Continue to Site

Welcome to EDAboard.com

Welcome to our site! EDAboard.com is an international Electronics Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.

finding fft in cadence

Status
Not open for further replies.

arya.jagadeesh

Member level 2
Member level 2
Joined
Jul 16, 2013
Messages
44
Helped
4
Reputation
8
Reaction score
4
Trophy points
8
Location
INDIA
Activity points
285
i am trying to find the dynamic parameters of adc so i want to do the dft(fft)

my sampling frequency is 55.9MHz and inp is 16.6MHz and using 1024 samples and
according to formula f_in/f_s = m_C/M
m_c is the nearest prime and adjust f_in according to that and while trying to do this in cadence

dft window is asking from and to _________what is the value that i should give for above problem.........i have given the time period of my input signal and in the fft i got samples only at fundamental and harmonics and nothing else....so how this value is going to effect...

thank you
 

That is just your sampling window. So you may be simulating your ADC in a time interval but dft is going to calculate ft for the time interval specified by from-to parameters. Please refer to cadence documentation, most of the special functions are really well documented.

Also to give you a straight answer, let's say that the number of samples you got from coeherent sampling is M, then your time region should contain M points otherwise you'll see meaningless ft. So the time interval should be exactly M*f_s. Commonly used approach is giving a little time for the ADC or DAC to settle, to reduce the effects of startup problems, after then should your samples for dft begin. That's why there are parameters like from or to :D. It makes life easier.
 
is it M* 1/f_s

- - - Updated - - -

here i random ly took 1024 samples ....is there any constrain on it?
 

Yeah, it should have been 1/fs, must have been a typo sorry. No they shouldn't be random 1024 samples, they should be the consecutive 1024 samples, in a certain time interval. There are no constraints, why would there be? Only constraint is if you take more samples you can reduce the noise floor.
 

Status
Not open for further replies.

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Back
Top