12x S06J 405nm Diodes Slowly Current Slowly Decreases

Hello everyone -

A lot of useful information here! I have some questions about the 405nm diodes.

I bought some diodes from DTR and pressed them into heat sinks. These are then screwed to a water cooled plate.
I have a current regulated power supply (not specifically made for diode driving)

I attached a single diode to the power supply and turned it on to about 35mA, right at the level where it begins to emit. I noticed, though, that after some time the light was getting fainter and fainter unit it was almost gone. All this time the current the diode was pulling decreased to almost 20mA.

I tried another diode which was rock solid at 35mA. I then move up to ~96mA, and began to see the diode move from 96mA to 94mA over the course of 2-3 minutes.

I do not have a capacitor soldered across the (+) and (-) terminal. Should I do this? I have seen some threads suggesting 2.2uF. It could be my power supply with voltage spikes.

Is this normal? Any other suggestions?
I have read through these thread so far:

405nm S06J getting dim diode
12X Diode Life
12X BR Diodes
Parts issue, Benboost and more...

Thanks

Either you are not operating your power supply correctly, or your power supply is defective.

Set both knobs to zero.
Connect load.
Increase voltage knob to max.
Now the current pot adjusts the current at the load.

If the current changed then either it is not running the diode in constant current or there is an issue with the power supply. Constant current should be constant. The only thing that should be variable would be the forward voltage based on resistance changes within the diode when the temperature changes.

I don't think it can be chalked up to running in constant voltage mode + diode heat. If voltage is constant, and the diode heats up, current draw should increase, not decrease. I also can't imagine threshold current of 35mA causing substantial heat.

It could be the resistance of the lead wires increasing with heat, coupled with constant voltage, because that would increase the voltage drop pre-diode, decreasing the voltage hitting the diode, and thus decreasing current draw. But again, what kind of horrible wires would you have to be using as PSU leads to experience a meaningful voltage drop at 35mA of current? That part doesn't add up to me.

I think you miss the point of DTR's response. If the current is changing, then your power supply is not operating in constant current mode. I would add one step to Cyparagon's procedure;

Set both knobs to zero.
Briefly short the output of the PS to ensure all capacitors are discharged.
Connect load.
Increase voltage knob to max.
Now the current pot adjusts the current at the load.

The internal caps always have a bleed resistor. Some voltage may remain, but it will be far less than a volt, which is harmless to the diode. shorting the output isn't necessary, but if you're ultra-paranoid, it won't hurt.

Cyparagon said:

The internal caps always have a bleed resistor. Some voltage may remain, but it will be far less than a volt, which is harmless to the diode. shorting the output isn't necessary, but if you're ultra-paranoid, it won't hurt.

Click to expand...

I think all power supplies have bleeders but not all power supplies will go to zero volts instantly when switched off. I have an old HP dual output PS and it takes about 6 seconds to go to zero volts if it's switched off without a load. This bit me once when I was testing a 30mW red diode (remember those?). I set the current @ zero and the voltage @ ~ 12V switched it off to connect the diode and *flash*. When I switched the PS back on and dialed up a little current I had a lovely (and expensive) red LED.

Better safe than sorry!

WizardG said:

I think you miss the point of DTR's response. If the current is changing, then your power supply is not operating in constant current mode.

Click to expand...

I think you missed the point of MY response.... there's no dispute from me that this PS is not delivering constant current. That's plain obvious from the fact that....THE CURRENT ISN'T CONSTANT

The point of my response was that the PS operating in constant voltage mode doesn't entirely explain the phenomena either. Constant voltage mode, just like constant current mode, doesn't seem to be compatible with what this poster reported experiencing.

I do however think YOU missed the point of DTR's response... DTR did not in fact say that the power supply isn't operating in constant current mode. What he said was:

DTR said:

If the current changed then either it is not running the diode in constant current or there is an issue with the power supply.

Click to expand...

See that "or" that I've made obnoxiously large in the quote above? DTR suggested two explanations. Either, the power support is not running the diode in constant current, or there's some other issue with the power supply itself. Options, two different possibilities - lack of constant current operation, or some other PS defect / fault / problem.

In my response, if you care to read it (and think about it) instead of impulsively posting an obnoxious response without putting thought into what you write, you'll recognize that I was suggesting that this issue doesn't seem to be explained by constant voltage mode rather than constant current mode. IE, suggesting that the problem is something else, some other defect, some other problem with the setup or with the PS itself. IE, suggesting that the issue falls into the second of the two possibilities DTR was referring to.

Get it?

Your point is well taken rhd.

"I was suggesting that this issue doesn't seem to be explained by constant voltage mode rather than constant current mode. IE, suggesting that the problem is something else, some other defect, some other problem with the setup or with the PS itself. IE, suggesting that the issue falls into the second of the two possibilities DTR was referring to."

That second possibility being some issue with paperslammer's power supply.

My post wasn't meant to be obnoxious. I simply didn't (and still don't) know if paperslammer had his supply set up properly to deliver CC instead of CV.

Presuming for the sake of argument the PS is set up correctly the clue may be where the original post mentions that his PS wasn't originally meant for driving laser diodes. What is the stability of the PS in question? Laser diodes, like just about any semiconductor diode are very non-linear at low currents. A very small drift in the power supply could give the kind of results mentioned in the original post. Perhaps paperslammer ought to hook his supply to a dummy load and see if there is a similar drift.

Finally, rhd, the tone of your reply suggests to me you took considerable offense at the "obnoxious' nature of mine. I'm really just trying to help the OP figure out his mystery. I hope this response is more helpful than the first as I've suggested (politely, I think) one means to find out if the power supply itself is drifting.

I'm not going to apologize for my first response but I'll offer you a :beer:.

Thanks all for the suggestions.

Let me clarify a few things.

• I am turning to voltage to 3.8 volts, which just barely turns the diode on to ~35mA. Sometimes it is 32mA, sometimes 29mA.
• You are correct, I am not limiting the current, but I have a limit of 50mA set so that the supply cannot drift above 50mA. So yes, it is voltage driven right now.

But keeping the voltage at 3.8 volts, why am I seeing this current drop? I have seen one diode drop from 33.4mA down to 32.5mA slowly over about 10 minutes. As the diodes heats up, it should draw more current, as a previous poster also pointed out.

Currently I'm going to guess that it's the power supply oscillations that are slowly degrading the diodes. I'll try another power supply and see what happens.
Next step would be soldering a capacitor across the + and - pins of the diode. I will try 1uf.

What happens when you drive it normally in constant current with the supply set to 6V and the current to zero before connecting the diode and then connect it, turn the supply on and slowly turn the current up?

And what happens if you hook your power supply to a dummy load made of six 1n4001 diodes in series?