The CGC Communicator Newsletter Web Posting

PRECISION DTV FREQUENCY OFFSETS AND CLOCK JITTER

Do you have a DTV station that must hold a precision frequency offset (plus and minus 3 Hz) against the visual carrier of a lower adjacent channel NTSC station?  Are you curious about the term "clock jitter" even if you do not have a 3 Hz offset?  If so, this letter may help you improve the performance of your DTV station.

PROBLEM ONE - MAINTAINING A 3 HZ OFFSET

KPBS RF engineering and Bob Gonsett of Communications General Corp. (CGC) have uncovered a serious problem with KPBS's two identical DTV exciters. KPBS-DT is required to maintain a precision 5.082138 MHz offset to the visual carrier of a lower first adjacent channel NTSC station.  As a practical matter, this means our DTV pilot must be accurate to within 1.5 Hz and the LPTV's visual carrier must hold the same tight tolerance so each station is doing its part to meet the 3 Hz rule.

KPBS's DTV exciters have local oscillators that are GPS-locked, but it turns out that our DTV pilot frequency is also determined by the symbol clock rate, and the symbol clock is not GPS-locked.  Our symbol clock is locked to the incoming SMPTE 310M bitstream, and the SMPTE frequency tolerance (plus and minus 54 Hertz) is so loose that normal frequency variations or additions at the studio can knock the transmitted pilot frequency out of tolerance.  (We will not need high precision pilot stability in about three years when NTSC goes dark, but until then KPBS and many other DTV stations must comply with the FCC's plus and
minus 3 Hertz rule.)

Recently, a multiplexer installed in the ASI transport stream at our studio caused the pilot to shoot off frequency.  (The ASI feeds an STL at our studio and the resulting data is converted to SMPTE 310 at the transmitter site.)  By switching to a higher stability oscillator in the multiplexer, we were able to bring our pilot back on frequency -- but you may not be this lucky.  (Who would have imagined that a change at the studio could affect the transmitted pilot frequency?)


PROBLEM TWO - PILOT AND BITSTREAM JITTER

The problem I am about to describe applies to all DTV stations, even those who do not maintain a precision (3 Hz) frequency offset against an NTSC station.  If your station has sufficient short-term pilot frequency drift (jitter), data recovery will be compromised in fringe area TV receivers.  So, if you want the best possible signal to reach those receivers, you will need to make sure the pilot jitter is minimized (jitter should be held to less than 2 nanoseconds.)  KPBS's jitter is automatically minimized by the STL receiver circuitry.

CONCLUSIONS

It is important for all DTV stations to check how (or if) their exciter's symbol clock is frequency locked.  If it is locked to the incoming bitstream, as ours is, I suggest you accurately check your pilot frequency.  Tests should be done with main and backup exciters along with all variations of studio equipment affecting the bitstream rate.  KPBS has seen pilot frequency errors of 5 to 16 Hz caused by the studio ASI multiplexer.  Our DTV exciters are 5 years old, so there must be other"vintage" exciters out there that behave like ours.

I would like to hear any feedback you may have, including exciter manufacturer and model.  Please take the time to e-mail me.


Rockley Curless
KPBS - RF Engineering
RCurless (at) KPBS.org

Posted by Steve Blodgett
Earthsignals.com