What are the tricks to finding the diodes max???

[k-shell]

I am wondering how you figure out what a diodes max output is,

the only one i am aware of is finding the knee using a power meter.
but is that the safest way?

and is there any way to tell if you're pushing it too far? what signs do you look for if you are over powering it (or if its over heating)?

 

[suiraM]

Re: What are the tricks to finding the diodes max?

I've successfully used IR thermometers to track the internal temperature in the case. If it has a photodiode, you can forward bias that and measure the Vf drop to get the temperature. Otherwise, a thermistor or the like is as close as you get, I suppose. Thermal overload is one possible failure mode, but if you've got a rig with good heat sinking, that shouldn't be an issue.

Igor noted that a lot of BR diodes don't seem to have an actual knee, but I guess you could try delta temp over delta Ild (that's one I haven't seen anyone here trying to measure), front mirror albedo (just watch out so you don't hit the emitter, or you may have an interesting day), or even a more sensitive power meter with very careful steps (diverge the beam onto a Hamamatsu S1337-66BR photodiode; it is linear over 10 decades of intensity, even below 400nm- they use these guys in the LHC).

When you see the actual knee occuring on the BR diodes, the way I understand it is that you're really seeing COD happening in slow-motion, rather than an oversaturation knee of the sort you would see on a red diode. Diverge it onto a wall at a distance, and you should be able to see the defects forming. Of course, that would be destructive testing and useless for finding out how hard you can drive a still-living one.

If you're willing to sacrifice one, rig it up with a beamsplitter between the photodiode and a diverging lens for projection, then watch the temp, the Ild and the albedo with a datalogger. Once the projection shows damage starting to occur, you can search the dataset for some sign of it before it happened, and use that for the next diode. I'll see if I can find time to yank a beamsplitter from one of the blu-ray sleds and run this experiment for myself if anyone is interested, but it won't be this month.

 

[k-shell]

Re: What are the tricks to finding the diodes max?

I've successfully used IR thermometers to track the internal temperature in the case. If it has a photodiode, you can forward bias that and measure the Vf drop to get the temperature. Otherwise, a thermistor or the like is as close as you get, I suppose. Thermal overload is one possible failure mode, but if you've got a rig with good heat sinking, that shouldn't be an issue.

Igor noted that a lot of BR diodes don't seem to have an actual knee, but I guess you could try delta temp over delta Ild (that's one I haven't seen anyone here trying to measure), front mirror albedo (just watch out so you don't hit the emitter, or you may have an interesting day), or even a more sensitive power meter with very careful steps (diverge the beam onto a Hamamatsu S1337-66BR photodiode; it is linear over 10 decades of intensity, even below 400nm- they use these guys in the LHC).

When you see the actual knee occuring on the BR diodes, the way I understand it is that you're really seeing COD happening in slow-motion, rather than an oversaturation knee of the sort you would see on a red diode. Diverge it onto a wall at a distance, and you should be able to see the defects forming. Of course, that would be destructive testing and useless for finding out how hard you can drive a still-living one.

If you're willing to sacrifice one, rig it up with a beamsplitter between the photodiode and a diverging lens for projection, then watch the temp, the Ild and the albedo with a datalogger. Once the projection shows damage starting to occur, you can search the dataset for some sign of it before it happened, and use that for the next diode. I'll see if I can find time to yank a beamsplitter from one of the blu-ray sleds and run this experiment for myself if anyone is interested, but it won't be this month.

Good ideas, I really don’t want to sacrifice one just to push the others to their max, that’s a job for you guys besides, all diodes are different... right? So even if I destructively tested one, the other may not have the same I/(mW output) trend.
But I still would like to monitor one of my diodes as I approach 170 to 180 mA and know if I should stop at any point. I will have it heat sinked with thermal interface material (TIM) like arctic silver (between the diode/aixiz housing/heatsink). My duty cycles will be short and if possible I will have a DMM in series with the diode (when testing). I’m not sure if the DMM thing is something you guys do, but the configuration would be like this:

I use a Fluke 87 DMM and would trust its accuracy with my life!

So there’s no knee you say? Well. The members here have given a good few approximations as to how far they can be pushed. So maybe ill just stick with them. Besides, I’m not in any race for output power. At least not yet!

What does COD stand for?

What does Ild stand for?

And when you say they use the Hamamatsu S1337-66BR in the LHC, you're not referring to CERN are you???

Thanks for the info!
K-Shell

 

[suiraM]

Re: What are the tricks to finding the diodes max?

COD = Catastrophic Optic Damage.
Ild = Current (I) through laser diode (ld).

Hamamatsu supplies quality parts ranging from $2/ea and up, including optics and semiconductors that are used both at CERN and in the LHC.

The S1337-66BR costs on the order of $60 at my local supplier (which is usually overpriced), and is a highly linear photodiode on a ceramic base with a 5x5mm detector area. It is linear over 10 decades of light intensity, and has a sensitivity of 20% at 400nm compared to the peak (their datasheet contains the response curve, which is fairly smooth, as you'd expect for that sort of money). To lower the dark noise and improve the measurements, you can try sticking a tiny peltiér under it and taking it down to the lowest temp with kink-free curves.

I expect the trends would be similar between diodes, but the specifics would probably vary a lot.

If you have a 405nm attenuating filter or somesuch to spare, you can try this: get a cheap webcam that actually sees the 405nm. Unscrew the lens in a room with reasonably clean air, pointing the thing downward. Get the filter in front of where the lens used to be, and either tape (if it's an expensive one) or glue (otherwise) it to the body of the webcam. Point the laser straight at the detector chip. Power it up slowly. See if you find any points where the beam profile changes noticeably and repeatably. The first such point should be the lasing threshold.

Similarly, if you have a diffraction grating or prism, you can try directing the diffracted beam at the webcam in the same way, and see whether it happens to do any weird mode hopping (weirder than usual, that is; it will shift some due to the current and temperature variation, but that's likely to follow a reasonably predictable pattern) when the power gets very high.

Just don't run the diode so hard that you end up unintentionally sacrificing it.

Anyway, just some thoughts on things to try.

I'm working on setting up my diode lab now, so I'll probably get around to this stuff eventually, if nobody else does.

 

[Ryo]

Re: What are the tricks to finding the diodes max?

I've hooked up my setup the way you have shown, with a Fluke MM in series to measure the current.

 

[k-shell]

Re: What are the tricks to finding the diodes max?

I've hooked up my setup the way you have shown, with a Fluke MM in series to measure the current.

cool, yea i wasnt sure if that was a proper way of measuring it as i've always wondered if the meter its self has any significant resistance which might be enough to change the actual current when its removed/installed???

 

[pullbangdead]

Re: What are the tricks to finding the diodes max?

You can also just put a 1 ohm resistor in series where you have the DMM, and then measure the voltage across that resistor. That way, it's guaranteed to be only 1ohm resistance, and it's still a very accurate reading that is more in situ and less invasive.

 

[k-shell]

Re: What are the tricks to finding the diodes max?

You can also just put a 1 ohm resistor in series where you have the DMM, and then measure the voltage across that resistor.  That way, it's guaranteed to be only 1ohm resistance, and it's still a very accurate reading that is more in situ and less invasive.

So in that illustration I made, just replace the diode with a 1ohm (1/4 watt) resistor?

1ohm load simulates the diode and won't damage the driver?

 

[pullbangdead]

Re: What are the tricks to finding the diodes max?

No no, replace the DMM with a 1ohm resistor. A 1ohm resistor in series with the laser diode.

That way, you have the current going through the laser diode, you only add 1ohm of resistance to the overall load, and you can measure the voltage across that 1ohm resistor to find the current going through that resistor. And the current through the resistor, since it's in series with the laser diode, will be the same as the current going through the laser diode.

 

[k-shell]

Re: What are the tricks to finding the diodes max?

No no, replace the DMM with a 1ohm resistor.  A 1ohm resistor in series with the laser diode.  

That way, you have the current going through the laser diode, you only add 1ohm of resistance to the overall load, and you can measure the voltage across that 1ohm resistor to find the current going through that resistor.  And the current through the resistor, since it's in series with the laser diode, will be the same as the current going through the laser diode.

gotcha,

No no, replace the DMM with a 1ohm resistor. A 1ohm resistor in series with the laser diode.

That way, you have the current going through the laser diode, you only add 1ohm of resistance to the overall load, and you can measure the voltage across that 1ohm resistor to find the current going through that resistor. And the current through the resistor, since it's in series with the laser diode, will be the same as the current going through the laser diode.

gotcha,

measure here
+v -v
[ch9660] [ch9660]
Driver-----1ohm------Diode
|_________________|

Similar to this diagram? so you're measuring the voltage across the resistor then calculate current using V/R. or in this case V/1 right?

BTW thanks for the idea!

Similar to this diagram? so you're measuring the voltage across the resistor then calculate current using V/R=I. or in this case V/1=I right?

BTW thanks for the idea!