| Author |
Message |
nandu
Joined: 12 May 2005
Posts: 44
Helped: 1
Location: Timbuktu
|
 23 Mar 2008 21:32 DLL transfer function |
|
|
|
Hi all
I would like to know the loop gain of a typical charge pump DLL. Also, I have seen the gain term  being used often, but dont know its exact meaning. Does it depend on the number of delay stages in the DLL? And does  have any upper and lower limits in order to ensure stability. Im guessing it wont because a charge pump DLL will only be a 1st order system and hence wont oscillate.
|
|
| Back to top |
|
 |
CherryQ
Joined: 01 Feb 2008
Posts: 15
|
| 25 Mar 2008 15:04 DLL transfer function |
|
|
|
| No, this is the voltage to delay time gain of your VCDL. You can see its unit is sed/V. If you have N delay cells, each one has the same voltage to delay time gain, so Kvcdl is N times of that.
|
|
| Back to top |
|
|
lijianheng
Joined: 29 Sep 2006
Posts: 133
Helped: 10
|
| 02 Apr 2008 10:37 Re: DLL transfer function |
|
|
|
CherryQ is right. Here is a paper link maneatis96b.pdf about DLL. Hope this helps.
http://www.edaboard.com/viewtopic.php?p=564225#564225
|
|
| Back to top |
|
|
stoned
Joined: 23 Feb 2006
Posts: 52
Helped: 5
|
| 05 May 2008 3:23 Re: DLL transfer function |
|
|
|
| nandu wrote: |
Hi all
I would like to know the loop gain of a typical charge pump DLL. Also, I have seen the gain term being used often, but dont know its exact meaning. Does it depend on the number of delay stages in the DLL? And does have any upper and lower limits in order to ensure stability. Im guessing it wont because a charge pump DLL will only be a 1st order system and hence wont oscillate. |
Hi nandu,
U got it, DLL is 1st order system, so it is unconditional stable. And you should consider you delay time is satisfy you requred in every coner when your VCTRL is changing.
Stoned
|
|
| Back to top |
|
|
