5
\$\begingroup\$

There are many different electrical clocks for generating pulses with constant frequency (e.g. RC oscillators, crystals, OCXOs, MEMS oscillators, chip-scale atomic clocks).

If they are mounted on a vibrating device (e.g. an electrical motor of a car toy or drone), is their performance or synchronization influenced?

Which ones are more suitable if we have several of them that need to be completely synchronous and not influenced by vibration?

\$\endgroup\$
5
  • \$\begingroup\$ this may be an XY question ... asking about a perceived solution to an unspecified problem... what problem are you trying to solve? ... why do you need the clocks to be completely synchronous? \$\endgroup\$
    – jsotola
    Jan 18 at 18:31
  • 1
    \$\begingroup\$ @jsotola I read it more as, wondering if it's a problem at all, and if so, by how much. So, good answers would include information on vibration (acceleration and frequency), and in turn, what kinds of vibration (same metrics) typical devices can exhibit. Is that an accurate assessment, @azerila? \$\endgroup\$
    – Tim Williams
    Jan 18 at 19:10
  • 2
    \$\begingroup\$ I have a fun related anecdote here: my dad once built a device with an RF transceiver that was attached to the rotating driveshaft of a locomotive engine. When the drive motor span up, the center frequency of the signal transmitted by the device shifted significantly - far more than simple Doppler effects would suggest, and enough that it caused the RF link to be lost. The cause was the physical force applied to the onboard crystal resonator. They ended up having to widen the bandpass on the receiver to compensate for it. Probably not that likely in your scenario, though! \$\endgroup\$
    – Polynomial
    Jan 18 at 19:12
  • \$\begingroup\$ @TimWilliams yes, and how much the influence can be, or suggestions on electrical clocks that may be less influenced by vibration. \$\endgroup\$
    – azerila
    Jan 18 at 20:46
  • \$\begingroup\$ Independent, but "completely synchronous", clocks are not something you're going to achieve without some external mechanism to keep them in sync, unless you're thinking about building an atomic clock... \$\endgroup\$
    – brhans
    Jan 18 at 22:05

2 Answers 2

Reset to default
4
\$\begingroup\$

I would expect higher frequency crystals (MHz) to be much more immune than low frequencies (tens of kHz). Vibrations from machinery etc. don't tend to be extremely high frequency so the effect during a given cycle of the oscillator is small. However 'small' is not enough in applications with high vibration and sensitivity to phase noise such as some radar applications.

Many, if not most, consumer clocks use 32.768kHz tuning fork crystals because they are cheap and the oscillators can easily be made to be low power, even with 1970s technology. There does not appear to be a lot of information out there on their sensitivity. They are known to be prone to damage in ultrasonic cleaners and have a significant moving mass so I would expect high sensitivity to a narrow spectrum and in certain axis.

There exist MEMS oscillators which use a much higher frequency mechanical oscillation and divide down the output to emulate a 32kHz oscillator, however even they caution against ultrasonic cleaning of the product. Modern IC technology has yielded products that run on very low current as well (albeit with some other environmental sensitivity as a result).

If you need clocks to be 'completely' synchronous, you can provide synchronization by communicating between them (eg. via a wire or via some protocol such as PTP over a communication channel) or by using a common clock source or synchronization source such as from a GNSS ( Global navigation satellite system) system.

\$\endgroup\$
2
  • \$\begingroup\$ Is there a reason high frequency crystals may be more immune? \$\endgroup\$
    – azerila
    Jan 18 at 19:08
  • 1
    \$\begingroup\$ Yes, explained in the second sentence of the first paragraph. \$\endgroup\$
    – Spehro Pefhany
    Jan 18 at 19:09
2
\$\begingroup\$

Yes, at least crystals and ceramic resonators are sensitive to shocks and vibrations. Some capacitors exhibit piezoelectric effects as well, and as they convert mechanical vibrations into electrical signals, any circuit with such a capacitor on sensitive oscillator path will affect the oscillator.

Based on your previous question, you need 1 us accuracy in 5 hours which is about 0.06 PPB.

\$\endgroup\$
1
  • \$\begingroup\$ Even "benign" components like resistors can be affected by vibration/shock. Any component on a PCB that suffers even microscopic movement can affect a critical circuit capacitance and alter oscillation frequencies. Depending on how stable, accurate, synchornized the clocks need to be, there can be many devils in the details. \$\endgroup\$
    – AndyW
    Jan 18 at 20:30

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.