The Designer's Guide Community Forum

The Designer's Guide Community Forum
https://designers-guide.org/forum/YaBB.pl
Simulators >> Circuit Simulators >> group delay calculation in Cadance
https://designers-guide.org/forum/YaBB.pl?num=1320769482

Message started by raja.cedt on Nov 8^th, 2011, 8:24am

Title: group delay calculation in Cadance
Post by raja.cedt on Nov 8^th, 2011, 8:24am

hello all,
Generally i find group delay through calculatorin spectre , but i have tried with finding the derivative of the Phase, both are not matching. can any one please tell me why this is happening to me.

Thanks,
raj.

Title: Re: group delay calculation in Cadance
Post by buddypoor on Nov 8^th, 2011, 8:49am

Raj,

there are two possible error sources:
* The x-axis (frequency) must be w not f
* This axis must scaled linear

There is a 3rd error source: The phase must be scaled in rad (not degrees).

Title: Re: group delay calculation in Cadance
Post by raja.cedt on Nov 8^th, 2011, 8:53am

hello buddypoor,
I have taken care of W by diving by 2*pi, but why axis must be linear?

Thanks,
raj.

Title: Re: group delay calculation in Cadance
Post by buddypoor on Nov 8^th, 2011, 9:20am

Dividing by 2*Pi ???
For my opinion: w=2*Pi*f

It must be linear because the program calculates the differential quotient of the function to be analyzed.

Title: Re: group delay calculation in Cadance
Post by raja.cedt on Nov 8^th, 2011, 11:33am

hello buddypoor,
it's 2*pi becaz when you find derivative you will get dΦ/df, si want to find dΦ/dω, so divided with 1/6.28.

Thanks,
Raj.

Title: Re: group delay calculation in Cadance
Post by raja.cedt on Nov 8^th, 2011, 12:47pm

hello buddypoor,
One intersting thing, i have calculated the ration of group delay given by simulator and my method(mean derivative of the phase), interstingly it is 360, i am expecting 2*3.14 ...

Thanks,
raj.

Title: Re: group delay calculation in Cadance
Post by buddypoor on Nov 8^th, 2011, 3:02pm

raja.cedt wrote on Nov 8^th, 2011, 12:47pm:

Is this a question?
Sorry, but I do not understand the content of your posting.
The result must be in seconds (or mikroseconds).
And - what do you mean with "mean derivative"?
The derivative of the phase function again is a function of frequency.
Please, clarify.

Title: Re: group delay calculation in Cadance
Post by rfidea on Nov 12^th, 2011, 6:49am

The relationship is simple. Eg. Virtuoso from CDS is using degrees for phase and Hertz for frequency. This gives

Tg = -d(fi,rad)/dw = - d(fi,deg*2*PI/360)/d(2*PI*f) = -d(fi,deg)/df / 360

So, you should divide the derivative from Virtuoso with 360 to get the group delay.

Title: Re: group delay calculation in Cadance
Post by raja.cedt on Nov 12^th, 2011, 9:42am

hello rfidea,
thanks a lot, i forgot this!!!!

Thanks,
Raj.

Title: Re: group delay calculation in Cadance
Post by rfidea on Nov 12^th, 2011, 11:25am

Yeah, sometime the answer is to simple...

Title: Re: group delay calculation in Cadance
Post by Mikay on Jan 22^nd, 2014, 8:20am

I noticed this is a quite old topic. But this will be unaccurate if the phase response is not linear. The group delay can be calculated in the spectre: (fi,deg)/f/360.

Mikay

rfidea wrote on Nov 12^th, 2011, 6:49am: