PSS Analysis of a multiphase system

[Sam]

Dear All,

I am trying to perform a PSS analysis of a multiphase DC-DC converter. However, the analysis is failing due to convergence problem.

But the PSS (and followed by PAC) on single-phase topology works just fine and gives correct result.

I am suspecting due to multi-phasing the PSS is failing. There are 5 phases in the system with fundamental period of 2 usec. So, each channel is phase-shifted by 400 nsec. And the output ripple period is 400 nsec. I have specified the beat frequency as 50 MHz (1/20nsec.) is PSS analysis setup.

If the problem is happening due to multi-phases, then what is the solution?

Regards
Sam.

 

[Andrew Beckett]

On 06/04/13 05:30, Sam wrote:

Dear All,

I am trying to perform a PSS analysis of a multiphase DC-DC converter. However, the analysis is failing due to convergence problem.

But the PSS (and followed by PAC) on single-phase topology works just fine and gives correct result.

I am suspecting due to multi-phasing the PSS is failing. There are 5
phases in the system with fundamental period of 2 usec. So, each channel

is phase-shifted by 400 nsec. And the output ripple period is 400 nsec.
I have specified the beat frequency as 50 MHz (1/20nsec.) is PSS
analysis setup.

If the problem is happening due to multi-phases, then what is the solution?

Regards
Sam.

Sam,

If you have a fundamental period of 2us, shouldn't the fundamental
frequency ("beat frequency") be 500kHz?

You need a period of the common frequency to be able to contain an
integer number of cycles of all the periods.

Regards,

Andrew.

 

[Sam]

On Tuesday, June 4, 2013 3:06:00 PM UTC+5:30, Andrew Beckett wrote:

On 06/04/13 05:30, Sam wrote:

Dear All,



I am trying to perform a PSS analysis of a multiphase DC-DC converter. However, the analysis is failing due to convergence problem.



But the PSS (and followed by PAC) on single-phase topology works just fine and gives correct result.



I am suspecting due to multi-phasing the PSS is failing. There are 5

phases in the system with fundamental period of 2 usec. So, each channel

is phase-shifted by 400 nsec. And the output ripple period is 400 nsec.

I have specified the beat frequency as 50 MHz (1/20nsec.) is PSS

analysis setup.



If the problem is happening due to multi-phases, then what is the solution?



Regards

Sam.





Sam,



If you have a fundamental period of 2us, shouldn't the fundamental

frequency ("beat frequency") be 500kHz?



You need a period of the common frequency to be able to contain an

integer number of cycles of all the periods.



Regards,



Andrew.

Hi Andrew,

Sorry, actually while posting, being unmindful, I have written 50 MHz. In the setup, I really put 500 KHz as the beat frequency !

So, please suggest any solution!

 

[Sam]

On Tuesday, June 4, 2013 3:06:00 PM UTC+5:30, Andrew Beckett wrote:

On 06/04/13 05:30, Sam wrote:

Dear All,



I am trying to perform a PSS analysis of a multiphase DC-DC converter. However, the analysis is failing due to convergence problem.



But the PSS (and followed by PAC) on single-phase topology works just fine and gives correct result.



I am suspecting due to multi-phasing the PSS is failing. There are 5

phases in the system with fundamental period of 2 usec. So, each channel

is phase-shifted by 400 nsec. And the output ripple period is 400 nsec.

I have specified the beat frequency as 50 MHz (1/20nsec.) is PSS

analysis setup.



If the problem is happening due to multi-phases, then what is the solution?



Regards

Sam.





Sam,



If you have a fundamental period of 2us, shouldn't the fundamental

frequency ("beat frequency") be 500kHz?



You need a period of the common frequency to be able to contain an

integer number of cycles of all the periods.



Regards,



Andrew.

Hi Andrew,

After setting the parameter "steadyratio" (in accuracy setting) to 0.065 (default is 0.01 in case of Conservative "errpreset"), the PSS analysis converged. Is this setting fine to to accept the result?

Sam.

 

[Andrew Beckett]

On 06/07/13 10:23, Sam wrote:

Hi Andrew,

After setting the parameter "steadyratio" (in accuracy setting) to
0.065 (default is 0.01 in case of Conservative "errpreset"), the PSS

analysis converged. Is this setting fine to to accept the result?

Sam.

Yes, that should be fine.

Andrew.