Half-Sin pulse shapping

[NorbertoBarroca]

I'm trying to implement an Half Sin Pulse shapping filter like the one shown bellow.

My input values are given by:

My filter function is:

However when I make the convolution between the input values and the filter the result is no the expected I alway have zero can someone help me?

Message was edited by: Norberto Barroca

[AWesterman-disa]

Hi Norberto

Would recommend change to definition of pulse for clarity, at 2TC 4TC etc output is defined by 2nd condition but mathematically it could be either 1 or -1.

The main problem however seems to be the way mathcad divides the period of the integral.

It is reaching a limit at which it cannot resolve the information & everything collapses to the 0 result that you get.

Not too useful in your case where you actually want to integrate between infinities.

Example sheet gives same output for I =100, 455 & 460 *Tc (3.1830988618x10^-7s) but 0s for the 461* Tc case.

It might be that for a different system (eg not mathcad 12 ...) it will run to different limits, I'll be interested to hear comments about this.

Note that I've tried changing TOL & CTOL values - does improve the accuracy of the result given but not the level at which it 'breaks' & gives a zero answer.

Regards

Andy