Frequency in a force-travel curve

[Marc77]

Hello all,

I'm having difficulties understanding something... Let's say I have calculated a force vs time curve looking more or less like a saw-tooth curve, for instance a lever with a 1-DOF revolute joint (the pivot) and below the pivot the lever has a spring loaded plunger moving over a couple bumps. (hope you can visualize this...) Now, if I want to re-create the force-travel behaviour in a hand controlled system(a physically controlled lever fed with the data from my calculated force vs time), I need to be able to figure out a way of finding the necessary minimum response time for that system. Since it's hand controlled it need to be fast since the operator might decide to jerk the lever(I know that last sentence sounds stupid...).

According to my boss it's possible to do a FFT of my calculated curve in order to find the minimum response time my real control system has to have, since the peaks of the saw tooth curve (my calculated curve) have frequency components that will be revealed after the FFT is performed. I however, don't really get this... The force vs time/saw tooth curve will have some sort of periodicity but I don't understand how you can find anything from the peaks of the saw tooth when looking at it in the frequency domain...

The force vs time curve is calculated via a Multi-Body system and the goal is to build a real system which lets you feel the result from the Multi Body system, for instance by having a servo controlling a 1-DOF revolute joint of a real lever...

Anyone have any ideas?

Best wishes

Marcus, Sweden

 

[Denial]

This is not my area, and so I do not really understand what your end objective is. If all you want to do is get a feeling for the significance of the various frequency components, I would work in the time domain rather than the frequency domain.

Take a single "pulse" of your infinitely-repeated curve and decompose it into its trigonometric components using Fourier Analysis. Then plot out the result for the first three terms, the first four terms, etc. At each stage you will be able to visualise how close you are to adequately reproducing your original pulse. When you have achieved an adequate reproduction, the frequency of the highest component used will lead you to the answer that I think you are looking for.

FWIW, the Fourier series for a saw-tooth whose pulse shape is an inclined straight line from (-pi,-pi/2) to (pi,pi/2) followed by a vertical straight line from (pi,pi/2) to (pi,-pi/2) is
(1/1)*sin(1x) - (1/2)*sin(2x) + (1/3)*sin(3x) - ...

 

[CESSNA1]

MARC77: I am not sure what you wnat to do either. I suspect your boss doesn't know also. An FFT or FT decomposes the time history in the sinusoids (frequencies) that make up the time history. It sounds like you want to manually replicate and automatic, mechanical function. If so, you have the automatic information, now you just have to train an operator to replicate it.

Regards
Dave

 

[Marc77]

Denial,

I think you hit the nail with your post, and that was my first thought as well when I talked to the man in charge. I've provided my boss with some raw data from a simulation that he will work on and we'll see what he comes up with. The reason for me not doing the job myself is that I haven't got the right software to do the job. But he has, at home...

I know it's difficult to understand what I'm talking about but I'll try to explain it... We want to be able to feel the result from a mechanical simulation without having to build costly prototypes that eats time and money...

If we have a real system that controls a lever(and yes a lever is what wer're building and selling at my company), and the behaviour of that lever is based on the results from a mechanical simulation model which, in turn, is based on the CAD geometry from our engineering design department, then you might be able to visualise what I'm aiming at...

It's basically a force-feedback system that is based not on your antics in a computer game but on the output from a simulation software... If we can build a real system that mimics the behaviour of the simulation (and those results are very accurate compared to real-life), say bye-bye to prototypes and say hello to glory and cost savings...

Best wishes

Marcus