PWM and it`s duration (picosecond pulse widths)

[Sveral]

Hello,
I wanted to ask, if anyone knows of a realistic circuit, which could be capable of producing a PWM of 1 pico second on and 350 ps off. Thanks to everyone taking the time to answer in advance.

 

[cnh1995]

Try a 555 astable multivibrator with a few pF capacitor and very small resistances. But I'm not sure if that's realistic since other parasitic effects might be significant at this small time scale.

 

[anorlunda]

What kind of load are you driving with this circuit? That too will be very significant.

 

[phinds]

I haven't looked at the specs in a lot of years but I don't think a 555 can switch that fast.

 

[Averagesupernova]

A 555 timer will never achieve the frequency in question. Unless I misunderstand the OP, we are talking about gigahertz frequencies. PWM at this frequency is a tall order. I have to wonder what is behind this requirement.

 

[donpacino]

Hello,
I wanted to ask, if anyone knows of a realistic circuit, which could be capable of producing a PWM of 1 pico second on and 350 ps off. Thanks to everyone taking the time to answer in advance.

What are your signal and load requirements?
Also I usually try to ask a question like this. Are you SURE you need 1 pico second on times?

 

[Svein]

350ps off is a start. Your basic timer should therefore run at 1/350ps ≈ 2.86GHz. On the other hand, at normal signal propagation rate in electronics (20cm/ns), 1 ps equals 0.2mm. A NOR gate with both inputs conected to the oscillator with 0.2mm signal length difference could thus shave off a 1ps pulse.

No, I do not know any logic family fast enough for that. Even the old MECL 3 would have problems with that speed. You need to design a fast ASIC...

 

[Borek]

AFAIK switching times of so called "fast transistors" are in the ten of ns range at best, orders of magnitude too slow. 555 is in the same category at best.

 

[phinds]

@Sveral you still have not told us what it is that you are actually trying to DO. Why do you need a 1ps wide pulse?

 

[Sveral]

What are your signal and load requirements?
Also I usually try to ask a question like this. Are you SURE you need 1 pico second on times?

Yes, I am positive. It is needed to test a research, that was performed in the US for not fully explained reasons. It is made to create a PWM circuit, which would drive a 50 mW diode, which would prove or disprove a theory.

 

[Sveral]

@Sveral you still have not told us what it is that you are actually trying to DO. Why do you need a 1ps wide pulse?

Do, please, take a look a bit higher.

 

[donpacino]

Ok, Well like others have said this is not an easy feat. You will either need to hire someone to build an asic for you, or pursue similar options yourself.

your typical diode can be driven by ms times.

What is the load on the circuit? What are the rise and fall time requirements. You mention PWM, does this need to go from 0-100% duty cycle. With what resolution?

 

[Sveral]

Ok, Well like others have said this is not an easy feat. You will either need to hire someone to build an asic for you, or pursue similar options yourself.

your typical diode can be driven by ms times.

What is the load on the circuit? What are the rise and fall time requirements. You mention PWM, does this need to go from 0-100% duty cycle. With what resolution?

100 mJ per
macropulse at a repetition rate of 30 Hz. Each
macropulse is composed of È104 micropulses
of 1 ps duration, spaced 350 ps apart. The
spectral width is 50 to
90 cm^-1. Honestly, there`s not much detail I can give, because I`m not really knowledgeable about this topic, except for the fact, that a free electron laser was used so the resolution should have been rather high.

 

[donpacino]

Honestly, there`s not much detail I can give, because I`m not really knowledgeable about this topic, except for the fact, that a free electron laser was used so the resolution should have been rather high.

Why are you doing this? Are you a student? is this for a company? is this for fun?

This is a rather large difficult task you are asking about.

 

[Sveral]

Why are you doing this? Are you a student? is this for a company? is this for fun?

This is a rather large difficult task you are asking about.

I know, that it`s on the ridiculously difficult scale of things, but...it`s for a research, I`m a student and it`s for fun and research.

 

[Sveral]

Why are you doing this? Are you a student? is this for a company? is this for fun?

This is a rather large difficult task you are asking about.

The only way I`ve read it definitely could be done is with a free electron laser, but that`s not really an option...

 

[Tom.G]

http://www.aml.engineering.columbia.edu/ntm/level3/ch02/html/l3c02s03.html
http://www.aml.engineering.columbia.edu/ntm/level3/ch05/html/l3c05s02.html

Above found through: https://www.google.com/search?source=&q=1ps+laser+pulse

 

[berkeman]

Yes, I am positive. It is needed to test a research, that was performed in the US for not fully explained reasons

Could you please provide links to the papers which describe the previous work? What is your budget for equipment for setting up this experiment?

 

[donpacino]

I know, that it`s on the ridiculously difficult scale of things, but...it`s for a research, I`m a student and it`s for fun and research.

Have you done anything similar before?

assuming this is for that laser, its a fairly significant undertaking. You said you're not really knowledgeable on the subject. Have you worked with lasers or circuitry before?

When you say for fun and research, is this your dip into the world of lasers? Did you just read a paper and decide hey this looks cool I'll do it.

One needs to walk before they can run. I hate to discourage you, but you may want to try something simpler. As you can tell from our responses, this is a very difficult task. It is not a task someone who does not have a background in the field should dive into.

note: don't take this as me saying not to do any project. I'm saying ps times are pushing the limits of modern circuity. ns times, while not trivial, would be much easier to design and use. us would be even more preferable.
I do not know enough about lasers to speak to entry level projects

 

[CWatters]

Google found..
https://photonics.ixblue.com/files/files/pdf/EPG-200.pdf

25ps - 200ps Electrical Pulse/Impulse Generator

The EPG-200 is a specially-designed short electrical pulse generator. The output pulse can be as short as 25 ps, and with the high-voltage option, it can drive most off-the-shelf laser diodes or LN modulators to generate short optical pulses. The tunability of the pulse-width and frequency repetition rate offers flexibility in laser machining application, where control of laser pulse energy is required.

That's as near as I can find to what you asked about. I think if you want shorter pulses you will have to get a bit of kit that actually produces the optical pulses as well.

 

[Borek]

I wonder how many frequencies (and what frequencies) one would need to mix to produce such pulses (1 ps/350 ps). Obviously for the pulse to be perfectly square wave one will need infinitely many, so the question is about practical limit of using such approach for pulse generation.

My gut feeling is that it could work for duty cycle close to 50%.

 

[CWatters]

I wonder how many frequencies (and what frequencies) one would need to mix to produce such pulses (1 ps/350 ps). Obviously for the pulse to be perfectly square wave one will need infinitely many, so the question is about practical limit of using such approach for pulse generation.

I'm a bit rusty but wouldn't they all have to be in the right phase relationship? If they are generated by separate oscillators that might be hard if post #7 is correct.