Pulse width of irregular wave

[Holzapfel]

Hello everyone,

I'm trying to inject pulse with different pulse width and height in to my adder circuit though carry_in, and record the height/width of pulse occurs at carry_out, and then build regression model... This is background.
I'm facing some troubles, in simulation I get a lot of responses like following figure, I'm wondering whether it is possible to do some approximations to calculate the pulse width for this wave?

Thanks a lot for help!
Regards,
Holz

 

[c_mitra]

Usually we understand a rectangular (really trapezoidal) shaped waveform as a pulse. If my understanding is correct, a pulse is having both a rising and a falling edge. A capacitor discharge (an exponential decay waveform) is not usually understood as a pulse. The rising edge is expected as a turn-on and the falling edge is expected as a turn-off signal to some device. Thus pulses are supposed to work in the digital environment but the "pulse" of your signal stumps me!

 

[Holzapfel]

Usually we understand a rectangular (really trapezoidal) shaped waveform as a pulse. If my understanding is correct, a pulse is having both a rising and a falling edge. A capacitor discharge (an exponential decay waveform) is not usually understood as a pulse. The rising edge is expected as a turn-on and the falling edge is expected as a turn-off signal to some device. Thus pulses are supposed to work in the digital environment but the "pulse" of your signal stumps me!

Hi, thanks for reply. Yes I know this "pulse" is really irregular... and I'm now looking for how can I approximate this irregular wave into a square wave.

 

[KlausST]

Hi,

it could help if you tell
* where the input signal comes from (to better understand it´s waveform)
* what the output is used for ( to decide a "rule" for the beginning and the end)

Klaus

 

[Holzapfel]

Hi,

it could help if you tell
* where the input signal comes from (to better understand it´s waveform)
* what the output is used for ( to decide a "rule" for the beginning and the end)

Klaus

Hello,
the input wave is trapezoidal wave, rise time=30ps, fall time=30ps,plateau is 20ps,with amplitude=500m, it is the input signal of a 1bit adder's carry_in, so the pulse width of trapezoidal wave is 50ps.
The output as shown in the figure, is the wave occurs at carry_out. I need to record the amplitude/pulse width of carry_out, then with a sweep of plateau, I can get the model of how plateau affects amplitude/pulse width at carry_out, use this model when I extend the adder from 1bit to nbits to predict the amplitude/pulse width at carry_out_Nbits.
For begin and end, this is no special rule, since I just run transient simulation in virtuoso.
Thanks for help!

Holz