Question on PLL simulation

[ranier]

Hi

I am designing the PLL using 0.13um process.( I used the hspice)
(VCO 150Mhz ~ 300MHz, input 11~22MHz ,PM = about 58 )

I found curious thing, when I simulated PLL.

When I used ".tran 0.1ns 10us' , PLL works well.

But When I used " .tran 0.01ns 10us' , Loop filter voltrage show oscillation.
( so output jitter very high)

I don't know difference of two conditions.

and which condition do I believe 0.1ns or 0.01ns ?

Thank you,
JH.

 

[qqmz]

could the pll work when you simulated at the step 0.01n ?

 

[sunking]

simulate a long time and try

 

[dingjingfeng]

I think the result of the 0.01 is more reliable

 

[borislee]

it is hard to say that 0.1n or 0.01n is more believable if no other information is provided.
try using ".option acurrate"
if the version of hspice is not older than 04.09, ".option runlvl=6" or ".option runlvl=5" can be also tried.

 

[ranier]

Hi

I solved the problem.
I applied VNTOL=1e6 RELTOL=7.5e-04 in the option.

But I still don't know the difference 0.1 and 0.01 .

Thank you.

 

[mpig09]

What is your type of VCO ?
Does the VCO is locked your desired frequency ?
If not, please simulate a long time.

Do you use post=1 or post=2 ?

 

[amarnath]

the difference is between the simulation step.when u say .tran 0.1 10us,the transient simulation step is 0.1us
ans when u say 0.01us the step is 0.01us.so the second case the results will be more accurate in that u will be able to examine in more detail the outputs of your simulation.

regards
amarnath