Author Topic: Do I need load capacitors for XTAL with CL 4pF to 10pF?  (Read 834 times)

0 Members and 1 Guest are viewing this topic.

Offline ttodorovTopic starter

  • Regular Contributor
  • *
  • Posts: 67
  • Country: us
Do I need load capacitors for XTAL with CL 4pF to 10pF?
« on: April 28, 2018, 08:04:13 am »
Hi,

I am choosing a quatz crystal for a design I am working on and looking at Mouser and DigiKey, I see a lot of options with load capacitance of 10pF and below. What I am used to are crystals with CL of 18 to 22pF and these were easy: CL=C/2-Cstray, with recommended values for stray capacitance from traces and the like around 10pF. But with the new ones I am left scratching my head :-//

Do I need any ceramic caps with CL of the crystal at 10pF? And if the stray capacitance of the board is 10pF, how do I use a crystal with CL=4pF?

I looked for application notes for crystals, but couldn't find anything specific or new enough (on sites like Abracon). I would appreciate some help, if anyone can answer that... Thx.

Best regards!
 

Offline Ian.M

  • Super Contributor
  • ***
  • Posts: 12860
Re: Do I need load capacitors for XTAL with CL 4pF to 10pF?
« Reply #1 on: April 28, 2018, 08:51:50 am »
You *ALWAYS* need the recommended load capacitance, whether as physical capacitors or as PCB trace capacitance.   Also in the classic single gate Pierce oscillator configuration, the capacitors are essential to provide 180 deg phase inversion across the crystal.   

With Cl=4pF, you need 8pF either end of it as the load capacitors are in series from the crystal's viewpoint.  As double sided 1.5mm FR4 will give you about 5pF* per cm2 of trace area above a ground plane, neglecting capacitance between traces, that's easily achievable if you keep the crystal layout tight near the IC pins.   If you come in under the target capacitance, just expand the crystal pads, but limit their solderable area with soldermask.

* varying the geometry from long and skinny to short and stubby will significantly affect this value, I got it from a microstrip calculator that includes fringing fields, using a 2mm trace width.   Your crystal traces will be mush narrower but 2mm isn't unreasonable as an average including the pads.
« Last Edit: April 28, 2018, 08:59:22 am by Ian.M »
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf