How to make a delta function signal in a circuit?

[dilloncyh]

Basically I want to test my analog circuit using a forcing function that has a form of a delta function. The function generator I use outputs sine wave, triangular wave and square wave (+ve and -ve output in one period). Are there any ways to produce a square wave that has an output for like 5% of the time, with the other 45% of the time with output zero for half a period, so it acts like a delta function? I only have the most basic circuit components like L,C,R, diode, op-amp, logic gates, etc.

thanks

 

[zoki85]

Are you positive your function generator doesn't have a burst mode option?
http://cp.literature.agilent.com/litweb/pdf/5988-7507EN.pdf

 

[dilloncyh]

Are you positive your function generator doesn't have a burst mode option?
http://cp.literature.agilent.com/litweb/pdf/5988-7507EN.pdf

I guess the generators in my lab don;t have the function you mentioned. I asked the demonstrator during the lab, and he said all I could do was to use another model of function generator that produced +ve and 0 output alternately, but this is still not what I want as I need a narrow 'spike'.

 

[dilloncyh]

I have an idea: have two square wave output signals from the same function generator, and introduce a phase shift to one signal. I should get zero output when the two signals have opposite signs, and the amplitude should double when the two signals have the same signs. Does it work? But I need to think of a way to introduce a phase shift to one of the signal without affecting its amplitude

 

[tech99]

I think you could make a monostable having the required pulse duration and trigger it with the leading edge of the square wave from the generator.

 

[jim hardy]

a square wave that has an output for like 5% of the time, with the other 45% of the time with output zero for half a period

If i understand
you have described a pulse that is high for 2.5% of a whole period and low the rest of the time.

See if your generator has a pulse setting and a knob to set pulse width.

I looked up "Delta Function" at http://mathworld.wolfram.com/DeltaFunction.html

The delta function can be viewed as the derivative of the Heaviside step function,

It's easy to "differentiate" a square wave. Place a small capacitor and resistor in series across your function generator, capacitor first, voltage across resistor will be a series of spikes resembling but not exactly equal to the derivative .

Size R and C for a decay time that you like.

It might be close enough for practical purposes.

The true derivative of a true step would have infinite amplitude , which is a math phenomenon that i don't think you can realize with real parts.
Try above with a 'scope and see what you think.

 

[zoki85]

Regarding differentiating a square wave signal, op-amp differentiator should be also considered since OP has op-amp

 

[milesyoung]

Basically I want to test my analog circuit using a forcing function that has a form of a delta function.

An ideal impulse is, in theory, a great test signal, due to its flat amplitude spectrum over all frequencies, but it's not a physically realizable signal, since it has infinite power.

What is it you actually want to accomplish by applying this signal to your system?

In any case, for whatever your application is, you don't actually need an ideal impulse. Your system will have some finite bandwidth, so you just need to approximate a flat amplitude spectrum in the band it's responsive to. Chirp and PRBS signals are common for analog and digital systems, respectively.

Usually, though, depending on your application, you can get good results with a step signal. You could, for instance, use the high-frequency content of the edges of a low-frequency pulse train.