Article View: sci.electronics
Article #15791Re: Clocks
From: Glenn A Sullivan
Date: Sat, 02 Jan 1993 05:52
Date: Sat, 02 Jan 1993 05:52
43 lines
2720 bytes
2720 bytes
flux@sys6626.bison.mb.ca (John Kamchen) writes: ......... > What I need is a 14Mhz clock that doesn't jitter the length of a video > line. > An LM1881 provides the syncs, and the clock should be stable as possible. > Looks like this > --____---------------------------------------------- > _-____-_-_-_-_-_-_-_-_-_-_-_- and so on this is the 14Mhz clock > Does not jitter over the length of a video line! Assume less than 5 nanosec jitter out of 62Usec line length. This gives about 5 picoseconds jitter per cycle. Basically we need an oscillator with LOW PHASE NOISE, since phase noise is Frequency Modulation (PM and FM are related via integration). You need a high-Q tank circuit that IS >>>> NOT OVERDRIVEN <<<<. So use a transistor oscillator, with biasing adjusted to keep the signal swings well away from saturation or cutoff. And find a way to keep the maximum signal from building up and up, which would distort. With "linear" amplification that is barely barely adequate to oscillate, any off-frequency energy (phase noise) actually gets ATTENUATED on each pass around the tuned loop. HP uses a nice trick in the LC oscillators for their 3330 synthesizer: (1) they use a differential pair to form the oscillator, with the emitter resistor providing a very stable current (2) the oscillator is LOADED intentionally with a resistor, so that as the signal swing increases as stored energy increases, eventually the resistor is absorbing any additional energy dumped into the tank. Thus the operating swing is fairly well defined. Since HP hangs this resistor upon the tank circuit, and resistors themselves generate noise, you want a low value resistor to avoid directly inducing lots of Boltzmann/Johnson/thermal noise (phase noise) into the tank. So use series capacitors ( 1.1*C in series with 10*C) with the damping resistor hanging across the 10*C capacitor. Just when you thought analog design was totally passe', and that DSP methods would replace all!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! Allen Sullivan analog design expert, video engineer, RF trainee, system analysis practitioner, DSP algorithm inventor, and hiker to where 300 foot waterfalls .....fall. Arizona in the winter!
Technical Details
Message-ID:
<1993Jan2.055211.26448@ennews.eas.asu.edu>Path:
arc1-ca.pugleaf.net!sparky!uunet!gatech!asuvax!ennews!enuxha.eas.asu.edu!gsullivaReferences:
<e26gwB4w165w@sys6626.bison.mb.ca>Show Full Headers
Complete Article Headers
Newsgroups: sci.electronics Path: arc1-ca.pugleaf.net!sparky!uunet!gatech!asuvax!ennews!enuxha.eas.asu.edu!gsulliva From: gsulliva@enuxha.eas.asu.edu (Glenn A Sullivan) Subject: Re: Clocks Message-ID: <1993Jan2.055211.26448@ennews.eas.asu.edu> Summary: Building a LOW-JITTER 14MHz oscillator Sender: news@ennews.eas.asu.edu (USENET News System) Organization: Arizona State University References: <e26gwB4w165w@sys6626.bison.mb.ca> Date: Sat, 2 Jan 1993 05:52:11 GMT Lines: 45