How can I make a regular laser diode pulse in the nanosecond range?

To get an easily measurable photoacoustic effect, you might need a rather powerful light pulse. I'd recommend having a look at LIDAR light sources, they're typcally 900nm (not quite 850), are cheap, put out tens of watts and have ns rise and fall times.

A quick perusing of the internets turned up the TPGAD1S09H which is affordable and in stock. While normally driven with a dedicated GaN FET switch and bias/pulse shaping network, in the interests of simplicity, you may be able to use a chunky MOSFET gate driver instead as it'll already have additional logic to sharpen up your control signal's rising and falling edges, the UCC27614DSGR is one such module (although any fast low-side driver could likely be substituted)

In the diagram above, above all else make the red loop path as compact as you possibly can otherwise you're going to have trouble getting decent rise and fall times.

Cb should be sized so that it only holds enough energy for one pulse in order to minimise the risk of cooking the laser should your CPU decide to just leave the thing permanently on - note also that EN is tied to the driver's local supply? We're exploiting EN's under-voltage lockout capability as an additional safety measure.

Rb should be small enough to charge up Cb between pulses but be large enough to act as a fault current limiter (start with 50mA, that laser can't dissipate that much average power, it is pretty small). Ri should be set to limit peak laser current to something sensible (couple amps in this case), start with half an ohm and go from there.

Needless to say, simulate everything as best you can before turning anything on! High power in small spaces can release much magic smoke very fast...

You can get a well-specified TT 850nm VCSEL laser diode from an industrial supplier like Digikey for about the same price as the mystery Amazon product. Not sure if that matches your optical and power requirements, though.

Rise/fall time of 100ps and threshold current of less than 3mA so it should be easy to drive. Eg. Logic gate + resistor. 74LVC logic from 5V can drive 32mA in a few ns. Parallel a few outputs in the same package and get ~100mA.

Please let me know if this sort of DIY thing is possible

It's easily possible with modern gates such as an EXOR. Both inputs are fed the same signal except one of the inputs is slightly delayed by a few tens of nano second. This is easily done with an RC circuit. I've made 2 ns pulse generators using similar methods (see further below) but, when you get to this sort of speed you have to shop around for the right parts but, nevertheless, the solution wasn't a wallet-buster.

DIY - yes! but make a proper circuit board with a local voltage regulator and the usual decoupling capacitors on gate supplies. Regarding the laser driver, use a GaN MOSFET. If you really need to push things then maybe choose an EPC2038. I used this device to generate an 80 volt pulse onto a load. Rise time was about 1 ns and pulse width a few nanoseconds.

For the digital pulse generator (combined with GaN driver) I used the LMG1020 from TI: -

The +IN pin and -IN pin were used with an RC network to generate a thin pulse from a regular square wave: -

But, like I said earlier, for a pulse of width circa 60 ns you can get-away with fast logic devices that are more commonly available. However, I'd still use a GaN FET for the laser interface.

I wanted to try making my own nanosecond pulsed laser from a regular 850nm wavelength laser diode from Amazon

Don't buy stuff like this from Amazon unless you are prepared to waste your time and money. The Amazon page delivers nothing about how quickly the laser can be activated/deactivated and, it's quite unlikely that it will meet your needs. Use reputable suppliers who produce a data sheet for their devices and, have a recognizable quality system, technical support and provenance.

I fear you need need to spend more money. Let us reason together. Your proposed laser puts out about 5 mW CW. If you modulate it with a 70 nsec pulse, your pulse energy will be .005 W x .00000007 sec, or about 0.35 nJ per pulse. Are you really intending to work in the nanojoule range? (That's an honest question, by the way, although I only ask it because I have my suspicions about the answer.) Now, it's true that you can probably overdrive it some because the duty cycle is so low that thermal problems aren't likely. Nonetheless, if you had hopes of producing a i KHz train of 50 J, 70 nsec pulses (total power 3.5 mW, less than the CW output), I'm afraid it doesn't work that way. Plus, beam delivery seems to me to be an issue. I assume you need a focussed beam at the target, and this sort of laser is not known for its great beam quality. It's generally convenient to couple the laser into a fiber optic, and then feed that to a compact optics package at the target area, so you need to consider that.

If I'm wrong, and you can use that module, then I think your path is pretty simple. I'd start by operating the laser and measuring the operating current. I'd expect something in the range of 50 mA, and if it's less than 100 mA you should be OK. So what you do is go on eBay and get yourself a cheap pulse generator. You should be able to get something in your pulse rate/width range for less than 100 bucks, and possiby under 50. Look for an output amplitude of 5 volts into 50 ohms. Your laser diode will have an operating voltage of about 2 volts and current of 50 mA, so 5 volts @100 mA will do nicely. Use a very short connection from the pulse generator to the diode, a resistor sized to give the measured current at about 3 volts, and a reverse diode like a 1N4148 (NOT A iN400X !!!!!) soldered to the diode to protect against reflections. A cheap pulse generator will typically have rise/fall times in the 5 nsec range, so if you keep your cable length under a few feet you should be OK, although shorter is definitely better. Less than a foot is best. Use twisted pair or coax. Also be aware that if you have programmed a 5 volt output the generator may well put out 10 volts into an open load (it assumes a 50% drop into 50 ohms) so your setup will actually draw more current than you expected since its effective resistance is greater than 50 ohms. Buying a couple of extra modules as insurance is always a good idea when trying to do something you don't know much about and trying to simultaneously do it on the cheap. They don't call it a learning curve for no reason.

In power applications (cutting metal etc.), a Diode-Pumped Solid State (DPSS) nanosecond-order laser works like an optical capacitor. An array of matched LEDs are permanently powered and, in simple terms, the beam is kept on a path to build up strength by continually bouncing through gain modules between end mirrors. Periodically, the beam is redirected out of the path and down a fibre-optic cable to its destination. So the electronics isn't switching the LEDs at all, they're always on. It's the redirection using adjustable polarisation of liquid crystal that's the fast part.

So trying to switch your LED on and off at those rates isn't using the same principles as a DPSS.

You could have a go - using something that can create the 66 ns pulses (Arduino?) and increasing the current until it breaks. I've successfully pulsed, well, modulated a cheap laser pointer diode at a few MHz. The average power will be incredibly low, imperceptible to the eye, will that be enough?

If this is for a bio experiment, not a laser project, you probably want a aource with accurate power, stability, pulse width specifications, that's stable over temperature etc, so you can concentrate on the rest of the experiment. You're about to take a side trip into laser land. At minimum you'll need $1000s of test gear (some of it also from Thor labs) to verify the laser is producing what you need, and it might still be unreliable with cheap diodes. I'd say it's only worthwhile if you need more than 10 or 20 lasers.

You have to make sure that your 850 nm LED is capable of pulsing at that rate. The rise and fall times are also important. A LED used in Optical Fibre communications meet these requirements.

I found this article on using a CD/DVD/BluRay optical pickup unit: maybe food for thought.

Hacking CD/DVD/Blu-ray for Biosensing AND ALSO

Open access

___________________

Other:

It was a long time ago in a galaxy far far away that I was sitting in a Philips lab with an engineering sample of the first generation DVD ROM mechanism from Panasonic in Japan, trying to get a CD player servo to focus on and track a DVD.. Heavy machined blocks of aluminium rather than plastic, I seem to remember discovering by accident that one way of seeking on the drive was to bounce the sledge off one of the endstops, heading off fast in the wrong direction sometimes landed you where you wanted to be..