Fast And Gate, 1 Ghz

That would be like showing up at the Daytona race track in an AMC Gremlin :-)

You'd probably be looking at low capacitance RF transistors like the BFP620:

Although, depending on your precision requirements even the old BFS17 might suffice. It can do a little under a nanosecond. When using really hot RF transistors be careful, they can be oscillating without you even knowing they do.

I wonder if you could hack it all with something like this?

-Lasse

A PECL NOR is an AND for low-true... that's quite fast. ...Jim Thompson

Yeah, but you can't get them for a buck :-)

If the laser pulse comes in with gusto I'd be tempted to just amplify it as needed (maybe with a limiter) and run it point blank into the base of the RF transistor.

The laser is 4 millijoules. give or take. The NB4N316M looks good, I think we may start there...

Steve

I probably have some 48 year old MECL in my junk box ;-) ...Jim Thompson

Got to overcome that 25mV hysteresis though. For this app a hysteresis might not be all that great because it'll skew your timing. Maybe Lasse knows a version sans hysteresis?

Then there's the question on how to tap into the PD. If you can get >>25mV across 50-100ohms there, fine. Otherwise you'd have to roll a TIA but that's not a really big deal.

I'd go for one of the faster Analog Devices PECL-output comparators, ADCMP567 maybe, driving a pair of fast PNP transistors as the current steering thing. I've done it with MMBR521s with 33 ohm base resistors from the PECL drive. Some relatively slow closed-loop positive current source can drive the emitters of the pair, through a ferrite bead, or even just a resistor to a decently positive voltage. 5 or 10 mA is about right.

The LVDS receivers make nice comparators for some applications, but are relatively slow and have a lot of built-in-on-purpose offset.

John

One problem is that (AIUI) it needs to separate the pulse height from the duration. Gated integrators don't do that--they integrate, so they get the area under the pulse.

One possibility would be a PECL or LVDS line receiver with a constant-fraction discriminator trigger to drive pulse stretcher A, plus a linear amp driving pulse stretcher B. That way, you can get the average height and the duration separately. The constant-fraction discriminator makes sure that the pulse height variation doesn't change the measured pulse length very much.

Cheers

Phil Hobbs

Yes, either that or just limit the signal. If the laser pulse is small the simplest thing would be an amp with enough gain to get you past

If $10-20 extra are ok and power consumption isn't a concern a nice PECL solution should work. However, a hysteresis is not so cool here and many line receivers have some Schmitt behavior.

Most CFDs discriminate the leading edge of a pulse, and only do that well for some fixed width. I've managed to make a CFD that's pretty good over a 3:1 or so width range and a healthy range of pulse heights (for readout of an MCP/2d-delay-line-detector imaging thing) but that won't do width.

If one did a fast peak holder, the main pulse could be run through a delay line until the peak detector settled. Then compare the delayed pulse to 1/2 of the detected peak amplitude.

It's getting complicated.

John

What if Steve would take one or two cheapo MMIC and railed the last one into a Schottky so the amplitude is always the same?

Then amplitude variations would change the width.

John

Ok, what if he'd take three MMIC for a whole lotta gain?

:-)

If you need super precision a comparator without hysteresis would be the ticket. Provided that the PD setup itself is faster than the MMIC. If not you might as well use MMIC.

You would be better off in an asphalt truck. :(

, I

I just happend to see that datasheet and think that with the current mode output it might make for a simple solution.

analog has some comparators with ~1mV hysteresis and ECL,PECL or CML output just need to find one without onchip termination

-Lasse

Ok, the course test is to focus the beam in saline and see if it breaks down and creates plasma.

I did find one 3 Ghz clipper-limiter on ebay, but it limits at 22 dbm, a little too high for this guy:

Thorlabs FDS02

Now before somebody tells me to buy their mounted and biased one, I've used them before, and 2" of stripline, 1 choke, a 12V battery, and a switch is not my idea of a extra 100$.

Mini Circuits, as far as limiters go, was NOT encouraging. So we're looking at counting, or current switching, or pulse stretching.

As no good deed goes unpunished, I just got asked to build it. Hes a good friend, and so that request had only one possible answer... .

This oughta be fun...

Steve

I find their stuff to be rather high-price.

:-)

As Lasse mentioned Analog Devices offers good comparators. This thing ought to do the job, about $7 in small qty and Digikey has stock:

Digikey P/N is: ADCMP567BCPZ-ND

How many beers will he buy you?