crystal oscillator 起振后的怪现象

vera_lyh

最近在做一个晶体震荡器,在spectre里仿真时,电压已经稳定,但电流却不断增加.这是怎么回事?

在hspice里面仿真时,可以起振,但起振后,放大器输出与晶振连接的地方振荡的中心电平会逐渐变低,导致后面的反向器无法翻转.设计要求负载电容用moscap在片内实现.用理想电容时,不会有这种现象.请问如何解决这个问题.

谢谢!

jonus

沒有附圖難以瞭解Crystal是RLC model?
是用什麼樣的放大器？
放大器的工作點是在1/2VC
Gain,  Driving current...enough?

vera_lyh

首先谢谢楼上的答复.由于我们内网和外网是分开的,自己不能抓图.只能描述一下了.

Crystal是用RLC model的.放大器是两个管子做成的反相器.
工作点是在1/2 vdd
工作电流和增益应该足够了.因为可以起振.只是起振后反相器输出的电平会下降.

vera_lyh

电路结构如下图

其中,Rf由传输门构成,C1和C2由moscap形成,晶体由RLC来建模.

intercon

CRYSTAL 是多少兆的

prof3

1. Check amplitude of V(C1) and V(C2). Is rail to rail?
2. Check CM of V(C1) and CM of V(C2), Is that equal?
3. Check amplitude of V(C2)  see if stable( means no more increasing already)
4. Normally, current will be up and down, up and down and stable in average.
5. If properly design, MOSCAP and Ideal CAP does not have that much difference in current but they do  impact the stability of oscillation frequency since CV of MOSCAP.

vera_lyh

谢谢各位的答复.
crystal是12M的.
vc1 and vc2 are not rail to rail (in hspice), and are both rail to rail and even larger than rail to rail (in spectre)
CM of vc1 and vc2 are first equal, but suddenly one increase and the other decrease.
here are some pictures, but they are not all, I will provide the other later.

the top curve is the vc2, the middle curve is the output after buffers, the bottom curve is the total current in spectre

the top curve is the output after buffer, the bottom curve is the total current, and the transient current has been more than 100mA (in spectre)

the top curve is the output after buffers, the bottom curve is the total current in hspice
Thank you very much!

vera_lyh

simulation results in hspice

                                                tt

                                       SS

thank you very much!

rainheart

这应该是个pierce 振荡电路吧

prof3

1. What is the size of the two inverter?
2. Enlarge the waveforms after 7.5ms (this is ridiculous big in current) and see what the waveform looks like