Multiple crystal clock to reduce jitter

[annex666]

I first must start by admitting I have very little knowledge of clock circuit designs so bare with me.

I had an idea to reduce the effect of vibrations on crystals and therefore reduce jitter caused by mechanical vibrations. Although I'm not sure if this idea would have any advantage, or even if it will work I'll explain it anyway.

The idea is to design a clock circuit that uses two crystals that are resonating in phase and to align them on the PCB so that they are in the same plane of oscillation, but are opposing each other.

When the crystals are acted upon by an external mechanical vibration the effect should be less pronounced than a single crystal as the vibration would be cancelled.

It's the same principle as a differential line - common mode noise being cancelled out by a differential input stage.

Will this idea even work?
Do dual (or more) crystal clocks exhist?

[Elso Kwak]

Hi Annex666,
I am not aware of any clock using a dual or two crystal(s).
Any mechanical vibration can be damped with Blue Tac or Poster Buddy. The effect is rather small I feel.......
Wildmonkeysects had an idea for a differential oscillator using one crystal:
http://www.diyaudio.com/forums/showt...5748#post25748

[peranders]

Originally posted by annex666
I first must start by admitting I have very little knowledge of clock circuit designs so bare with me.

I had an idea to reduce the effect of vibrations on crystals and therefore reduce jitter caused by mechanical vibrations. Although I'm not sure if this idea would have any advantage, or even if it will work I'll explain it anyway.

The idea is to design a clock circuit that uses two crystals that are resonating in phase and to align them on the PCB so that they are in the same plane of oscillation, but are opposing each other.

Will this idea even work?
Do dual (or more) crystal clocks exhist?

I think this is a good idea in theory but not in real life. First you must analyze WHAT causes the jitter, then remove or reduce it. Mounting the whole oscillator in some kind of suspension can be a good start and also make it small.

[budsmoker]

i believe frequency jitter in cdps is mainly caused by noise on the power supply affecting the squaring circuit which creates a squarewave from the natural oscillation of the crystal.

[Steve Eddy]

Originally posted by peranders
I think this is a good idea in theory but not in real life. First you must analyze WHAT causes the jitter, then remove or reduce it. Mounting the whole oscillator in some kind of suspension can be a good start and also make it small.

Yup. And ultimately, the only place where jitter REALLY counts is when the final conversion is made by the DAC chip. You can feed the DAC chip the most perfect clock signal one could imagine but that doesn't mean that the conversion will take place perfectly.

It'd be interesting to see how much effect vibration has on jitter though. Be nice to have that Miller jitter analyzer that Stereophile uses which employs a special digital signal fed to the DAC and the analog output is used to produce the jitter's spectrum (which I think is far more telling than some singular ns or ps jitter spec) and suspend a CDP or DAC over a loudspeaker driver and do a frequency sweep and see if any snakes pop up out of the grass.

That's the thing with jitter. Even if two devices have the same basic jitter spec, they can have two wildly different spectra just as two amplifiers might have the same THD+N spec can have wildly different distortion spectra and sound just as different.

se

[Jocko Homo]

The magnitude of any jitter caused by vibrations is going to be much less than any jitter of Gaussian nature that all oscillators have.......not to mention jitter caused by ground bounce and other crap.

You assume that both crystals will react the same way to mechanical vibration.....I really doubt that. They all have an acoustic resonance that varies from crystal to crystal, and I have found that almpst no amount of "goop" stuck to their case has any effect.

They make isolation mounts for crystal in applications where mechanical vibration is a problem.


Jocko

[Elso Kwak]

Hi Annex666,
You could try the famous Jon Risch sandbag trick......
http://db.audioasylum.com/cgi/m.pl?f...ag&r=&session=
and
http://db.audioasylum.com/cgi/m.pl?f...ag&r=&session=

I believe it was Peter Daniel who showed a picture of a Madrigal clockboard suspended by rubber springs on this forum.

[annex666]

Originally posted by Jocko Homo


You assume that both crystals will react the same way to mechanical vibration.....I really doubt that.

Jocko

Like I said I'm no expert, but from what I have read the shape of the crystal defines the frequency of oscillation (depending on the primary mode of oscillation) and therefore they are produced to very tight tolerances - controling their dimensions to effectively "tune" them - would this not make them more inclined to act the same under the same mechanical vibration?

One question I have is this, are there any clock designs that use two crystals in another type of design?