Recommendation for Near IR Work

[aromaru]

As this seemed a more geek oriented discussion area, I hope it isn't some offense to post this here.

I need to pulse approximately 9058.33 A, up to 500mJ, under conditions that might involve (depending on testing) a rarefied atmosphere. This is something I only need to do on a specific project, and I don't want to break the bank, as I have little there now.

I have been considering a DIY approach using Osram's SPL PL90 diode (standard operation: pulses of 100 ns pulse width at 1 kHz rate with 5 A operating current at TA=25°C, mighty overkill unadjusted for, I know); however, I don't know what circuit would be appropriate for what I would like to do (anything more brings costs to an abusively high range).

I would like some recommendations regarding the most economically feasible way to proceed.

[mixedgas]

1 Khz is a problem short of using a commercial pulse driver. Very high rate.
100 nanoseconds - You need a 10 Mhz driver bandwidth at 5 amps? (Make that 20 Mhz to be sure it can do it)

BTW, we use nanometers now, so divide by 10 , ie 908.533
Only about 10 of us would know what Angstroms are here.

Considering diodes drift in wavelength at .1-.4 nm per degree C.

So you need a grating stabilized pulsed diode at 5 amps.

That is a bit ambitious. Exciting a metal vapor or polarizing a gas for imaging?

www.avtechpulse.com

http://www.directedenergy.com/pdf/pc...data_sheet.pdf

Your not looking at exactly a inexpensive driver here!


Steve

[aromaru]

Thank you for the starting point, regarding an avalanche transistor, something I think I've only heard of once a decade ago.

Sorry about the angstroms, I've been using older spectral reference charts and hadn't thought it might be an issue; though, I am aware that nanometers are more common discussing wavelengths in general.

The 1kHz is actually more than needed; I was using the SOC from the SPL 90 guide -- hence my mention of overkill.
__

Post editing:

I am aware of range drift, but I need to be close enough to, to cover the next question, allow iodine to hit it's emission range before it escapes its copper companion.

Thank you, Steve,

-Aaron

[aromaru]

I've tried using the selection servlet recently, but the only stuff that came up in that range was a might, a very significant might, bit more expensive than I can afford, hence my seeking a circuit.

Thank you, though, for that extra bit of effort

- Aaron

[andythemechanic]

Hi,

just two quick notes:
1. 500mJ with ns Pulses out of a single Diode is simply not possible. With the mentioned diode you get something like 2.5µJ @ 100ns. I actually doubt that it is possible at all with any diode laser.
2. Diodes have a broad spectrum, several nm spectral width and the spectral position of it varys greatly with current and temperature. Even a feedback stabilized high power diode has a spectral width of ~0.3nm. Stabilizing a ns pulsed diode is quite difficult and expensive and will cost output power.

Sorry I don't want to destroy your project but that might be things to consider before building the power supply for the diode and buying that expensive device.

Andreas

[mixedgas]

Having locked gas lasers to iodine transitions in a previous life, this is not a job for a drifting diode laser. The near IR and visible Iodine transitions are broad compared to most materials, but not that broad.

Why 500 mJ per pulse? At 1 Khz, that is a huge amount of energy. You are possibly looking at a tunable nanosecond or femtosecond lab laser system with a external amplifier. These take up large amounts of optical benches.

Ie a Merlin, Spitfire, Tsunami, Mira, Opel etc

I know some people you can talk to, who use such systems, and can possibly take a better look at what you wish to do. Drop me a PM and I'll refer you to a professor who might have a better idea of what to do. He "owns" nano and femto systems. He works in a entirely different field, but would be a good person to start with.

For the most part, I've only ever worked on the pump end of fast systems. You need better advice then I can give.

Steve

[aromaru]

The 1kHz is actually more than needed; I was using the SOC from the SPL 90 guide -- hence my mention of overkill.
...
I am aware of range drift

The work I am interested in is in a similar vain of the attached file with a different target. Lower power can be matched by increased energy to the substrate. The extreme numbers are the claim of the manufacturer and the outer limit of usefulness to me.

[mixedgas]

I finally got around to reading all this.

SPL-90 has less average power then a 100 mCd red T1-3/4s LED. Its a rangefinding diode in a T1 plastic package.

That is a PEAK power at one shot on the data sheet. The .1 duty cycle is the problem.

Here, for a 905 nm wavelength and high energy, this is more what you need, something like a

http://www.coherent.com/Products/ind...BR-Ring-Series, Only you may need Kerr Lens Modelocking or Qswitching added to it.

That needs a coherent Verdi for a pump.

I suggest you collaborate with a professor who already has a laser CVD setup.

I'm sorry, if you only have a few thousand dollars for a system, you had best rent or borrow system time.

A typical high power diode array has a linewidth of 3-4 nanometers, so only a tiny fraction of the power will be in your pump band. This is assuming you wish to try active photo- disassociation and not just pure power ablation.

If tuning to hit a transition, You need a tunable Ti:Sapphire, Optical Parametric Oscillator, or Fiber laser.

Your idea has strong merit. However you need a narrow linewidth laser.

Just so you know, if your device existed, which it does not, its core would look like this, it would have a feedback and tuning mechanism attached, and would be water cooled.

http://cgi.ebay.com/100W-808nm-QCW-H...item415a09c740

Spontaneous emission would prevent such a high power diode array having its wavelength narrowed by grating feedback. The feedback would most likely blow the face off the diode at some point.

Good Luck,

Steve

[aromaru]

I greatly appreciate the work you've put in on putting together the requirements of what is needed, and I think I can obtain exactly what is necessary with modification on the techniques of what is discussed in the paper.

The work you've put into this is greatly appreciated. I have substrate trickery I've been working on might compensate for the range values I'm looking at, as the energy of transition from being at or under the value is more than sufficient to obtain the new bond without sacrificing the much of the existing one.

[andythemechanic]

Just so you know, if your device existed, which it does not, its core would look like this, it would have a feedback and tuning mechanism attached, and would be water cooled.

http://cgi.ebay.com/100W-808nm-QCW-H...item415a09c740

Spontaneous emission would prevent such a high power diode array having its wavelength narrowed by grating feedback. The feedback would most likely blow the face off the diode at some point.

Hi Steve,

it is possible to narrow the spectrum via feedback at such a device. In the company I work now since almost a year we do this all the time.
Have a look at the datasheet here: http://dilas.de/gdresources/download...dth_MF_TD1.pdf

Linewidth is however limited by diode imperfections like smile and also by the efficiency requirements since locking will reduce the efficency.

Regards
Andreas