¿ªÔÆÌåÓý

Help! How do I do find maximum signal easily!


 

I'm trying to find an easy way to see if my signal saturates any OpAmp output, or violates some level to the next stage.

Right now, I'm using input values predetermined from matching to the existing breadboard circuit I know the levels are ok, but now I want to redesign using simulation.

Problem is...I have a VERY complex time input signal yet has fairly 'flat' harmonic content, so historically I have been able to use AC analysis and get very accurate simulation. for example, a representative data point using most recent LTspice I predict 35 ppm noise at 1kHz [that is, noise to signal levels] and MEASURE 36 ppm. So am very happy with the simulation and models. [the noise pretty much matches within 10% over the rest of the spectrum]

Here is the problem. I want to redesign the circuit, but not build it to verify operating levels. So how do I bounce between this frequency domain and the time domain WITHOUT taking days for each run?

What I mean is, I can represent the input waveform in a .wav form or PWL model, but we're talking 4400 UNIQUE data points that repeat and repeat. then, do .tran analysis, but that takes iteration after iteration to get a stable repeatable output signal.

Or, I could 'post-process' an AC analysis by just 'assuming' everything is copacetic, run the AC analysis, and find the harmonic values/phase angles and use octave to plot the signal. And, octave will tell me if I've violated maximum signal level.

How do I make the AC simulation put out specific frequency/phase values, maybe not the ones it used during the AC analysis? And how accurate will they be? The signal may be complex enough that a few degrees could cause a catastrophe, but I guess I could cross that bridge later, getting close right now would be a big help.

Is there a way to make a file of table of results? I give frequency points [not in the AC anlayis] and LTspice gives me amplitude/phase in a file, so I can run that through octave. Is there a way to 'force' the frequency data points of LTspice - there's a lot fo them?

Or, is there a way to stay in the time domain easily quickly generate the waveform as I change a component value someplace?

Help!

Join [email protected] to automatically receive all group messages.