[time-nuts] signal from DirecTV
SAIDJACK at aol.com
SAIDJACK at aol.com
Mon Sep 22 13:51:49 EDT 2008
Hello Andrew,
unfortunately I am not sure what you mean by Sat signal/code frequency.
The local STB PLL generated 27MHz on the video signal will be derived from,
and locked to the PCR time stamp inside the MPEG stream of the broadcaster.
The broadcaster will likely use an in-house frequency standard to generate
this, but it is entirely up to them what they are using. My comment was that
this signal was not very stable from what I have seen.
If you are interested in the RF frequency of the Sat signal, things get very
complicated because the Sat signal is mixed with an LO at the LNB, and so
not knowing the LO offset error inside the LNB you would have to hack into the
LNB to get the RF frequency of the Sat. This is not as simple as using a
27MHz bandpass: even the IF is in the 800MHz to 2GHz range, and the RF is
somewhere above 30GHz I think.
bye,
Said
In a message dated 9/22/2008 09:50:01 Pacific Daylight Time, novick at nist.gov
writes:
Said,
If I measure the output of the STB, am I not just measuring the PLL in the
box?
You mentioned in your original post that the outputs of different STBs
behave differently. If I'm looking for the frequency offset and stability of the
satellite signal/code frequency, then it seems like I need to tap off of the
signal at the antenna input. Maybe with a 27 MHz bandpass to get rid of the
modulation information?
Thanks for any thoughts on this.
Andrew
At 02:30 PM 9/19/2008, you wrote:
Hello Andrew,
you can find the NXP chips on most any TV capture card for PC's. Philips
also used to make and sell large eval boards, but they are only available for
customers as far as I know.
You may be in luck to find an embedded capture card on mouser etc. that uses
one of the NXP chips.
Some set-top boxes (such as Directv's older units with CVBS inputs) also may
have them built-in.
Lastly many of the DVD-based video recorders will have them, especially if
these are from Philips/Magnavox.
Alternatively, you can probably wire it up fairly quickly (since you are
only interested in the LLC signal (27MHz clock output), but I think you would
have to have a micro initialize the chip via I2C.
Good luck,
bye,
Said
In a message dated 9/19/2008 11:35:09 Pacific Daylight Time, novick at nist.gov
writes:
Hello!
I haven't had a lot of time to look into this...is there a kit board for
any of the SAA chips?
Did you layout a circuit?
Thanks for any help!
Andrew Novick
NIST
novick at nist.gov
303-497-3378
>Message: 2
>Date: Sat, 10 May 2008 17:51:53 EDT
>From: SAIDJACK at aol.com
>Subject: Re: [time-nuts] Pulling a signal from DirecTV
>To: time-nuts at febo.com
>Message-ID: <d2b.2580607a.355772f9 at aol.com>
>Content-Type: text/plain; charset="US-ASCII"
>
>Yes, the frequency accuracy is not very good.
>
>We generated 27MHz out of the CVBS coming from a Philips Directv STB.
>
>To do this, just hook up an NXP video decoder to the CVBS/YC signal (such
as
>SAA7111, SAA7114 etc) and that will give you a line-locked 27MHz.
>
>
>Significant medium term drift, measurable on 1/10Hz if I remember
correctly.
>Jumps every so often ( a couple of times per hour) to re-sync.
>
>Depends also on the algorithms and PLL inside the STB, different vendor's
>units all behave differently from what I could see.
>
>bye,
>Said
>
>
>In a message dated 5/9/2008 08:42:08 Pacific Daylight Time, novick at nist.gov
>writes:
>
>Does anyone have experience getting a measurable frequency (in phase with
>the broadcast FQ) from a consumer DirecTV box?
>
____________________________________
Looking for simple solutions to your real-life financial challenges? _Check
out WalletPop for the latest news and information, tips and calculators_
(http://pr.atwola.com/promoclk/100000075x1209382257x1200540686/aol?redir=http://www
.walletpop.com/?NCID=emlcntuswall00000001) .
**************Looking for simple solutions to your real-life financial
challenges? Check out WalletPop for the latest news and information, tips and
calculators. (http://www.walletpop.com/?NCID=emlcntuswall00000001)
More information about the time-nuts
mailing list