ΣSigmaStudio Ask AI Русский Log in

ADAU1701 12.288MHz clock don't always start=

SigmaDSP & SigmaStudio Korey Oct 1, 2012 00:00 6 replies Original on forum ↗
Subscribe
Korey Question
Oct 1, 2012 00:00

Hi,

I'm using a ADAU1701 with a 12.288MHz clock and sometimes it starts and other times it don't. I have a osc crystal with a 7pf shunt capacitance and using 22pf and 100 ohm resistor as the datasheet requires. The PLL_LF has the right components as per spec also. The clock will sometimes come up and other times it don't. Sometimes it comes up and then drops out.

Any light you can shed on the issue would be helpful.

Thank you for your time and best regards.

Korey

Audio Signal Processors

SigmaDSP Audio Processors

adau1701

KJBob Reply
Oct 2, 2012 00:00

Hi Korey,

One of Murphy's famous laws states, "A self-starting oscillator won't." Crystal oscillators seem to be the worst of the bunch, they get downright fussy. Brett G. posted an example of a PCB layout that appears perfectly fine yet proved unreliable at https://ez.analog.com/message/33073#33073

You'll find quite a theoretical treatment of this subject (perhaps more than anyone wants to know!) in (sorry) Microchip's app note AN849, which you can view here:

http://ww1.microchip.com/downloads/en/appnotes/00849a.pdf

Bob

Korey Reply
Oct 2, 2012 00:00

Thanks Bob,

I have vias at the caps so I will give this a try.

Korey

ADIApproved Reply
Oct 2, 2012 00:00

I have seen other cases where poor grounding or power supply bypassing caused excess noise on the chip supply which makes it difficult for the oscillator to maintain a constant frequency and for the PLL to lock. Solid ground and tight power supply bypassing can help a great deal in resolving these kinds of problems as well. Here's a list of a few best practices to follow: http://ez.analog.com/docs/DOC-1889

I agree with Bob that the vias could be causing reflections that are creating clock edges where they should not be, which could further "confuse" the system and prevent it from starting up reliably.

Korey Reply
Oct 2, 2012 00:00

THank you both for your input. I have cut the vias and installed jumpers directly to the regulator ground and the clock seems to starting consistantly. I also changed to a 27pf and 33pf to put my load capacitance above 18pf with additive stray capacitance. This was a little confusing on the datasheet since it shows 22pf and the would give 10pf load plus the stray capacitance. I did see a post that kind of addressed and passed it to the particular xtal being used. So, in summary, I:

1) Tied the 'rock' load caps to ground directly without going through vias and ensured no sharp (90 degree) bends in any traces.

2) Changed to a 7pf shunt 'rock'

3) Changed the load capacitors to meet load capacitances of 18-25pf and also offset them per xtal tips.

Thanks for your help. We will respin and see how it goes!

Korey

ADIApproved Reply
Oct 11, 2012 00:00

Excellent news. Thanks Korey. Please let us know how the respin works.

EZAdministrator Reply
Aug 2, 2018 00:00

This question has been assumed as answered either offline via email or with a multi-part answer. This question has now been closed out. If you have an inquiry related to this topic please post a new question in the applicable product forum.

Thank you,
EZ Admin