Laser diode wont work with certain power supplies

[blasterman]

Hey guys, one thing I use for testing power LED's is a big old thermistor.

LED's, especially 3watt Cree's are very sensitive to 'thermal inrush' as constant current supplies initially stabilize. A 700watt, 48volt constant current LED supply is *lethal* to a string of 3watt LEDs if connected 'hot'. They won't even flicker - just blow. However, a big thermistor in line allows me to 'hot' jump strings of LEDs with no problems. Even if the supply is rated for a minimum of 9volts like many are, and I connect it to a single LED at 3 Vf the LED power up normally where as without the thermistor it would go up in flames instantly. Obviously the residual current gets wasted as heat.

Wondering if a big thermistor would stabilize things for green diodes. They are certainly cheap enough.

 

[Cyparagon]

But then how can you ever be safe in turning on a diode if the regulator doesn't regulate for enough time to possibly kill a diode?

Not all supplies have a current spike. For those that do (or if you're not sure), just start the supply with the voltage at zero. There will be no spike that way.

Couldn't this be achieved by throwing in an inductor into the output of a DDL driver?

And then when you turn it off, inductive kickback kills your diode. The lm317 is good for constant current - no spikes. It's only some bench power supplies that have an initial spike, and that's only during initial power on.

A 700watt, 48volt constant current LED supply is *lethal* to a string of 3watt LEDs if connected 'hot'.

So don't connect them with the driver powered up. I don't see the point of wasting energy with the thermistor.

 

[WarFreak131]

If inductive kickback is a problem, couldn't you put a diode in anti-parallel with the inductor to protect it?

 

[blasterman]

So don't connect them with the driver powered up. I don't see the point of wasting energy with the thermistor.

Thanks for not responding to my suggestion - I'm glad you have this all figured out.

A bit hypocritical we're referring to the lm317, which is at best a 'ghetto' form of current regulation in terms of efficiency, but then call a thermistor a 'waste of energy'.

I suggested the thermistor as a means of correcting the transient problem, and once the thermistor is running energy waste is mimimal.

FYI - I use thermistors as per LED scenarios as a form of testing. If I have 22 3-watt LEDs in a string running at 48 volts stepping down the line to resolve a flaky solder with a current regulated driver resolves the problem quickly. This is not possible to do without a thermistor and often no pratical with a multimeter because the problem only occurs under load. While green LEDs are far more sensitive to voltage spikes than LEDs, the problem is the same.

 

[Cyparagon]

If inductive kickback is a problem, couldn't you put a diode in anti-parallel with the inductor to protect it?

Why do you insist on adding all this extra crap when it isn't needed? You're missing the point - the point is you don't need to improve the current regulation with an lm317. It is already good enough. A secondary point was that inductors in constant current systems will cause problems. Even if you manage to suppress the kickback, it can cause oscillations.

I suggested the thermistor as a means of correcting the transient problem

Fair enough. It will cause an error in load voltage reading though. In my opinion, it's easier to power on with voltage at zero.

 

[WarFreak131]

Why do you insist on adding all this extra crap when it isn't needed?

Becuase I don't want to destroy a diode, especially if it is expensive. It would be great if the power supply was set back to zero every time I was done working, but sometimes people forget.

the point is you don't need to improve the current regulation with an lm317. It is already good enough.

Well according to that guy Bob who seems to know his stuff, it may take a few microseconds for the regulation to kick in, which leaves potential time to damage the diode.

Even if you manage to suppress the kickback, it can cause oscillations.

Which is also solved rather simply. Adding a parallel capacitor to ground will suppress oscillations.

 

[Cyparagon]

Bob is talking about (some) bench power supplies.

JUST a capacitor is sufficient to stabilize the lm317 to begin with. Look at the datasheet.

 

[Bob_Boyce]

Bob is talking about (some) bench power supplies.

JUST a capacitor is sufficient to stabilize the lm317 to begin with. Look at the datasheet.

Absolutely correct Cyparagon. I was referring to the very common low quality chinese adjustable so-called "lab" or "bench" supplies that a number of ppl buy and expect to perform like the more expensive real thing. Some of those are so bad that they will even put out a spike when powered up, even when the voltage and current controls are turned all the way down. I evaluated several dozen from various manufacturers while I was in China from late October to mid December 2012

While I was in Shenzen China, I toured one "manufacturer" that was building claimed 60A 12V power supplies that looked really nice on the outside. On the inside they used a 10A toroidal power transformer feeding a pair of 3A bridge rectifiers in parallel just screwed to the bottom of a thin powdercoated steel cabinet with self-tapping drill-screws. The cabinet firmed up and felt more solid once the flimsy halves were screwed together. No heat sink compound on the rectifiers, and no filter caps connected to the output. They used a piece of what looked like 20 guage wire for the shunt on a 60A analog ammeter, with the length of wire calibrated so that 15A of actual current would read as 60A on the meter. It had a 15A output breaker on the back, a sure clue to anyone that knew enough to look that it was not a 60 amp power supply. Needless to say, we did not buy any power supplies from them.

LM317 based current regulators may be wasteful of energy if input voltage is much higher than that required to maintain good regulation, but with proper implementation they do work quite well.

Bob

 

[WarFreak131]

Bob, can you comment on the BK Precision model 1550?

 

[Bob_Boyce]

Bob, can you comment on the BK Precision model 1550?

I'm sorry, no I cannot. I did not test any BK power supplies, however, BK has been a trusted supplier of test equipment for many decades.

The purpose of our testing in China was to find a reliable supplier of good enough quality bench power supplies to support our expansion into the asian market. We met with many power supply manufacturers there and I was not very happy with what we were finding. Very few met our specifications, as we needed to be able to drive high powered laser diodes directly with no other driver circuitry. Soft start and surge suppression was of utmost importance. Mastech was the company that we chose to buy from. There were a few others that tested well, but quality control became the deciding factor in the end.

Bob