The Designer's Guide Community Forum

The Designer's Guide Community Forum
https://designers-guide.org/forum/YaBB.pl
Design >> Analog Design >> PFD characteristics
https://designers-guide.org/forum/YaBB.pl?num=1259664998

Message started by ReadWDX on Dec 1^st, 2009, 2:56am

Title: PFD characteristics
Post by ReadWDX on Dec 1^st, 2009, 2:56am

hi,all
now I want to design a RF PLL, when I design PFD, I do not know how to simulate PFD characteristical curve,some has done it? I need help, thanks!

Title: Re: PFD characteristics
Post by Mayank on Dec 1^st, 2009, 3:56am

Hello RDX, :P

On a standalone basis, you can simulate your PFD so that the UP & the DNbar or UPbar & DN signals math exactly....ie transition at vdd/2.

The Best way to characterize Phase-detector ckt of a PLL is to plug the PFD and Charge-Pump together.
1. Give two signals on input freq and fed-back Freq with a certain difference in Time-period & that difference as the delay between the clocks transitions. ---> The idea is to basically produce transitions with decreasing/ increasing difference b/w ref. freq node & PLL fed-back node.
2. Now measure the CPUMP current corresponding to each commparison. You will get a staircase kind of response...
3. This will help you estimate the gain of your P-D ckt as well as Static Phase Error.
4. If you want to obtain the gain plot of PD ckt, vary the delay b/w the two clocks at ref & PLL fed-back node and measure the Cpump current o/p. Plot cpump current o/p against the phase error (360deg * delay / Tcomparison) -- you will obtain a K_PD plot as shown in books...My belief is that you however generally dont need that plot.

thanx
mayank.

Title: Re: PFD characteristics
Post by ReadWDX on Dec 1^st, 2009, 11:05pm

thanks Mayank very much

I have done as you suggest,I got the result is followed:
in the plot ,Iout is average of the Charge pump output cuurent,
If you can tell me the problem in my result,I would be grateful!
[img][/img]

Title: Re: PFD characteristics
Post by raja.cedt on Dec 2^nd, 2009, 12:05am

hi,
seems like your circuit is correct but may be you have given wrong sweep statement. by the what's your Icp? is it 180uA

Thanks,
Rajasekhar.

Title: Re: PFD characteristics
Post by Mayank on Dec 2^nd, 2009, 12:55am

Hello RDX,
The shape of the plots seem matching, so i guess your ckt functioanlity is correct....How are you calculating the avg current ?? & What is your X-axis (Delter) and in what units (if phase, degrees/readians) ??

mayank.

Title: Re: PFD characteristics
Post by raja.cedt on Dec 2^nd, 2009, 1:33am

hi,
i feel sweeping in the -ve direction is wrong and are you finding average over enough cycles.

Thanks,
Rajasekhar.

Title: Re: PFD characteristics
Post by ReadWDX on Dec 2^nd, 2009, 5:53pm

Thanks for all giving me avaluable advice!
in my design ,I take the charge pump current is 180uA,and the average current is calculated by " average"function in cadence calculator,the "delter" is the delay time of input signals of PFD,unit is secend.
when I calculated the average of current,I took 80 cycles ,Is it enough?

Title: Re: PFD characteristics
Post by raja.cedt on Dec 2^nd, 2009, 7:48pm

hi,
i feel 80 cycle's are too much.ok any how it's not a problem but have checked abut sweep comand once..i feel there is some problem.

Thanks,
Rajasekhar.

Title: Re: PFD characteristics
Post by Mayank on Dec 2^nd, 2009, 8:56pm

Hi,
I think the plot is a bit shifted....Ideally the plot should be a piecewise linear plot b/w +pi to --pi on x-axis if your x-axis is in radians. What is your time period 28ns-30ns ??

Can you attach the staircase response of charge-pump o/p voltage node with the i/p & fed-back pulses ?

thanx
mayank