.

- That's part of it. And also because the specs and tests for strong-signal
performance, D/U ratios and interference protection ratios just aren't
rigorous enough. In the UK, products are routinely tested for a -28.3dB
protection ratio against a single adjacent-channel DTT interferer at -35dBm
(or -37dB at -25dBm for analog), criterion "picture fail". These specs are
in fact adequate in most British scenarios where co-siting is almost
universal, analog 'big stick' ERPs run in the 500kW - 1MW range and
main-station DTTs make do with 5 & 10kW -- a 20dB difference. (And of
course <cheeky grin> we have COFDM). The guidelines are comfortably
exceeded by most units on the market, but they'd be woefully (IMO)
inadequate for the USA, where multiple transmitting sites for a given
market area are the rule. A multi-carrier IMD test, something more like the
CATV tuners' CSO/CTB figures, in addition to higher single interferer
protection ratios, would seem to be required. 

That's not to say they couldn't be met with single-conversion architecture
though.

The other part is cost. At least in the US, users seem prepared to pay for
quality. In the UK, consumers expect their DTT STBs to come free with a
pack of cornflakes, or at least to pay (at today's rates) no more than $75
for them, and to return them for a full refund if they don't like them.

Bert also wrote:

> In the Linx tests conducted in Chicago by Linx
> themselves, as I recall, every single failure was in
> sites and frequency bands where there was a strong
> adjacent NTSC channel. Analog adjacent channels would
> be particularly problematic, because of their stronger
> carrier and stronger peak signal power. This implies
> less than ideal IM distortion specs could be an issue.

ACI protection is a function of both tuner and demod. Wideband tuner RF AGC
reduces CNR of wanted signal in presence of strong unwanted; tuner channel
filtering allows skirts of analog signal to challenge ADC range; intermod
places third-order products from adjacent NTSC sound and vision (oops,
aural and visual) carriers into the wanted DTT. But then its down to the
demod device to handle the lower CNR and reduced ADC range, and to reject
what is now (as far as the demod is concerned) an in-band (CCI) interferer.

SJB

 
 
----------------------------------------------------------------------
You can UNSUBSCRIBE from the OpenDTV list in two ways:

- Using the UNSUBSCRIBE command in your user configuration settings at 
FreeLists.org 

- By sending a message to: opendtv-request@xxxxxxxxxxxxx with the word 
unsubscribe in the subject line.

Other related posts: