Simple current inrush circuit?

Hi,
I've been doing some research on building my 4w C-mount red laser and it seems that several people have killed these diodes at or below rated current. The most popular theory about these untimely (and expensive) deaths is current spikes at power up. At present there are a few drivers available with soft start but none so far are rated for the max allowable current of 5a (miniaturized drivers for handhelds). I was wondering if anyone knows of a simple circuit that could easily be miniaturized and placed inline between a driver and the diode that would offer the needed surge protection. I was thinking a simple capacitor/inductor circuit could be effective but was hoping someone with some experience and knowledge on this might offer a good schematic I could build from.

Thanks!

Jmillerdoc said:

The most popular theory about these untimely (and expensive) deaths is current spikes at power up.

Click to expand...

I think you're getting ahead of yourself. Perhaps you should run your driver on a test load and look for current spikes before you go trying to fix a problem that may not be there.

Possibly, but it's still worth looking into making a simple circuit. Especially for some of us who don't have a occilascope anymore (me). Seems a series inductor with a 1-10uF Tantalum cap across the +/- would offer excellent spike protection. I'm just not sure what value of inductor would be needed for a " spike". Calculating a value for an AC frequency is pretty simple but not too sure how one would do this for a single spike in current. Even a Varistor I believe may be useful but again, not too sure when were dealing with these voltages and currents. All my past experience with these circuits were in building tube amps and such dealing with much higher voltages.

Back to your premise, yes there may not be a problem, especially with drivers built with soft starts. What I really wish could be answered is if anyone has proven current inrush spikes are the cause of diode death. I've seen people talk about it but I haven't seen anyone show proof of this claim. Not saying they haven't, just saying I haven't seen it. So to go even beyond what you're saying, it would be even better to have knowledge that the problem I'm looking for is really even a problem. I even metion this because some diodes easily handle high frequency switching, etc with currents slamming then at thousands a time a second.

If you put a series inductor in then you'll generate a very high voltage whenever current is interrupted. Very not good for your diode. If you place a parallel cap in with that inductor to try and absorb voltage transients you'll get ringing at the resonant point of the LC circuit, also very not good for your diode.

They make devices for protecting laser diodes, I think it is called Lasorb. Failing that, you could place a large value capacitor across the diode terminals along with an antiparallel fast diode, and a parallel high value resistor (say 10k ohm). If thermal issues are not an issue you can put a negative temperature coefficient resistor in series with the diode, but remember to adjust driver current for the HOT resistance and not the cold resistance. I'd just go with the cap, resistor, and diode, honestly.

Jmillerdoc said:

What I really wish could be answered is if anyone has proven current inrush spikes are the cause of diode death.

Click to expand...

"Diode death" can be caused by several things, and each death may be different. If your driver is crappy enough that it sends a current spike on startup, may I suggest using a better driving instead of adding 7 weird filters? Without a scope, you cannot check for spikes. You cannot check for effectiveness of the measures against them either, so all your work is wild speculation at this point.

Rules of thumb only get you so far. You might think a cap "smooths things," but it may just end up destabilizing the feedback loop. It depends on the load, driver, leads, etc.

If you're serious about this, here's step 1: get a scope. They're $50 on ebay if you're patient.

You could read through this design, but if you have no way to observe the operation of your circuit or its inputs it's like guessing how steep a cliff is while walking towards it blindfolded.