I wonder, what's the best oscillator scheme to use with watch crystals(32kHz) to get optimal stability & precision? (provided that temperature is stable, digital correction is out of this topic).

Typical solution is pierce oscillator, but what's the optimal swing/voltage? I have a feeling that sine-wave generator should put less stress on the crystal.

So, what's the best & less stressful oscillator scheme for watch crystals?

That is for digital clock, so long-term stability is important (crystal will be temperature-stabilized & corrected for aging).

  • \$\begingroup\$ this seems like a heavy opinion question. What is better? What is optimal stability & precision? What defines best and least stressful? \$\endgroup\$
    – Kortuk
    Commented Jul 17, 2011 at 1:59
  • \$\begingroup\$ @Kortuk Well, this seems straightforward to me: The more stability the better. I guess driving crystal at excessive swing or improper load or improper waveform might cause it to age faster. \$\endgroup\$ Commented Jul 18, 2011 at 1:22
  • \$\begingroup\$ so cost plays no part? the best stability and life and we are golden? \$\endgroup\$
    – Kortuk
    Commented Jul 18, 2011 at 7:38
  • \$\begingroup\$ Yes, as long as we are not going over 1000$ in parts :-D \$\endgroup\$ Commented Jul 18, 2011 at 7:56
  • \$\begingroup\$ With that kind of budget ($1000), you can do some fancy things. Temperature control is well within range. Can you get a used HP cesmium clock for that money? \$\endgroup\$ Commented Jul 18, 2011 at 12:37

1 Answer 1


You don't say what the device this is going into. However, if there will be a micrcocontroller present then you likely don't need a separate crystal driver at all. Many microcontrollers (and just about all Microchip PICs) come with low power watch crystal oscillators. You add the crystal and two caps, and the micro does the rest.

Yes, driving the crystal with a sine wave is the best for the crystal.

  • \$\begingroup\$ That is for digital clock, so there will be micro. The question is can it be done better than micro. \$\endgroup\$ Commented Jul 16, 2011 at 22:50
  • \$\begingroup\$ @Bars: "Better" is in relation to some spec. If better includes low parts count, then probably no. \$\endgroup\$ Commented Jul 16, 2011 at 23:02
  • \$\begingroup\$ Surely, it's hard to do better than 0. Better = better long-term stability, less stress on the crystal \$\endgroup\$ Commented Jul 18, 2011 at 1:21

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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