Continue to Site

Welcome to EDAboard.com

Welcome to our site! EDAboard.com is an international Electronics Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.

Question about loop filter of CDR/PLL

Status
Not open for further replies.

usuikazkou

Junior Member level 2
Joined
Oct 4, 2016
Messages
20
Helped
1
Reputation
2
Reaction score
0
Trophy points
1
Activity points
166
Hi,
22.PNG
As I know this type of loop filter is very common, it can found from lot of textbooks and thesis.
eetop.PNG
But what is the purpose of above graph(a connection to node between resistor and capacitor)?
Is the transfer function still same as 1 path current injection? if not how to find the transfer function?
I have search alot of paper and textbook and I didn't found any discussion about it.


Thank you.
sorry for my poor english.
 

We need more information on the application to be certain but I would guess C1 is the charge reservoir from the lower charge pump and the R1/C2 are a simple RC filter. In general, two charge pumps driving each other doesn't make sense so that 'enable' signal must have some control over which is producing the error voltage. The topology is quite different to the top schematic.

Brian.
 

Having frequency and phase lock-loops with different transfer characteristic can make sense to improve the lock behavior. Building it with two charge pumps is just an implementation detail that shouldn't be overrated. Presume you know what's the difference between frequency and phase detector and what's their purpose in a PLL.

Without guessing too much, you can assume that both charge pumps are activated mutual exclusively, the frequency detector signal is only active in unlocked state and the phase detector in locked state. A standard PLL with combined phase/frequency detector can be expected to activate the charge pump up/down outputs in unlocked state with 100 % duty cycle. The present design does the same, but has the option to tune the frequency detector transfer characteristic differently.

To find the transfer characteristic, you need the quantitative parameters of all blocks and components.
 

It seems that frequency detector here is used to bring two signals close to each other. When frequency difference is small then phase detector is enabled to make a phase lock.
 

We need more information on the application to be certain but I would guess C1 is the charge reservoir from the lower charge pump and the R1/C2 are a simple RC filter. In general, two charge pumps driving each other doesn't make sense so that 'enable' signal must have some control over which is producing the error voltage. The topology is quite different to the top schematic.

Brian.

Thanks you for the reply,

pic.1
c4.PNG
pic.2
c3.PNG
pic.3
c5.PNG
pic.4
c6.PNG

I would like to design a clock and data recovery circuit, and I am confused by the dual loop system. Usually phase locked loop text book only discuss about single loop system as we see from pic.1 and the filter function is pic2. I would like to try to do a loop analysis about pic3's dual loop clock and data recovery architecture.
I have done single loop pll behavioral simulation using matlab (Pic.4).
I have no idea how to do the simulation using architecture pic.4, which have 2 charge pump connected to difference node on a single loop filter ,
and what the loop filter function should be ? (with 2 current injected into the loop filter, lets say I1 for PLL and I2 for FLL)
any reference about this?

reference:
pic3 taken from https://ieeexplore.ieee.org/document/4415629

thanks you.
 

I agree with the loop filter analysis, except for the erroneous R2 which should be R1. Why can't you perform the same analysis for the charge pump feeding C2 in this circuit? You get a similar second order transfer function, but without a zero.

Frequency detector mode and switching between FD and PD mode requires a state machine, also in the single loop PFD configuration and can't be modelled by a simple linear transfer function.
 

Yeah, the R2 should be replaced by R1 which looks like the pic below (taken from the Dean Banerjee PLL).
c7.PNG


Why can't you perform the same analysis for the charge pump feeding C2 in this circuit? You get a similar second order transfer function, but without a zero.
Sorry I don't get the point, do you mean I can run the loop analysis of the dual loop system as one loop system?

Thank you
 

As previously stated, the charge pumps will be activated alternatively, you can calculate transfer functions for both detectors separately.
 

As previously stated, the charge pumps will be activated alternatively, you can calculate transfer functions for both detectors separately.

I will try to do the analysis in this way
Thank you.

any alternative methods or ideas are welcome
 

Status
Not open for further replies.

Similar threads

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Back
Top