why is output of my MATLAB function in Simulink raises with slope?

Hello everyone,

I tried to implement my own pulse generator with variable inputs. When I simulate it and compare results with the build in Pulse Generator I see that my output becomes high with a slope, but Simulink's Pulse Generator becomes high immediately. How can I have this kind of output?

Simulation performed for 100 seconds, with Fixed-step, discrete solver and Fixed-step size has been taken "1".

function [y, countTime] = fcn(Amplitude, period, width, phaseDelay, timeIncremental, Clock, pre_y, preCountTime)
%Setting previous values.
y=pre_y;
countTime=preCountTime;
%y is high for 'width' time, so it can be low again
if countTime==width
    y=0;
end

%Counting the time between two high values
if y>0 || countTime>0
    countTime=countTime+timeIncremental;
end
%Output is high first time after 'phase delay' passes.
if Clock==phaseDelay
    y=Amplitude;
end
%Period has been reached, y is high again.
if countTime==period
    countTime=0;
    y=Amplitude;
end

[Aquatris]

That slope is probably due to sampling rate. I think you are using ode45, which is variable time step and it selects large time steps when you put your version of pulse and small time step when you put built-in one. If you examine your version in detail, you will see that the difference in time between the start of the slope and end of the slope is just 1 time step. I recommend you do the same test with fixed step solver and compare that way.

[Berkin Birol]

Thank you for your reply. The solver was fixed step and in the same solver Simulink's own pulse generator works just fine. I think I miss something due to my lack of experience in Simulink.

[Berkin Birol]

I solved the issue with zero order hold block and uploaded the block to: https://www.mathworks.com/matlabcentral/fileexchange/69483-customizable-pulse-generator.

But, I am still wondering if there is an another way(without ZOH) to do this.