The Designer's Guide Community Forum

The Designer's Guide Community Forum
https://designers-guide.org/forum/YaBB.pl
Simulators >> RF Simulators >> questions about the transient response of a pll
https://designers-guide.org/forum/YaBB.pl?num=1183472852

Message started by calven on Jul 3^rd, 2007, 7:27am

Title: questions about the transient response of a pll
Post by calven on Jul 3^rd, 2007, 7:27am

hi!
i am a freshman in pll design.i have designed a pll in transistor level.the simulation tool is spectre.the voltage domain model corresponding to the first waveform is implemented with veriloga presented as a comparision,and the simulation time is 30u.the circuit corresponding to the second waveform is implemented with transistors,and the simulation time is 3u.according to the waveform,we can see that the waveform settles to a ripple(and the curve is not a line,but at some instant the curve has different values).it is more obvious in the third figure,which is a part of the second figure zoomed out.what cause this?
additionally,the ripple will last for a long time.i have simulated the circuit with the simulation time of 50u.the amplitude of the ripple maybe decreases,but not apparently.however,the ripple of the veriloga model decrease quickly.the veriloga model and the transistor circuit basiclly have the same parameters,and what is the reason?
give me some explanation and advice,please.thanks a lot.i am sorry that i can not upload the image bigger
best regards

Title: Re: questions about the transient response of a pl
Post by calven on Jul 3^rd, 2007, 7:55am

the second figure is as follows:

Title: Re: questions about the transient response of a pl
Post by calven on Jul 3^rd, 2007, 7:56am

the third figure is as follows:

Title: Re: questions about the transient response of a pl
Post by Ken Kundert on Jul 3^rd, 2007, 2:06pm

First you should ask yourself: what is the source of the ripple? Then you should ask if that source is modeled in the Verilog-A model. Perhaps the ripple is due to a dead zone exhibited by the phase detector. If the verilog-a model did not include a dead zone, then you would not see the ripple.

-Ken

Title: Re: questions about the transient response of a pl
Post by gavi on Aug 14^th, 2007, 8:03am

Hi,

What kind of phase detector have you used?

Best regards,
Gavi

Title: Re: questions about the transient response of a pl
Post by cmc on Aug 14^th, 2007, 12:01pm

It is useful to simulate the PLL components individually, and then combine them one at a time in an open-loop sense and checking that the intermediate and output signals are what you expect. This way you can pinpoint sources of ripple, isolated from the loop dynamics.