Simple driver to single-pulse diode in nanoseconds?

Glad I found this forum; seems to be filled with useful info.

There appears to be quite a few driver designs around, and unless I overlooked, none seem to be designed to emit a single pulse with:

- Nano-scale rise/fall times
- Nano-scale duration of (single) pulse
- Micro-controlled


Has anyone come across any driver that can do the aforementioned?

Thank you


ps: Green diode extracted from eBay-type laser pointer!

raz said:

Glad I found this forum; seems to be filled with useful info.

There appears to be quite a few driver designs around, and unless I overlooked, none seem to be designed to emit a single pulse with:

- Nano-scale rise/fall times
- Nano-scale duration of (single) pulse
- Micro-controlled


Has anyone come across any driver that can do the aforementioned?

Thank you


ps: Green diode extracted from eBay-type laser pointer!

Click to expand...

not true... there arent green diodes available... what you have is an ir diode. so if you power it up it will look like it is off but really on.. do not stare in to it ever.

michael

It would be helpful if you describe your intended application. Most drivers we talk about here are capable of microsecond speeds, which are useful for laser projectors and such.

Going to the nanoseconds is a couple of orders of magnitude faster, but not impossible per se... so tell us a bit more!

Hey BenM,

Application is fault-injection on micro-controllers through optical means.
In other words, light affect the way transistors work.

Goal is to test the theory in a controlled environment: single pulse, small width, analyze which areas are affected and how it affects the micro-controller.

Has anyone driven the cheap "green" (ok, IR) diode off ebay by bypassing the driver that comes with the module? Any schematics around? The ones I've looked at seemed to be for other applications.

I believe this "green" laser is DPPS.


Thanks

Sounds like an intriguing application. Check this driver out:
Discrete high-speed laser modulator schematic
off the bat it can get you about 500ns pulse width, if you replace the transistors with RF transistors you could get where you want. It's not micro-controlled but can accept a control signal from a microcontroller.
When you are building a circuit to operate this fast, you should use very close spacing of small surface mount components, bypass capacitors, and mount the diode right to your prototype PCB or use inductance-free stripline to the diode.

Edit: another option, if you don't care about precise control is to set up a capacitor charge/discharge circuit so that a capacitor can be charged to a pre-determined level and then discharged through the diode via a fast switching thyristor.

raz said:

Has anyone driven the cheap "green" (ok, IR) diode off ebay by bypassing the driver that comes with the module? Any schematics around? The ones I've looked at seemed to be for other applications.

I believe this "green" laser is DPPS.

Click to expand...

I have - its quite simple. Just choose a current equal to what the original driver drew from the battery and use your own driver to supply that. Its quite simple to do.

As far as schematics go: The ones i use aren't geared towards high speed operation. They work up to 100 kHz or so, perhaps pushable to a megahertz, but thats a very far cry from nanosecond pulses.

I suppose you could use the standard 2-transistor current sources for very high speeds with the right transistors though, so its just a matter of supplying a fast enough pulse. With transistors like BFR96S it would be feasible to get them to work in the nanosecond range are current sources. The capacity of the diode junction could pose a problem though.

Usually the way to high speed pulses, is a Bias T and a pulse generator from HP, HighLand Technology or AvTECH etc.

The Bias T lets you put a DC bias onto the diode, and then AC couple the pulse into the diode. The DC sets right below the lasing threshold.
The Bias T also normalizes things to 50 ohms, allowing use of a commercial pulse generator.

My good friend Dr Lava has it right, a current source that is redirected away from the diode as needed. Usually 3-4 pairs of RF grade diversion devices are used to reduce capacitance at shorter pulses.

Watch out, some place around 1 Ghz to 2 Ghz the diode becomes a very nasty load without proper impedance matching, and becomes a LED for the duration of the pulse.

Also watch your average power, the current density of many diodes is already at the internal limits of the die, you need to back off to 1. avoid the resonance effects above, and b to keep the carrier density down.

As sonnet and OC48 use these kinds of pulses, driver chips are out there, but you really need a special diode package at that point.

Of if your desired bandwidth is under 20 mhz or 100 mhz, you can consider a acousto-optic modulator (usually 7-20 mhz) or a electro-optic modulator (100 mhz or more bandwidth)

As you wish for green, you may need to consider the size and storage time of the Ir to green conversion process, and use either a "microchip" laser, a very short cavity, intracavity passive qswitching, or a external modulation. Microchip lasers have extremely short cavities and "pop" out a self qswitched 3 to 7 nanosecond pulse a few microseconds after the pump diode turns on. The ND:vandate in a cheap laser pointer module has a storage time of 800 uSec or so, and thus you may need to hold the pump diode just below lasing threshold and then drive it to lasing....

What are you probing? if Silicon, 750-820 nm IR is just fine, but otherwise you might need to tune your wavelength to match a bandgap of your semiconductor.

see for example:

Avtech Electrosystems - Nanosecond Waveform Generators

Highland Technology Precision Electronic Instruments

Bias Ts are available from many vendors including Thorlabs...

http://www.ixysrf.com/products/mosfet_driver_ics.html

Steve

The above is good advise - i would avoid using a dpss laser. There is some buildup/delay between applying current and green light coming out. Normally this is not noticable (in a laser pointer etc).

Using a 'standard' green dpss diode/crystal/etc set i've noticed this delay with scanners myself. Its not a lot of time, but when you scan a fast yellow pattern with some blanking, its noticable that the red laser fires up slightly faster than the green one.

I'm not sure about how long the delay really is, but its defenitely in the order of many microseconds to make it visible like that. Holding the dpss just on the threshold instead of turning it off completely might remedy this for scanners, but i'm quite sure the lag would still be a problem in nanosecond applications.

Don't know if this is what you're looking for, but try this:
Optherm

A 12x BRD burner writes 432 million bits a second, which is 432MHz. Take the reciprocal to get one pulse per 2.3 nanoseconds. I wonder how viable hacking a simple burner would be. Just tell it to "write" 010101010101 ad infinitum?

Perhaps its feasible to do just that, with the existing drive circuitry. I'm sure there are plenty of safeguards preventing it, but if you could overwrite those in firmware, the hardware from a bluray burner should be capable of doing rapid pulses at desired times.