Designer’s Guide Community

manodipan Community Member Offline Posts: 83	THD of Switched capacitor filter Feb 17^th, 2009, 3:36am Hi friends, I am simulating a switched cap filter operating at 10 MHz,now for THD calculation harmonics are to be taken upto what frequency??in continuous domain we consider upto infinity.but in this case output is sampled at 10MHz....thanks a lot...
Back to top	IP Logged

manodipan

Community Member

Offline

Posts: 83

Re: THD of Switched capacitor filter
Reply #1 - Feb 17^th, 2009, 4:24am

Hi ,
I have some more doubts......does amount of distortion has some relation to order of filter??
also for some spec of SC filter does suppose 2nd order distortion improve as we change the fundamental freq towards dc...what i have seen is that for suppose 300KHz cutoff ,better 2nd order distortion at 10KHz input than 100KHz for same amplitude of input signal....

Back to top

IP Logged

subgold

Community Member

Offline

Posts: 97

Re: THD of Switched capacitor filter
Reply #2 - Feb 18^th, 2009, 4:12am

manodipan wrote on Feb 17^th, 2009, 4:24am:

Hi ,
I have some more doubts......does amount of distortion has some relation to order of filter??

it will have some influence.

manodipan wrote on Feb 17^th, 2009, 4:24am:

also for some spec of SC filter does suppose 2nd order distortion improve as we change the fundamental freq towards dc...what i have seen is that for suppose 300KHz cutoff ,better 2nd order distortion at 10KHz input than 100KHz �for same amplitude of input signal....

it is not good to choose the sampling freq as a harmonic of your input. try to use another input freq.

Back to top

IP Logged

Visjnoe

Senior Member

Offline

Posts: 233

Re: THD of Switched capacitor filter
Reply #3 - Feb 18^th, 2009, 11:29am

Dear Manopidan,

1. Yes, in theory you should consider all harmonics 'upto infinity' to calculate THD,
but in practice you can end the summation after the first n (n=3,n=5,n=10?) since
the higher harmonics will have less and less power and eventually slide under the
noise floor.

2. Sampling makes no difference. If your harmonic > Fs, than it will simply fold back
into your baseband ([-Fs/2,Fs/2]).

3. Use an RF simulator if you can, you will gain a lot of simulation time versus
the transient analysis approach.

Regards

Peter

Back to top

IP Logged