Measuring circuit delay and impulse response

Hi,

Is there a way to measure a circuit's delay and the impulse response?

For instance is there a way where I can take S-parameter measurements and
inverse it using IFFT at a certain frequency range?  Will it be an accurate
measurement?

On 4/4/2011 10:17 AM, pacman101 wrote:
> Hi,
>
> Is there a way to measure a circuit's delay and the impulse response?
>
> For instance is there a way where I can take S-parameter measurements and
> inverse it using IFFT at a certain frequency range?  Will it be an accurate
> measurement?

1. Yes. 2. No.  2. Estimate?  Maybe.

Take S-parameter measurements? ..... sure

Inverse FFT "it" ... what's "it"?

Will "it" be an accurate measurement?  You're making the measurement, so 
will it?  What's "accurate"?

Q: Would an IFFT of [unknown] be an accurate *estimate* of [unknown]?
A: Unknown.

Sorry to be so terse but the answers are in the spirit of the questions. 
  If you want more serious answers then ......

S parameters are generally based on single-frequency steady-state 
measurements of circuits - so in that sense related to single-frequency 
Fourier Transforms of signals which result from steady-state analysis 
(i.e. an infinite convolution with a sinusoid). How might one relate 
that to a transient measure? Any suggestion of a connection would be 
conjecture.

What is the objective here?

Fred

On Mon, 04 Apr 2011 12:17:06 -0500, pacman101 wrote:

> Hi,
> 
> Is there a way to measure a circuit's delay and the impulse response?
> 
> For instance is there a way where I can take S-parameter measurements
> and inverse it using IFFT at a certain frequency range?  Will it be an
> accurate measurement?

Do you really mean "S-parameter", as in getting RF reflection 
coefficients?  You could, with a bit of math (and careful attention to 
your P's and Q's) translate this into a transfer function in the Fourier 
domain, and from there into an estimate of the impulse response.

If you mean "frequency response measurements", i.e. gain and phase 
responses, then your path is more direct but it'll still be something of 
an estimate.

The delay at any specific frequency can be readily estimated from the 
slope of the phase response.

-- 
http://www.wescottdesign.com

On 4/4/2011 10:17 AM, pacman101 wrote:
> Hi,
>
> Is there a way to measure a circuit's delay and the impulse response?
>
> For instance is there a way where I can take S-parameter measurements and
> inverse it using IFFT at a certain frequency range?  Will it be an accurate
> measurement?

You want time domain data, why try to get it in the frequency domain? 
Ideally you want the TDR's brother, the TDT.

-- 
Rob Gaddi, Highland Technology
Email address is currently out of order