crystal oscillator 起振后的怪现象

DukeLeon

进来学习，我的project也要用到同样的Pierce Crystal Oscillator

vera_lyh

Thank you for all your replies.
The size of the inverter is NMOS: 4*48/0.7 PMOS: 4*9/0.7. I know the ratio of their size is a little strange, but otherwise I couldn't get a VCM near the Vdd/2.
Some of my colleagues thought that the reason may come from the discontinuity of the model.

dongzz201

NMOS --- 48/0.7, M=4;
PMOS --- 9/0.7, M=4;

Maybe you should decrease the size of NMOS & increase the PMOS size.
For example, let NMOS size : 4*12/0.7, & PMOS : 4*24/0.7

you can try.

vera_lyh

Thanks.
I will have a try.

superrabbit

应该是PMOS的宽长比大些的~~



"
hiNMOS --- 48/0.7, M=4;
% k. K6 t4 T- ?" E0 NPMOS --- 9/0.7, M=4;
+ N* w" y% w/ ~% [" l0 x" [  I& M3 B0 k' i6 _) J$ a
Maybe you should decrease the size of NMOS & increase the PMOS size.' Z9 O9 k/ [) Z# Z5 W
For example, let NMOS size : 4*12/0.7, & PMOS : 4*24/0.7 : K7 h) o9 ~8 o% p/ T
- K- d/ n# x% L% j' {, R! F/ k
you can try.
"

vera_lyh

Thank you.

prof3

To tell the truth, your crystal ciruit design is still at very low level because
(1) Inverter based XTAL circuit is very power consuming and distorted.
(2) Inverter based XTAL circuit has very bad frequency stability of oscillation and therfore high jitter
(3) Inverter based XTAL circuit also impacts on XTAL life( too high DL)

only advantage is --- simple


May I suggest
(1) Still keep pierce type, but change to class A amplifier
(2) I understand you wanna keep VCM at VDD/2, this is good. Few people has this idea. You can do it using another DC loop to make it instead of using abnormal PMOS/NMOS ratio, just as you did now.
(3) If you wanna more complicate, I also sugget you add amplitude control.

If you did all of three, you are now at least entering advanced level.

Of course, you don't care about the performance. Pls, just keep inverter!

vera_lyh

Thank you very much.

All suggestions are very good points. I'll try them.

vera_lyh

BTW, what's DL?
Thanks!

haibianxia

好像是MOSCAP有寄生的电阻，影响了反馈回路