1
\$\begingroup\$

I am trying to build a simple 32kHz crystal oscillator but I am seeing something strange. I am using a simple circuit in that the crystal is in feedback around a current starved inverter.

However, I am seeing something strange. The frequency of operation is correct but at edge of the clock, there a multiple transitions before it finally stays high (or low). I.e. the voltages are bouncing"

scope output

\$\endgroup\$
11
  • 1
    \$\begingroup\$ Post links to oscilloscope output and to schematic and we'll edit them into the post. \$\endgroup\$
    – AndrejaKo
    Jun 1, 2012 at 17:26
  • 1
    \$\begingroup\$ That behavior is called 'ringing', if i interpret your description correctly. \$\endgroup\$
    – dext0rb
    Jun 1, 2012 at 17:37
  • 2
    \$\begingroup\$ Also, please post a photograph of your circuit and scope probe. \$\endgroup\$
    – markrages
    Jun 1, 2012 at 17:56
  • 1
    \$\begingroup\$ I'm not confident enough to post this as an answer, but empirically I have found 32kHz crystals to be extremely sensitive to additional load capacitance and/or any form of resistive load. \$\endgroup\$ Jun 1, 2012 at 18:13
  • 1
    \$\begingroup\$ I would say its not ringing but chatter due to slow rising edges on inputs to your inverters. Are you using Schmitt-trigger inverters? A schematic of your complete circuit is needed to give any more certain analysis. \$\endgroup\$
    – The Photon
    Jun 1, 2012 at 22:33

1 Answer 1

1
\$\begingroup\$

If the metal can of oscillator is not grounded, then it works as a capacitive load parallel to crystal. The MHz bursts are probably matching the time delay of second positive feedback of invertor when this capacitance is involved.

http://www.telluriantech.com/Effects_of_Grounding_on_Ceramic_SMD_Crystal.pdf

Replace crystal with 2-10pF variable cap and see if the oscillations will match the frequency of bursts.

Answer: Ground the can. It will move bursts to GHz. Also the crystal pins can be grounded through small capacitors.

\$\endgroup\$
2
  • \$\begingroup\$ Thank you for the response. I am assuming that by can you mean the housing of the crystal...correct? There already pulling capacitors in parrallel with the crystal that go to ground. \$\endgroup\$
    – user10090
    Jun 2, 2012 at 0:58
  • 1
    \$\begingroup\$ Yes. The housing. Another mistake can be driving the crystal with too high power. The typical power should be in microwatt range. The crystal can be destroyed mechanically by milliwatt power. \$\endgroup\$
    – user924
    Jun 2, 2012 at 1:07

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.