Ok.. So, the driver is rock steady. I tested the entire input voltage range and let it run for half an hour. The current was the same the entire time, no matter what i did, and the output is very clean.
So good news: It wasn't my driver..
Bad news: It wasn't my driver..
The reason for the current shooting up before, was the diode not being a diode anymore, and it's Vf dropping below the driver input voltage.
The diode died at 169mA immediatelly after a power-up because of the power-up related stresses, combined with an unusually high optical output, and not because of a fresh battery causing a higher current.
When i first made this one, i set it to 180mA, and it was doing 180mW. I was scared the optical power alone would kill it, so i lowered it to 170mA, where it started out at 169mW.
169mW after a plastic lens means a lot more behind it. And maybe all this optical power, combined with reflections from the lens, could be the reason, or at least a part of the reason for the slow deterioration and death.
Your high current diodes have a slightly lower efficency and put out less power. Maybe the lower optical flux at the die is what's keeping them alive..
My other 170mA diode is also not showing any signs of deterioration yet, and it also has a slightly lower efficiency. Maybe the average diodes will turn out to be a tiny bit sturdier. But it did go through less use.. So i guess we'll see.
I will suspend the high current experiments for a few days, and then start again, with the AR coated lenses..
Or maybe i should run one without a lens..... Only use a lens to measure the start power, and then once every day, to check for any deterioration..