2
\$\begingroup\$

As part of a project I am working on, I need to bring up a 4060-based 32.768kHz crystal oscillator circuit and ensure it works without breaking crystals or having cold start issues. I can build a board that has pots on it for adjusting Rs and Rp in the schematic below, but should I just pick load capacitors that match my 12.5pF/1µW 32.768kHz tuning-fork crystal (a rather generic Raltron part, if you're deathly curious), or should I add trimmers for the load capacitors as well? Should I be focusing on only one resistor, or taking the resistor values as givens, for that matter?

Also, are there other tests I should do during the bring-up process besides making sure it starts up at cold temps and works across the supply voltage range I'm dealing with without drifting excessively or frying crystals with excessive drive? I don't have an active or other ultra-high-Z probe, either, so processes based on direct crystal-node measurements are not useful for me.

schematic

simulate this circuit – Schematic created using CircuitLab

Bonus: the MC14060 datasheet specifies a 20pF and an 82pF load capacitor (as shown in the schematic above) for making the 4060 oscillate with a watch crystal. Given a 12.5pF watch crystal, though, this seems like an awful lot of load capacitance to me, but my understanding of this sort of thing is somewhat limited. What's going on with their suggestion here?

\$\endgroup\$
3
  • \$\begingroup\$ Maybe this will help. \$\endgroup\$ – Andy aka Mar 8 at 10:06
  • \$\begingroup\$ What do you mean by "bring up"? \$\endgroup\$ – Hearth Mar 8 at 16:10
  • \$\begingroup\$ @Hearth -- "bring up" as in "get running reliably" \$\endgroup\$ – ThreePhaseEel Mar 8 at 23:31
3
\$\begingroup\$

The effective load is 82pF in series with 20pF or about 16pF, plus stray capacitance, so maybe 18 or 20pF.

That amount of difference will probably cause 20 or 30ppm additional error in your crystal frequency.

It sounds like you've done some reading about it already. You can try lower value capacitors, test at temperature extremes with resistors that are significantly higher/lower than your design values, calculate worst-case dissipation from ESR and so on. Since you don't have the equipment to actually confirm margin on the 1uW maximum crystal dissipation.

If you really need reliability then you might think about just purchasing a packaged 32.768kHz oscillator for about $1. in singles.

\$\endgroup\$
2
\$\begingroup\$

IF you care about tolerances using the Raltron load equivalent circuit Caps (in series) that match your specs and not MOT's. (sorry ON Semi ;)

You may have also noticed 32kHz resonators in stock at D-K are only stocked from 3 pF to 12.5 pF which requires some calculation to choose the C values. The bigger C value parts are old like the MC datasheet and not in stock for other applications besides yours. ( like wide tuning ranges with a Varicap)

\$\endgroup\$

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.