4
\$\begingroup\$

I'm using a 32kHz crystal oscillator with timer1 on a PIC16F1825 to provide more accurate timing for events (I want to do something every 20 minutes). The crystal circuit looks easy enough to design but I'm not sure whether a resistor is necessary to reduce the drive level. I'm running at 3.0V.

enter image description here

I'm using an ABS25-32.768KHZ-4-T crystal from Abracon as the crystal. It has a CL of 12.5pF. Thus I figure C1 and C2 at 18pF. The problem is Rs. The crystal maximum drive level is specified at 1uW and the ESR is 50k but I cannot figure out if the PIC (in LP mode) will overdrive the crystal if Rs is zero. Can I assume that LP will be the "right level". I've seen a number of circuits where Rs is 100k and a number where it isn't present. Thus my confusion.

\$\endgroup\$
2
  • 1
    \$\begingroup\$ ti.com/lit/an/szza043/szza043.pdf page 13 onwards \$\endgroup\$
    – Andy aka
    Commented Oct 31, 2015 at 12:27
  • \$\begingroup\$ Interesting, thanks. I will admit there's a certain amount of laziness involved in the question - ie: I know I should qualify my own circuitry but I'm being made do the PCB asap! I will put a space for Rs on the PCB and try different values. \$\endgroup\$
    – carveone
    Commented Oct 31, 2015 at 13:25

1 Answer 1

Reset to default
1
\$\begingroup\$

The effective output resistance of the internal cmos gates has not been stated .So I suggest you leave provision for the output resistor .Remember that a resistor has much better tolerances than the internal gates .If you are doing a big prod run the validity of getting it right is great .If you do overdrive the Xtal you could blow it up like you said ,I havnt managed this on 3V3 .You could get lousey accuracy ,3rd overtone operation which will really screw things up.More subtle you could get unreliable starting because the osc needs some of the phase shift produced by C2 and the total effective resistance .If you are a bit worried about all this then check the osc waveform on a scope .Too much feedback tends to square things up and too little makes the sinewave too perfect.

\$\endgroup\$
1
  • \$\begingroup\$ That was it exactly - I don't know the levels from the PIC. I don't see them in the datasheet. I'm loath to just guess so I'll have to do as you say and check the output on a scope (adjusting for scope capacitance I suppose). I was kinda vaguely hoping this was a pre-solved problem but I do believe that I will need the resistor. Now to squash it in there on the PCB! \$\endgroup\$
    – carveone
    Commented Nov 1, 2015 at 16:42

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.