Low-Pass Filtering a system, Time constant issue

[dolle39]

Hi!

I want to low-pass filter some simulations that I made. The indata to the simulation is built up by several modules and hence at the edges transients will occur. I want to filter these out. The indata in sampled at 20 Hz. How should I set the time constant for the filter? I mean I know that I should relate that to the time constant of the system in order to not filter away important data. But what is the time constant of a system really? Consider a car steering system. Is the time constant the time it takes for the system to manouver from max left to max right? Or is it the time it takes to manouver from +10% to -10%? Some guideluines would be appreciated.

 

[monchosoad]

I don't understand why do you need a LPF to do that (i mean, you "can't" do that). If you have a digital signal (the indata sampled at 20hz) just take the important data and throw away the transients. Or you can use a cyclic prefix technique or something like that.
Now, if you need to design a LPF to do another thing, the time constant is well defined in first-order systems (it is defined in superior orders too). The definition depends of the context.

 

[skeptic2]

What is the maximum rate of change of your data?