Path: news.mathworks.com!not-for-mail From: "Phil Goddard" <phil@goddardconsulting.ca> Newsgroups: comp.soft-sys.matlab Subject: Result of Derivative 2 times then integrator 2 times? Date: Fri, 19 Jul 2013 21:22:16 +0000 (UTC) Organization: Goddard Consulting Lines: 19 Message-ID: <kscam8$jlp$1@newscl01ah.mathworks.com> References: <ksajbv$4f3$1@newscl01ah.mathworks.com> Reply-To: "Phil Goddard" <phil@goddardconsulting.ca> NNTP-Posting-Host: www-03-blr.mathworks.com Content-Type: text/plain; charset=UTF-8; format=flowed Content-Transfer-Encoding: 8bit X-Trace: newscl01ah.mathworks.com 1374268936 20153 172.30.248.48 (19 Jul 2013 21:22:16 GMT) X-Complaints-To: news@mathworks.com NNTP-Posting-Date: Fri, 19 Jul 2013 21:22:16 +0000 (UTC) X-Newsreader: MATLAB Central Newsreader 26433 Xref: news.mathworks.com comp.soft-sys.matlab:799594 > If we take the derivative of an input signal 2 times and then take the integrator 2 times, we are supposed to get the original input signal. That's only (possibly) true if the original signal is differentiable, which a step is not. > > So how can I solve the problem? I mean in the Simulink simulation, that I can get output = step signal ??? Not the large value. This isn't a Simulink issue, it's a general mathematics issue. Working backwards from your (desired) output, you need a signal which integrates to a step. That would be a durac delta function. Prior to that you are asking for a signal that integrates to get the durac delta function. I'm not sure (mathematically) what that looks like. So in effect you are asking for Simulink to do something that even writing down on paper isn't really well formulated. You may have better success choosing an input signal that is better defined over the range of interest, such as a sine wave, but even then numerical issues (of taking the derivative) may arise. Phil.