Fast buffers for non-linear loads

[f95toli]

I am currently using a pulse generator with a reasonably fast rise-time (something like 1 ns) for some experiments. It works well as long as I am using e.g. 50 Ohm loads but the problem is that it can't source much current, meaning it doesn't work well for lower impedances.
Now, my problem is that I am currently doing some experiments where I need to send fast pulses to a diode (well, the IF port of a mixer; but it essentially behaves like a diode); it sort of works as it is but I need more current in order to improve the performance; I probably need something like 50-100 mA.

The idea is to put some kind of fast buffer in between the pulse generator and the load. The question is what to use?
I would like to keep the rise time below 10 ns (the faster the better)

Would a high speed buffer like OPA693 be OK for something like this; or should I go for a MOSFET driver (which would only give a fixed voltage; but can source a lot of current)?

 

[ShadowPho1]

Here's a link I found while googling quickly:

http://www.slac.stanford.edu/cgi-wrap/getdoc/slac-pub-13269.pdf

Not sure if it would be more help.

 

[berkeman]

I am currently using a pulse generator with a reasonably fast rise-time (something like 1 ns) for some experiments. It works well as long as I am using e.g. 50 Ohm loads but the problem is that it can't source much current, meaning it doesn't work well for lower impedances.
Now, my problem is that I am currently doing some experiments where I need to send fast pulses to a diode (well, the IF port of a mixer; but it essentially behaves like a diode); it sort of works as it is but I need more current in order to improve the performance; I probably need something like 50-100 mA.

The idea is to put some kind of fast pulses in between the pulse generator and the load. The question is what to use?
I would like to keep the rise time below 10 ns (the faster the better)

Would a high speed buffer like OPA693 be OK for something like this; or should I go for a MOSFET driver (which would only give a fixed voltage; but can source a lot of current)?

Can you just use a 74AC245-style buffer? It won't quite do 100mA in 1ns, but it should do many 10's of mA in 5ns or so. Would the voltages be compatible?

 

[berkeman]

Can you just use a 74AC245-style buffer? It won't quite do 100mA in 1ns, but it should do many 10's of mA in 5ns or so. Would the voltages be compatible?

BTW, you can parallel up several 245 drivers to get more current...

 

[f95toli]

Can you just use a 74AC245-style buffer? It won't quite do 100mA in 1ns, but it should do many 10's of mA in 5ns or so. Would the voltages be compatible?

That might work. I can adjust the output voltage of the pulse generator so that is not really a problem. But are 74AC25 buffers really suitable for "strange" loads like this?

The reason I was looking at OPA693 is that it is not only fast but can also drive capacitive loads quite well, although I am a bit worried about ringing since it is basically a high-speed op-amp.

 

[vk6kro]

I wonder how you would go using MMICs

See this page:
http://www.minikits.com.au/doc/MMICSb.PDF
Not suggesting you get them from there, though.

These wideband amplifiers work very well and I have seen the MAR6 used as a buffer to build up power levels. You could parallel a couple of them up if you were careful.

The MAV11 gives an output of 17.5 dBm or about 56 mW at 1000 MHz (across 50 ohms?) so it is pretty fast.

They are generally quite cheap too.

Note that you have to provide a load for them and this should be a surface mount type resistor(s) to get low inductance.

Edit: you might like to have a look at the following page:
http://www.downeastmicrowave.com/cat-frame.htm