Fast LEDs for optical pulse generation?

Question

Hi:Anyone working with fast LED pulsers lately?  I'm looking for LEDs that can deliver 10-100ns pulses.  Hopefully some SMT package can do this. Prefer visible range, but near IR might be Ok.  The closer I can get to 10ns the better.Of course, I would like a lot of photons.  I am looking into transmission line pulsing circuits to deliver very short several amp pulses.The purpose is to create a short optical marker pulse for setting up gated intensified cameras.A slightly different project would be to be able to attempt to substitute an LED for a pulsed YAG laser to create pulses of light to illuminate very high velocity events such as fuel injection spray visualization.For this a

John Larkin · Accepted Answer

LEDs tend to be slow, but you might get below 100 ns driving a small led or an array. You'd likely need a push-pull driver to blast the diode capacitance both ways.

A cheap 850 nm VCSEL laser can deliver a milliwatt or two of optical power with risetimes in the 200 ps range. Drive one from a TinyLogic gate through a resistor and it should be easy to get, say, a 3 ns light pulse. ECL, maybe an EL89, would be even faster.

How much optical power do you need? An ICCD shouldn't need a lot of light.

Whose ICCD camera? We make the timing electronics for brand A.

John

bill.sloman · Answer

Been there, done that - with HP-2400 opto-couplers. A bit of reverse bias (from a diode clamped inductor) helped appreciably in getting a worst case toleranced 10Mbit/sec link running. The boss ran it it up to

18MBit/sec (over 18 metres of ribbon cable) before the one he tried fell over. I read him the riot act - if the sales guys had ever foound out about it, that 18MHz would have ended up in the specs.

Sounds good to me.

Reverse biassing an LED into breakdown is reputed to give similarly fast edges - I'll dive into my paper files next week and should be able to find the reference. The paper didn't say anything about how long the LED lasted when abused in this way.

Chris Carlen · Answer

Is it the lead inductance, or the semiconductor physics (large junction C, etc.) that causes this?  I suppose then that the larger the LED die and light output capability, the slower the thing is.The VCSEL might be the way to go.  I actually have a visible one.  I shall look in to these further.We have a lot of Xybion, some Princeton Instruments, and maybe another.-- Good day!________________________________________Christopher R. CarlenPrincipal Laser&Electronics TechnologistSandia National Laboratories CA USANOTE, delete texts: "RemoveThis" and"BOGUS" from email address to reply.

Fast LEDs for optical pulse generation?

Join the Discussion

Didn't find your answer?