PHR-803T batch testing.. Windowless 803Ts?

I know that my post was a gross oversimplification, but it is useful for the non-scientific mind to have such broad strokes. We field questions about how much power one can expect at a given input. Info like this shows us that we cannot answer that question with any real accuracy. We hear about lasers being sold as xxxmW based on the drlava or the Zom-B graph. This shows that statements about output without a power meter are "guesses" at best. We have observed significant differences in output even among reputable builders. This explains why.

Add to this the many potential problems associated with harvesting the diodes and differences in power meters, and the potential range widens more!

Peace,
dave

I didn't say it was an oversimplification Dave.. I just posted the data, you summed it up....


It is true, that the power can not be guessed from a graph. Especially, when a graph was based just on one diode, that could be on either end of the spectrum..

Lava's PHR must have been a very good one, because it did 80mW at only 90mA.. That's pretty much a "freak" diode!
And keep in mind, that i calculated the above powers to match Glenn's Kenometer. On my meter, the diodes in this batch go from 70-95mW and the average @109mA is 84mW..


Luckily, we will soon know which is correct.. Or rather, we will establish our own "correct"...

AixiZ acrylic lens Igor?

IgorT said:

I didn't say it was an oversimplification Dave.. I just posted the data, you summed it up....

Click to expand...

Thank you. I realize that. Admitting it is an oversimplification is just me paying homage to the fact that MUCH more information can be "mined" from the data you have collected.

Things are so much easier without science getting in the way

Peace,
dave

jayrob said:

AixiZ acrylic lens Igor?

Click to expand...

Yeah. And all with the same lens, because i found the plastic lenses vary somewhat..
If the above measurements were with the glass lens, the average with a plastic lens would be much lower, than it is now.


At first i wanted to use a different current this time, but i left everything the same, so that i can compare with the previous batch, and because i have a system for diode selection, and a different current would just confuse me.. The measurements at 109mA give me a clue about which diodes to use, in order to reach a certain power level.

And even then, the results at a higher current can vary.. For example, i had two diodes, one did 83 and the other 84mW at 109mA.. When i set them to 143mA i expected to get a similar power.. But one did 112mW and the other 128mW.. (my meter, numbers on a Kenometer = +7%)

IgorT said:

[quote author=jayrob link=1214268330/60#66 date=1218388423]AixiZ acrylic lens Igor?

Click to expand...

Yeah. And all with the same lens, because i found the plastic lenses vary somewhat.[/quote]

Damnable confounding variable are EVERYWHERE!

I'm going back to using JuJu

It's soooo much easier -- bleed a chicken = get a date -- life is good

Peace,
dave

daguin said:

Damnable confounding variable are EVERYWHERE!

I'm going back to using JuJu
It's soooo much easier -- bleed a chicken = get a date -- life is good

Click to expand...

Haha.. Yeah, i kinda lost 5mW off my 6x diode. I got really scared, that my $350 diode is dying, and at only 174mW! :'(

But then i remembered i switched the lens.. The new lens was clean, but only gave me 169mW.. I managed to find another lens, that brought me back to 174mW..

I added some measurements at higher currents to the list.. Most are at 125mA, if the current is different, it is mentioned next to the measurement.


I also calculated the average of the previous batch. It's 90.48mW @ 109mA.. So the average is pretty consistent between the two batches..

At 109mA you get 90mW +- 15mW, but there was one diode, that only did 69mW in the last batch, which is 22.5mW below the batch average.



I also found one diode (the highest efficiency one), that actually climbs as it warms up. Last night, when i made it, it was very cold, and it always started out at 166mW (155mW on my meter) and slowly climbed to 176mW (165mW on my meter). I just measured it again, and it now starts out at 173mW, and climbs..

I don't mind the high power, but i think i shouldn't leave this one on for too long at once, so it doesn't commit COD.

Visually it starts out slightly bellow my 203mW 6x laser, and slowly matches it's beam brightness. In the end it's hard to tell them appart and a few times i even confused them, as the PHR seemed bighter!



EDIT: I checked the current on the weird diode, to see if it could be climbing for some reason and causing the power to climb, but it's completelly stable, i measured the forward voltage of the diode, and it's normal for the current (5.6V) and drops as it heat's up, just like it should. But the power goes up instead of down. Basically, power going into the diode goes down, but the power coming out goes up!

However, it does seem to have stabilised now. It starts out at 175mW, goes to 177mW and then back to 175mW, like i would expect.. But if i force cool it, it goes back to 166mW, the exact opposite of what every other diode i remember would do. My 6x increases in power by 2mW if i cool it in the same way.

It was very cold last night, when i made this laser. Now it's warmer, so it starts out at the temperature it first had to reach last night... Maybe that's why i haven't noticed this before.. Need to test some others under the same conditions.


EDIT EDIT: After climbing to almost 180mW several times, the weird diode is doing quite stable 176mW now (164.4mW on my meter).

I've batch tested multiple diodes from multiple group buys, and collected all data together. Here are my results. 24 (of the 26) diodes survived long enough to get some accurate measurements.

Power graph: (the three weak ones are KES-400A, ignore them)


Here's a list with raw data (sorted by P[sub]max[/sub])

Legend:

• I[sub]th[/sub] Threshold current in mA. The current at which the laser starts lasing (when the laser emits about 0.1-0.3mW)
• I[sub]max[/sub] Maximum recommended current found using my algorithm, in mA
• P[sub]max[/sub] Optical power at Imax, in mW
• @##mA Optical power at this supply current, in mW
• [ch916][ch951] Delta efficiency: slope of straight part of the curve, expressed in mW per mA

I've also made some graphs.

Sorted by Threshold current:


Sorted by I[sub]max[/sub]:


Sorted by P[sub]max[/sub]:


Sorted by Delta Efficiency (times 100 to match scale)



At last, I could only find a very weak correlation between I[sub]th[/sub], [ch916][ch951] and I[sub]max[/sub]. This is probably due to the big inaccuracy of the I[sub]th[/sub] measures (it's hard to find it as it requires extreme low mW measurements). I do consider this the holy grail as it lets you find the maximum power without using an expensive laser power meter, by using measurements that can be obtained with far cheaper equipment. I'm still working on improving the correlation formulas.

In addition to P vs. I, you should also try plotting dP/dI vs. I.  It will give a much clearer picture of when you hit a kink, and may even show some other kinks or slope changes you can't see.

Also, you just stopped when you hit the first kink in the chart, correct?  


Another idea about your calculations: a possibly better way to calculate efficiency since you clearly don't mind the trouble of taking measurements and doing calculations. Assuming the mirrors on both ends of the cavity are the same reflectivity (sure, why not?), multiply the slope of your power out/current in number by 2 (equal light out both sides. With the total power out, in units watts/amps, multiply by q/(hv) in units of Coulombs/joule. (q is electron charge, h is Planck's constant, v is photon frequency). This will give the "differential quantum efficiency", literally the number of photons emitted in the laser mode per electron injected into the device. Especially if you have a way to measure the exact frequency/wavelength, it could change the numbers slightly for the efficiencies. If you don't have a way and just put in 405nm for all of them, it won't change the order of the efficiencies, just give a number that is a different way to express it.

pullbangdead said:

In addition to P vs. I, you should also try plotting dP/dI vs. I. It will give a much clearer picture of when you hit a kink, and may even show some other kinks or slope changes you can't see.

Click to expand...

That's just the differential of the graph curve. I've got something that will even more accurately show upcoming kinks. In my "algorithm" I draw the (smoothed) curve of [ch951]-[ch951](n-1), or more primitively, P/I-P(n-1)/I(n-1). When a kink is coming, the function starts dropping faster until it drops below zero, that's the knee itself. Sometimes it drops linearly until it is below zero and starts dropping faster past that. I use the zero crossing itself as the maximum recommended current.

I started thinking after reading a thread by billg519. As my weak diodes are almost useless anyway I gave it a try and increased the current through the knee and above.

I just found out that one diode I deemed weak actually has some serious mode issues and is a lot stronger. I will immediately test the other weakies as I am not so afraid anymore of blowing them.



[edit]
Second one is even weirder. It actually drops in power after 100mA but eventually reached a stunning 220mA249mA!! (although with reduced efficiency)
After 100mA data samples every 1mA (phew, just measured 141 ~200 samples in total)


The discontinuities are from periods where I stopped for a few minutes and the laser warmed up.

My PDA battery is almost empty so I will test the final "weak" one in a few hours.

I made a video of the mode shifts of this diode. While doing so, I noticed at around 210mA there is another mode shift, so this knee there might actally also be fake and this might be a freak diode.

[Edit]I was right about the mode hop but it's not a freak. I'm happy anyway because I'VE GOT A BLU-RAY AS POWERFUL AS A LONG OPEN CAN!! The mode hop is even temperature dependent too. I updated the graph in the post above. (Press shift+F5 to refresh the image)

In the video I tune the current down from 220mA to 0mA. IRL it is much more noticable, but watch first the left growing a tiny bit the left, then the whole left side shrinks, then the whole blob moves left on.

[media]http://www.youtube.com/v/1WI_i2dz72U[/media]

While I was recording videos anyway, I took a shot of my current measurement setup:

[media]http://www.youtube.com/v/ukgxcHwStiU[/media]

From one of my textbooks:

"... actual P-I characteristics can have kinks, and the tend to be nonlinear. The kinks can indicate a switching between lateral or axial modes or an additional parasitic mirror in the device. ... Premature saturation of the output power may indicate the existence of current leakage paths that "turn-on" at higher current levels or excessive heating of gain material."

Yep, you found some mode shifting going on. Cool stuff, eh?

Premature saturation of the output power may indicate the existence of current leakage paths

I assume Premature means below the nominal current

Correct, before the theoretical max output power. Since these have all been through QC and binning at a factory somewhere, I doubt that part has as much to do with it. These are all going well above the stated power output, so who knows really what that theoretical max is for a given diode. I think that comes more into play at the factory when a few die from a given wafer will saturate and level off well below the printed datasheet values for power.

Once we're above the datasheet, which at 150mW CW we are, I don't guess there's really any way to know where is premature and where isn't for a given diode. So I guess there's really no way to know when you get to a kink if it's going to be the ultimate output knee or just a kink on the way to a higher output, except educated guesses based on experience or going past it to test it.