Question about making RF pulse

[panoum]

I am working on a project which involves producing an RF pulse of about 5-10 nsec duration.The carrier frequency should be about 2-3 GHz.What kind of switch should I use?How should I drive it?Has anybody got any ideas or made anything similar that works fine?

 

[flatulent]

This calls for PIN diodes driven by ECL type logic. The Agilent web site has application notes on these diodes. Ordinary switches for microwaves are not fast enough.

 

[flyhigh]

Hi,

besides PIN diods, you can sometimes think on using step-recovery diodes as a base of short pulse genereator. They offer much shorter pulses (depending on a frequency of fundamental you can get even hundreds of ps) and are much more expensive then PINs, so this is just in the case you need something faster in the future.

flyhigh

 

[sergio mariotti]

Please explain if you are asking about a true "pulse" or a "pulse modulation" of a 2-3 Ghz CW signal.
The first case call for the use of a comb generator and involve fast SRD diodes or fast varactors. Agilent produce connectorized comb generators.
To drive it, many people use Hybrid Power Amplifiers, generally they may deliver more power than monolitic Amplifiers.
If you are speaking about pulse modulation, my opinios is that PIN modulator are often too slow.
I know that only Narda produce one PIN diode modulator able to be fast enough.
GaAs FET switch are fast enough but they may manage only "litte" amount of RF power.

 

[panoum]

I want to send some periods of a simple 2-3 GHz sinusodial signal.Could I use a counter?

 

[sergio mariotti]

?? a counter? are you thinking to the gating possibility of some counters?
I don't know every counter on the marlet, but i don't think is possible.
You need an electronic switch, and... if the "gate" must be coherent (sincronous) with the phase of CW, an electronic circuitry is necessary.

 

[radiohead]

@ flatulent:

You propose to use fast PIN diodes for switching. Can you give some typenumber which have a reverse recovery time that is small enough to generate 5 ns RF pulses? I would use the PIN diodes as a fast gate between the oscillator and the output, the design focus goes to isolation.

Thanks!

 

[flatulent]

PIN diodes are optimized for use as switches.

Another way to generate this fast pulse is to use transmission line pulsers.

http://www.hp.woodshot.com/hprfhelp/lit/diodelit.htm has application literature on PIN diodes and other types.

Figure 4 in www.hp.woodshot.com/hprfhelp/4_downld/lit/diodelit/an929.pdf shows bias required for switching speeds down to 2 ns.

 

[radiohead]

Thanks flatulent! Respect!