Laser Diode Driver IC with analog modulation capability up to 2Ghz ?

The only way I've seen this done is by directly injecting the RF into the diode. (I've only done it up to several hundred MHz.... the limit of my frequency generator.) AC couple into the DC current line, then a 50 ohm 'termniation' resistor in parallel with an inductor (to allow the DC current to flow) and this feeds the diode.

George H.

How much current do you need? What kind of laser?

An analog laser diode driver is as simple as a signal and a resistor.

If you need gain, there are tons of cheap MMICs that have 6-12 GHz bandwidth. Just stick the laser diode in series with the output pullup resistor, or AC couple the MMIC output into the laser and provide a separate DC bias path. Check out the MiniCircuits parts.

Most lasers don't modulate very well. They tend to have mode jumps, "kinks", in their current-to-light curve. The high-end optical RF links use CW lasers and lithium niobate modulators.

John

VCSELs can be fairly benign in that respect but they do have mild polarization mode hops.

Doing 2 GHz with a TO-can is no joke, even if the diode chip itself is fast enough, which it probably isn't. The problems are both electrical and optical. The main optical reason is that the level populations in the upper and lower laser states often don't respond fast enough. Ordinary cleaved-cavity Fabry-Perot lasers IME have relaxation peaks near 1 GHz, so they don't respond terribly well at 2 GHz. You'll get some response, for sure, but it's very hard to get nice looking edges or nice open eye diagrams at 2 GHz with those lasers. (Even 1.25 Gb/s is a bit of a trick.)

There are VCSELs and drivers available commercially that work well up to

On the other hand, if you just want to get 10% sinusoidal modulation at

Cheers

Phil Hobbs

Strong modulation should in itself reduce or eliminate mode hops (see Phil Hobbs' book). AIUI the "modes" then don't have time to get going. Needs to be a few hundred MHz or above.

It's like going over a washboard road. Above a certain speed the rattling in the car is "gone". But the car's handling gets iffy and it is much more prone to unauthorized departures from the road :-)

I've always wanted to try this up at higher frequencies. We put sidebands on a Sanyo 74DL201 (?) and observe it with a confocal Fabry- Perot. (It's an easy way to 'put a yard stick' on the spectrum as I'm sure you know.) The diodes have relaxation oscillations right about

George H.

t -

up

ser

mit

en

cal

in

is

ESD

xt -

Using a transformer to inject a little AC on a DC signal is an old trick for doing high frequency PSRR testing. Often RF generators can't supple the DC to run the DUT, so you roll your own transfer. minicircuits has nothing above 1.4GHz.

Hi Rafael, You would be smart to think about suggestions of the others also. I haven't worked up above 200MHz and... So the transformer idea is also a good one. How much modulation do you want?

The impedance of the diode (At DC) is pretty easy to estimate... It's

George H.

up

er

up

ser

mit

en

cal

in

is

Opps that number is a DL-7140-201. (Which are about to go out of production.. sigh.)

ESD

xt -

Bias tees are fine too. Note that the impedance of the LD chip will be very very low at 2 GHz--the total impedance will be dominated by the lead inductance. Another approach is wiring an RF transistor in parallel with the diode and driving that. It does require some attention to prevent oscillations though.

Cheers

Phil Hobbs