impulse

[meowth08]

What's the shape of an impulse?

Is it triangular, trapezoidal, rectangular, or none of these?

 

[ericgibbs]

What's the shape of an impulse?

Is it triangular, trapezoidal, rectangular, or none of these?

hi m80,

Look at this link.
**broken link removed**

 

[meowth08]

"The waveform shape, peak voltage, impedance, and application of the pulse varies among standards."

 

[MrAl]

What's the shape of an impulse?

Is it triangular, trapezoidal, rectangular, or none of these?

Hi,

Do you mean a theoretical impulse, such as what might be used to find the impulse response of a network?

 

[crutschow]

An impulse can be represented by the Dirac delta function. I believe that has a gaussian shape.

 

[meowth08]

To MrAl

Yes sir, the one used for stability testing.

 

[MrAl]

Hi again,

The kind of impulse i was talking about was the theoretical one used in electrical circuit theory, which is an infinitely high pulse with zero pulse width.
This is impossible to generate in reality, but an approximation can be used. The theoretical impulse has unit area so we can approximate it with a pulse
that is say 1Gv high and 1ns wide. That keeps the area equal to 1, but then so does 100Megavolts high and 10ns wide, 10megavolts high and 100ns wide,
etc., etc. So a 1000v pulse would have to be 1ms wide to approximate the impulse, and the output result delayed by 1ms.
Not sure if that helps you, but that's a theoretical view and some ways to think about it in reality.
The distribution idea is harder to picture, but basically the pulse ends up with more distribution near zero and almost none anywhere else.

 

[meowth08]

Hello,

I'm ok now with knowing that the area it covers is equal to one. This is consistent with crutschow's dirac delta function. However, I'm still wondering what is the lowest approximation we can give to the pulse height.

 

[crutschow]

Hello,

I'm ok now with knowing that the area it covers is equal to one. This is consistent with crutschow's dirac delta function. However, I'm still wondering what is the lowest approximation we can give to the pulse height.

A lower height means a longer width to maintain the unit area. So that will depend upon the frequency response of the circuit you are applying the impulse to. The low frequency components of the pulse, as determined by its width, should be much less than the low frequency response of the circuit (so as far as the circuit response is concerned, it still likes like an impulse). This can be experimentally determined by increasing the pulse width until the output response of the circuit changes shape. You should maintain the width below that point.

 

[meowth08]

Thank you very much. Another bunch of knowledge. I'm laughing at myself right now. Sometimes, I only need common sense.

 

[crutschow]

Thank you very much. Another bunch of knowledge. I'm laughing at myself right now. Sometimes, I only need common sense.

That's a good comment, since I think much of understanding technical stuff involves common sense (or logic to perhaps be more specific). Sometimes we ignore common sense since we think the subject is too complex to be understood by just analysis and logic. But I find it much better to intuitively (for lack of a better word) understand something then to try to use the equations that describe the subject, which generally leave me with no feeling for what's happening. Of course the equations are needed when we need to analytically calculate what's happening under specific conditions.

 

[MrAl]

Hi,


I couldnt agree more, but there's the flip side to this kind of thought process too. That's where we use equations to gain more common sense. Common sense is like intuition, and equations can often lead to more intuition, or put another way, a deeper understanding of the system. Before we start we may or may not know some general things about the system, but once we start running some equations and maybe start investigating the limits associated with those equations we start to get a better 'feel' for what is really going on behind the scenes.

 

[meowth08]

From my post about the shape of an impulse, sir eric shared a link and I got this. "The waveform shape, peak voltage, impedance, and application of the pulse varies among standards." However, after knowing that the area it covers is a unit area, another question came out of my mind. That is the lowest approximation we can give to the impulse. Failing to connect the information that I have which is the quoted one and the area, I still raised that question. From the reply of crutschow I realized that the answer to my question was already within the information I have. I regard it as a lack of common sense.

my opinion:

The use of equations is to dig deeper into the details. Supposing we have a control system with transfer function T(s)=P(s)/Q(s) and we are trying to get its impulse response to determine if the system is stable or not. From Q(s), we can determine the location of the poles which determines whether the system is stable, marginally stable, or unstable. But this I think is not enough. We have to test and try different values of Q(s) so that we can compare. What I am saying is, from two stable systems, there is a relatively more stable one which can be used in decision making. We can use equations to compute for time domain equivalent of t(s) so we can plot them and see what system decays to zero faster. That is the advantage of using equations.

On my part, I know I failed at first but realizing where went wrong in the least amount of time is for me already an achievement.

 

[MrAl]

Hi,

In theory you can apply a pulse of 0.1 volt, but that's only in a perfectly linear system. The input/output relationship is such that if you have to apply a 1000v pulse you can decrease that to 100v and multiply the test result by 10. That's another trick, but as i mentioned that's for a perfectly linear system which you may or may not have.

Yes there is such a thing as marginally stable, and that would make another system look better if it was more stable than that. One of the things you can look for is the decay time. In a tunable system we can tune the gain from low to high to make the system stable, marginally stable, and stable. As the system starts to become more unstable the decay time gets longer and longer, until it actually doesnt decay anymore but actually starts to grow (totally unstable). So if you had two systems that were similar and one responded with a quick decay and the other with a slow decay it would make some sense to deem the faster system more stable.
To be sure through you do have to vary the system through its expected range of gain or whatever else might change in the real life application.

There's also the amplitude itself to think about. You dont want a huge response either