How Can I Build a High Frequency Pulse Generator with Specific Characteristics?

[Soliton_steve]

Hi,

I'm trying to build a high frequency generator. The pulse needs to have atleast 1ns rise/fall time, variable amplitude (~0-8V), 50 Ohm output impedance, and have very little jitter/be stable and I need to be able to set the pulse width and delay from ~5-1000ns.

Cheers,

Steve

 

[vk6kro]

Yes, I'd like one of those too.

In principle, you turn on a gate to give 5 volts out. Then you count pulses from a clock for the required pulse time and then you turn off the gate. Then you count a different number of pulses from the clock to get the delay time before you turn the gate on again.

However, you would have problems getting suitable chips to do it.
Taking your extreme values, 5 nS on 5 nS off, that is 100 MHz which means your clock speed would be at least 200 MHz to get two rising edges in a 5 nS pulse. There are not a lot of chips that can do programmed counting at a 200 MHz clock rate.

Then you need to control the counter to get the right number of counts and you need to display this somehow.

Such an instrument could be developed given $50000 or bought for about $5000 if one already exists. See Agilent (HP). They have some nice function generators.

If your requirements were not really that stringent, maybe you would like something like this:
http://alternatezone.com/electronics/hsfg.htm

 

[mheslep]

Such an instrument could be developed given $50000 or bought for about $5000 if one already exists. See Agilent (HP). They have some nice function generators.

Hmm that's a little high for a 200MHz pulse generator.
http://www.google.com/products?q=pulse+generator+250+Mhz&hl=en
This pops up quickly
http://www.bellnw.com/products/1210/

 

[vk6kro]

Yes, second hand would be a good way to go. Note that the $34 and $53 ones were just books, not the instrument.

But it looks like you could get something useful for about $200 without even having to bid for it.
Like that
TEKTRONIX PG 502

You might also like to read this thread from June 22nd:
https://www.physicsforums.com/showthread.php?t=321383
Several nice instruments were located but were expensive if bought new.

 

[waht]

Hmm that's a little high for a 200MHz pulse generator.
http://www.google.com/products?q=pulse+generator+250+Mhz&hl=en
This pops up quickly
http://www.bellnw.com/products/1210/

some of these pulse generator in the link going for $600 up are at least 30 years old,

you could probably find a good HP 30 year old for half that on ebay,


Here is modern pulse generator:

http://www.home.agilent.com/agilent/product.jspx?nid=-536902258.536881947.00&cc=US&lc=eng

 

[vk6kro]

That's more like the prices we were getting before. $10000 to $16000.

what, are you in Australia? Where?

 

[waht]

That's more like the prices we were getting before. $10000 to $16000.

what, are you in Australia? Where?

I'm in the US.

The cost of good test equipment in general could be a down payment on a house.

 

[waht]

Here is a freely available service manual of HP 8012B pulse generator with circuit description and schematics.

It's 46 mb in pdf, and is an informative reference

http://bama.edebris.com/manuals/hp/8012b/

 

[mheslep]

some of these pulse generator in the link going for $600 up are at least 30 years old,

you could probably find a good HP 30 year old for half that on ebay,


Here is modern pulse generator:

http://www.home.agilent.com/agilent/product.jspx?nid=-536902258.536881947.00&cc=US&lc=eng

Well that's a pattern generator, different beast. Also, this is the usual 'system' based gear - a full blown computer system w/ 488 remote control, 'SRAM adapter cards', blah, blah, and it happens to have a pulse generator along for the ride.

 

[vk6kro]

If you were prepared to experiment a bit, there is a more crude way of getting short pulses.
I invented this years ago and find it very useful.

If you take a NAND gate and apply a square wave to one input directly and to the other input through an integrator, you get a downward going pulse at the output that depends, in width, on the time constant of the integrator. It starts at the rising edge of the input square wave.

The integrator can be just another NAND gate with the input pins tied together, a series resistor to the input of the NAND gate and a capacitor from output to input. Just like an OP amp.
You can make the resistor variable for variable pulse width and switch capacitors for bigger steps in pulse width.

If you take the output of this pulse generator and pass it through an inverter (which can be another NAND gate with the inputs tied together), you have a positive going pulse that repeats at the same rate as the square wave, but variable up to the period of the positive part of the square wave.

You do not get a pulse off the falling edge of the square wave with this arrangement.

 

[Bob S]

I suspect you will want to use discretes rather than logic arrays for your outputs. Do you know what style of output you want; NPN vs PNP, collector outputs (best for current sources) or emitter followers? Any particular circuit, like (common emitter) differential pair for example?