Programmable Devices
CPLDs, FPGAs, SoC FPGAs, Configuration, and Transceivers
21611 Discussions

VCO post-scale problem

Altera_Forum
Honored Contributor II
1,203 Views

I’m using the alt pll reconfig IP to generate different output frequency and confonting a problem which puzzles me long. 

The output frequency of PLL can be calculated as follows: 

Fout=F_in× (M/(N×C)) 

In which: 

F_in is the input frequency; 

M is the Feedback counter (M); 

N is the Pre-scale counter (N); 

C is Post-scale output counters (C0–C4) 

 

The inner diagram of pll: 

1.jpg 

 

The blue circles check out the M,N,C, please notes the red circle which checks out the VCO post-scale counter K. And I’ve got some doubts about this VCO post-scale counter K. 

 

Let 50Mhz clock as input clock and 126Mhz as output clock. Now the parameters is like below: 

2.jpg 

 

 

We can know from the picture that  

M=63,N=5,C=5 

And the output frequency is:50*63/(5*5)=126Mhz 

Please note the red circle which checks out the VCO post scale counter is 2. 

 

In some other case the I set the output frequency is 138Mhz, and now the parameter is: 

3.jpg 

M=69,N=5,C=5 

The output frequency is:50*69/(5*5)=138Mhz 

But pls look at the picture: 

 

In which the VCO post scale counter is 1 

 

I just want to ignore this VCO post scale counter, but in pll reconfig IP this para is used: 

4.jpg 

 

So,would anybody could tell me how to calculate this VCO post scale parameter?
0 Kudos
1 Reply
Altera_Forum
Honored Contributor II
504 Views

 

--- Quote Start ---  

 

So,would anybody could tell me how to calculate this VCO post scale parameter? 

--- Quote End ---  

 

 

IIRC, Cyclone IV has a nominal VCO range of 650 to 1300 MHz. When the VCO post scale counter is set the VCO range is divided by 2, i.e. 325 to 650 MHz. So if your calculated VCO frequency is >= 650MHz, use 1, if it is smaller, use 2. The interesting thing is that this has no influence on the calculation of the other counters. This is all from top of my head, so I hope I got things right, but I am quite sure that 650MHz was the "magic" limit. 

 

Regards, 

 

Thomas 

 

www.entner-electronics.com - Home of EEBlaster
0 Kudos
Reply